repo
stringlengths
6
65
file_url
stringlengths
81
311
file_path
stringlengths
6
227
content
stringlengths
0
32.8k
language
stringclasses
1 value
license
stringclasses
7 values
commit_sha
stringlengths
40
40
retrieved_at
stringdate
2026-01-04 15:31:58
2026-01-04 20:25:31
truncated
bool
2 classes
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/sequence/src/state/command_status.rs
default-plugins/sequence/src/state/command_status.rs
use zellij_tile::prelude::*; #[derive(Debug, Clone)] pub enum CommandStatus { Exited(Option<i32>, Option<PaneId>), Running(Option<PaneId>), Pending, Interrupted(Option<PaneId>), } impl CommandStatus { pub fn get_pane_id(&self) -> Option<PaneId> { match self { CommandStatus::Exited(_, pane_id) => *pane_id, CommandStatus::Running(pane_id) => *pane_id, CommandStatus::Pending => None, CommandStatus::Interrupted(pane_id) => *pane_id, } } } impl Default for CommandStatus { fn default() -> Self { CommandStatus::Pending } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/sequence/src/state/execution.rs
default-plugins/sequence/src/state/execution.rs
use crate::path_formatting::format_cwd; use crate::state::{ChainType, CommandEntry, CommandStatus}; use std::collections::BTreeMap; use std::path::PathBuf; use std::time::Duration; use zellij_tile::prelude::*; pub struct Execution { pub all_commands: Vec<CommandEntry>, pub current_running_command_index: usize, pub is_running: bool, pub sequence_id: u64, pub primary_pane_id_before_sequence: Option<PaneId>, } impl Execution { pub fn new() -> Self { Self { all_commands: vec![CommandEntry::default()], current_running_command_index: 0, is_running: false, sequence_id: 0, primary_pane_id_before_sequence: None, } } pub fn longest_cwd_display(&self, global_cwd: &Option<PathBuf>) -> String { self.all_commands .iter() .map(|cmd| { let cwd = cmd.get_cwd().or_else(|| global_cwd.clone()); if let Some(cwd) = &cwd { format_cwd(cwd) } else { "~".to_string() } }) .max_by_key(|s| s.len()) .unwrap_or_else(|| "~".to_string()) } pub fn remove_empty_commands(&mut self) { if self.all_commands.iter().len() > 1 { self.all_commands.retain(|c| !c.get_text().is_empty()); } } pub fn get_first_command(&self) -> Option<CommandEntry> { self.all_commands.iter().next().cloned() } pub fn set_command_status(&mut self, command_index: usize, status: CommandStatus) { self.all_commands .get_mut(command_index) .map(|c| c.set_status(status)); } pub fn set_current_running_command_status(&mut self, status: CommandStatus) { self.all_commands .get_mut(self.current_running_command_index) .map(|c| c.set_status(status)); } pub fn get_current_running_command_status(&self) -> Option<CommandStatus> { self.all_commands .get(self.current_running_command_index) .map(|c| c.get_status()) } pub fn can_run_sequence(&self) -> bool { self.all_commands.iter().any(|command| !command.is_empty()) } pub fn copy_to_clipboard(&self) { let text_to_copy = self .all_commands .iter() .map(|c| format!("{}", c.get_text())) .collect::<Vec<_>>() .join("\n"); if !text_to_copy.is_empty() { copy_to_clipboard(text_to_copy); } } pub fn update_pane_id_for_command(&mut self, pane_id: PaneId, command_text: &str) { for command in self.all_commands.iter_mut() { if let CommandStatus::Pending | CommandStatus::Running(None) = command.get_status() { let cmd_text = command.get_text(); if cmd_text == command_text { command.set_status(CommandStatus::Running(Some(pane_id))); break; } } } } pub fn update_exited_command_statuses(&mut self, pane_manifest: &PaneManifest) -> bool { let mut updated = false; for command in self.all_commands.iter_mut() { let status = command.get_status(); let pane_id_opt = match &status { CommandStatus::Running(pid) => *pid, CommandStatus::Exited(_, pid) => *pid, CommandStatus::Interrupted(pid) => *pid, _ => None, }; if let Some(pane_id) = pane_id_opt { let mut pane_found = false; for (_tab_index, panes) in &pane_manifest.panes { for pane_info in panes { let pane_matches = match pane_id { PaneId::Terminal(id) => !pane_info.is_plugin && pane_info.id == id, PaneId::Plugin(id) => pane_info.is_plugin && pane_info.id == id, }; if pane_matches { pane_found = true; if pane_info.exited { command.set_status(CommandStatus::Exited( pane_info.exit_status, Some(pane_id), )); updated = true; } break; } } if pane_found { break; } } if !pane_found && command.start_time.elapsed() > Duration::from_millis(400) { match command.get_status() { CommandStatus::Running(_) => { eprintln!( "Pane {:?} was closed while running, setting pane_id to None", pane_id ); command.set_status(CommandStatus::Running(None)); updated = true; }, CommandStatus::Exited(exit_code, _) => { eprintln!( "Pane {:?} was closed after exiting, setting pane_id to None", pane_id ); command.set_status(CommandStatus::Exited(exit_code, None)); updated = true; }, CommandStatus::Interrupted(_) => { eprintln!("Pane {:?} was closed after being interrupted, setting pane_id to None", pane_id); command.set_status(CommandStatus::Interrupted(None)); updated = true; }, _ => {}, } } } } updated } pub fn update_sequence_stopped_state(&mut self) -> bool { let mut needs_rerender = false; if self.is_running { let current_idx = self.current_running_command_index; if let Some(command) = self.all_commands.get_mut(current_idx) { if let CommandStatus::Exited(exit_code, _) = command.get_status() { let should_stop; if current_idx >= self.all_commands.len().saturating_sub(1) { should_stop = true; } else { if let Some(chain_type) = &self .all_commands .get(current_idx) .map(|c| c.get_chain_type()) { match chain_type { ChainType::And => { should_stop = exit_code.unwrap_or(0) != 0; }, ChainType::Or => { should_stop = exit_code.unwrap_or(0) == 0; }, ChainType::Then => { should_stop = false; }, ChainType::None => { should_stop = true; }, } } else { should_stop = true; } }; if should_stop { self.is_running = false; } } } needs_rerender = true; } needs_rerender } pub fn execute_command_sequence( &mut self, shell: &Option<PathBuf>, global_cwd: &Option<PathBuf>, primary_pane_id: Option<PaneId>, ) { use zellij_tile::prelude::actions::{Action, RunCommandAction}; self.all_commands.retain(|c| !c.is_empty()); let Some(first_active_sequence_command) = self.get_first_command() else { return; }; let shell = shell.clone().unwrap_or_else(|| PathBuf::from("/bin/bash")); let first_command = first_active_sequence_command.get_text(); let first_chain_type = first_active_sequence_command.get_chain_type(); let command_cwd = first_active_sequence_command .get_cwd() .or_else(|| global_cwd.clone()); let command = RunCommandAction { command: shell.clone(), args: vec!["-ic".to_string(), first_command.trim().to_string()], cwd: command_cwd, hold_on_close: true, ..Default::default() }; let placement = NewPanePlacement::InPlace { pane_id_to_replace: primary_pane_id, close_replaced_pane: false, }; let action = Action::NewBlockingPane { placement, command: Some(command), pane_name: Some(first_command.trim().to_string()), unblock_condition: first_chain_type.to_unblock_condition(), near_current_pane: true, }; self.sequence_id += 1; let mut context = BTreeMap::new(); context.insert("sequence_id".to_string(), self.sequence_id.to_string()); run_action(action, context); self.set_command_status(0, CommandStatus::Running(None)); } } impl Default for Execution { fn default() -> Self { Self::new() } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/sequence/src/state/positioning.rs
default-plugins/sequence/src/state/positioning.rs
use zellij_tile::prelude::*; pub fn reposition_plugin_for_sequence( plugin_id: u32, total_cols: usize, total_rows: usize, total_commands: usize, ) { // Width: 30% of viewport let width = std::cmp::max((total_cols * 30) / 100, 50); // Height: UI overhead + commands, capped at 25% of viewport let overhead_rows = 6; let height = (overhead_rows + total_commands).min((total_rows * 25) / 100); // Position: top-right with 1-space margins let x = total_cols.saturating_sub(width); let y = 1; change_floating_pane_coordinates_absolute( plugin_id, Some(x), Some(y), Some(width), Some(height), true, ); } pub fn change_floating_pane_coordinates_absolute( own_plugin_id: u32, x: Option<usize>, y: Option<usize>, width: Option<usize>, height: Option<usize>, should_be_pinned: bool, ) { let coordinates = FloatingPaneCoordinates::new( x.map(|x| x.to_string()), y.map(|y| y.to_string()), width.map(|width| width.to_string()), height.map(|height| height.to_string()), Some(should_be_pinned), ); if let Some(coordinates) = coordinates { change_floating_panes_coordinates(vec![(PaneId::Plugin(own_plugin_id), coordinates)]); } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/sequence/src/ui/truncation.rs
default-plugins/sequence/src/ui/truncation.rs
use unicode_width::UnicodeWidthStr; const BASE_SEPARATOR_WIDTH: usize = 2; pub fn calculate_available_cmd_width( cols: usize, folder_width: usize, overflow_indicator: Option<&String>, chain_width: usize, status_width: usize, ) -> usize { let mut separator_width = BASE_SEPARATOR_WIDTH; if overflow_indicator.is_some() { separator_width += 1; } if status_width > 0 { separator_width += 1; } let overflow_width = overflow_indicator.map(|s| s.chars().count()).unwrap_or(0); cols.saturating_sub(folder_width) .saturating_sub(chain_width) .saturating_sub(status_width) .saturating_sub(separator_width) .saturating_sub(overflow_width) .saturating_sub(2) .max(1) } pub fn truncate_middle( text: &str, max_width: usize, cursor_position: Option<usize>, ) -> (String, Option<usize>) { let text_width = text.width(); if text_width <= max_width { return (text.to_string(), cursor_position); } if max_width < 5 { return truncate_minimal(text, max_width, cursor_position); } if let Some(cursor_char_idx) = cursor_position { return truncate_with_cursor(text, max_width, cursor_char_idx); } truncate_no_cursor(text, max_width) } fn truncate_minimal( text: &str, max_width: usize, cursor_position: Option<usize>, ) -> (String, Option<usize>) { let mut result = String::new(); let mut current_width = 0; let mut new_cursor_pos = None; let mut char_pos = 0; for ch in text.chars() { let ch_width = ch.to_string().width(); if current_width + ch_width <= max_width { result.push(ch); if let Some(cursor_char) = cursor_position { if char_pos == cursor_char { new_cursor_pos = Some(char_pos); } } char_pos += 1; current_width += ch_width; } else { break; } } (result, new_cursor_pos) } fn truncate_no_cursor(text: &str, max_width: usize) -> (String, Option<usize>) { let available_for_text = max_width.saturating_sub(3); let left_width = available_for_text / 2; let right_width = available_for_text - left_width; let mut left_part = String::new(); let mut current_width = 0; for ch in text.chars() { let ch_width = ch.to_string().width(); if current_width + ch_width <= left_width { left_part.push(ch); current_width += ch_width; } else { break; } } let chars: Vec<char> = text.chars().collect(); let mut right_part = String::new(); let mut current_width = 0; for ch in chars.iter().rev() { let ch_width = ch.to_string().width(); if current_width + ch_width <= right_width { right_part.insert(0, *ch); current_width += ch_width; } else { break; } } (format!("{}...{}", left_part, right_part), None) } fn truncate_with_cursor( text: &str, max_width: usize, cursor_char_idx: usize, ) -> (String, Option<usize>) { let chars: Vec<char> = text.chars().collect(); let char_widths: Vec<usize> = chars.iter().map(|ch| ch.to_string().width()).collect(); let width_before_cursor: usize = char_widths[..cursor_char_idx].iter().sum(); let width_after_cursor: usize = char_widths[cursor_char_idx..].iter().sum(); let available_one_ellipsis = max_width.saturating_sub(3); let (start_idx, end_idx) = if width_before_cursor <= available_one_ellipsis { calculate_end_truncation(&chars, &char_widths, available_one_ellipsis) } else if width_after_cursor <= available_one_ellipsis { calculate_start_truncation(&chars, &char_widths, available_one_ellipsis) } else { calculate_middle_truncation(&char_widths, max_width, cursor_char_idx) }; let (start_idx, end_idx) = adjust_small_truncations(start_idx, end_idx, &char_widths, cursor_char_idx, &chars); let (start_idx, end_idx) = trim_excess(&char_widths, start_idx, end_idx, max_width, cursor_char_idx); build_truncated_result(&chars, start_idx, end_idx, cursor_char_idx) } fn calculate_end_truncation( chars: &[char], char_widths: &[usize], available: usize, ) -> (usize, usize) { let mut end_idx = 0; let mut width = 0; while end_idx < chars.len() && width + char_widths[end_idx] <= available { width += char_widths[end_idx]; end_idx += 1; } (0, end_idx) } fn calculate_start_truncation( chars: &[char], char_widths: &[usize], available: usize, ) -> (usize, usize) { let mut start_idx = chars.len(); let mut width = 0; while start_idx > 0 && width + char_widths[start_idx - 1] <= available { start_idx -= 1; width += char_widths[start_idx]; } (start_idx, chars.len()) } fn calculate_middle_truncation( char_widths: &[usize], max_width: usize, cursor_char_idx: usize, ) -> (usize, usize) { let available_both_ellipsis = max_width.saturating_sub(6); let target_before = available_both_ellipsis / 2; let target_after = available_both_ellipsis - target_before; let mut start_idx = cursor_char_idx; let mut width_before = 0; while start_idx > 0 && width_before + char_widths[start_idx - 1] <= target_before { start_idx -= 1; width_before += char_widths[start_idx]; } let mut end_idx = cursor_char_idx; let mut width_after = 0; while end_idx < char_widths.len() && width_after + char_widths[end_idx] <= target_after { width_after += char_widths[end_idx]; end_idx += 1; } let leftover_before = target_before.saturating_sub(width_before); let leftover_after = target_after.saturating_sub(width_after); if leftover_before > 0 { let mut extra = leftover_before; while end_idx < char_widths.len() && char_widths[end_idx] <= extra { extra -= char_widths[end_idx]; end_idx += 1; } } if leftover_after > 0 { let mut extra = leftover_after; while start_idx > 0 && char_widths[start_idx - 1] <= extra { start_idx -= 1; extra -= char_widths[start_idx]; } } (start_idx, end_idx) } fn adjust_small_truncations( start_idx: usize, end_idx: usize, char_widths: &[usize], cursor_char_idx: usize, chars: &[char], ) -> (usize, usize) { let width_truncated_start: usize = char_widths[..start_idx].iter().sum(); let width_truncated_end: usize = char_widths[end_idx..].iter().sum(); let mut start_idx = start_idx; let mut end_idx = end_idx; if width_truncated_start > 0 && width_truncated_start < 3 && end_idx < chars.len() { let gained = width_truncated_start; start_idx = 0; let mut removed = 0; while end_idx > cursor_char_idx + 1 && removed < gained { end_idx -= 1; removed += char_widths[end_idx]; } } else if width_truncated_end > 0 && width_truncated_end < 3 && start_idx > 0 { let gained = width_truncated_end; end_idx = chars.len(); let mut removed = 0; while start_idx < cursor_char_idx && removed < gained { removed += char_widths[start_idx]; start_idx += 1; } } (start_idx, end_idx) } fn trim_excess( char_widths: &[usize], start_idx: usize, end_idx: usize, max_width: usize, cursor_char_idx: usize, ) -> (usize, usize) { let truncate_start = start_idx > 0; let truncate_end = end_idx < char_widths.len(); let ellipsis_width = match (truncate_start, truncate_end) { (true, true) => 6, (true, false) | (false, true) => 3, (false, false) => 0, }; let visible_width: usize = char_widths[start_idx..end_idx].iter().sum(); let mut start_idx = start_idx; let mut end_idx = end_idx; if visible_width + ellipsis_width > max_width { let mut excess = visible_width + ellipsis_width - max_width; if truncate_end { while excess > 0 && end_idx > cursor_char_idx + 1 { end_idx -= 1; excess = excess.saturating_sub(char_widths[end_idx]); } } if excess > 0 && truncate_start { while excess > 0 && start_idx < cursor_char_idx { excess = excess.saturating_sub(char_widths[start_idx]); start_idx += 1; } } } (start_idx, end_idx) } fn build_truncated_result( chars: &[char], start_idx: usize, end_idx: usize, cursor_char_idx: usize, ) -> (String, Option<usize>) { let mut result = String::new(); let mut new_cursor_char_pos = 0; let mut current_char_pos = 0; if start_idx > 0 { result.push_str("..."); current_char_pos = 3; } for i in start_idx..end_idx { if i == cursor_char_idx { new_cursor_char_pos = current_char_pos; } result.push(chars[i]); current_char_pos += 1; } if cursor_char_idx >= end_idx { new_cursor_char_pos = current_char_pos; } if end_idx < chars.len() { result.push_str("..."); } (result, Some(new_cursor_char_pos)) } #[cfg(test)] mod tests { use super::*; #[test] fn test_truncate_middle_no_truncation_needed() { let text = "hello world"; let (result, cursor_pos) = truncate_middle(text, 20, Some(6)); assert_eq!(result, "hello world"); assert_eq!(cursor_pos, Some(6)); } #[test] fn test_truncate_middle_no_cursor() { let text = "this is a very long string that needs truncation"; let (result, cursor_pos) = truncate_middle(text, 20, None); assert!(result.contains("...")); assert_eq!(cursor_pos, None); } #[test] fn test_truncate_middle_cursor_at_start() { let text = "this is a very long string that needs truncation"; let (result, cursor_pos) = truncate_middle(text, 20, Some(0)); assert!(result.starts_with("this")); assert!(result.ends_with("...")); assert_eq!(cursor_pos, Some(0)); } #[test] fn test_truncate_middle_cursor_at_end() { let text = "this is a very long string that needs truncation"; let cursor_char = text.chars().count(); let (result, _cursor_pos) = truncate_middle(text, 20, Some(cursor_char)); assert!(result.starts_with("...")); assert!(result.ends_with("truncation")); } #[test] fn test_truncate_middle_cursor_in_middle() { let text = "this is a very long string that needs truncation"; let cursor_char = "this is a very ".chars().count(); let (result, cursor_pos) = truncate_middle(text, 20, Some(cursor_char)); assert!(result.contains("very")); assert!(cursor_pos.is_some()); } #[test] fn test_truncate_middle_wide_chars() { let text = "こんにちは世界"; let (result, _) = truncate_middle(text, 10, None); assert!(result.width() <= 10); } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/sequence/src/ui/text_input.rs
default-plugins/sequence/src/ui/text_input.rs
use zellij_tile::prelude::*; const MAX_UNDO_STACK_SIZE: usize = 100; /// Action returned by TextInput after handling a key event #[derive(Debug, Clone, PartialEq)] #[cfg_attr(test, derive(Eq))] pub enum InputAction { /// Continue editing Continue, /// User pressed Enter to submit Submit, /// User pressed Esc to cancel Cancel, /// User pressed Tab to request completion Complete, /// Key was not handled by the input NoAction, } /// A reusable text input component with cursor support and standard editing keybindings #[derive(Debug, Clone)] pub struct TextInput { buffer: String, cursor_position: usize, // Character position (0-based), NOT byte position undo_stack: Vec<(String, usize)>, // (buffer, cursor) snapshots redo_stack: Vec<(String, usize)>, last_edit_was_insert: bool, // For coalescing consecutive inserts } impl TextInput { /// Create a new TextInput with the given initial text /// Cursor is positioned at the end of the text pub fn new(initial_text: String) -> Self { let cursor_position = initial_text.chars().count(); Self { buffer: initial_text, cursor_position, undo_stack: Vec::new(), redo_stack: Vec::new(), last_edit_was_insert: false, } } /// Create an empty TextInput pub fn empty() -> Self { Self::new(String::new()) } /// Get the current text pub fn get_text(&self) -> &str { &self.buffer } /// Get the cursor position (in characters, not bytes) pub fn get_cursor_position(&self) -> usize { self.cursor_position } /// Check if the input is empty pub fn is_empty(&self) -> bool { self.buffer.is_empty() } /// Get a shorthand for cursor_position pub fn cursor_position(&self) -> usize { self.cursor_position } /// Get mutable access to the underlying buffer for direct manipulation pub fn get_text_mut(&mut self) -> &mut String { &mut self.buffer } /// Set the text and move cursor to the end pub fn set_text(&mut self, text: String) { self.break_coalescing(); self.save_undo_state(); self.cursor_position = text.chars().count(); self.buffer = text; } /// Set cursor position (clamped to text length) pub fn set_cursor_position(&mut self, pos: usize) { let text_len = self.buffer.chars().count(); self.cursor_position = pos.min(text_len); } /// Clear all text and reset cursor pub fn clear(&mut self) { self.break_coalescing(); self.save_undo_state(); self.buffer.clear(); self.cursor_position = 0; } /// Insert a character at the current cursor position pub fn insert_char(&mut self, c: char) { self.save_undo_state_unless_coalescing(); // Convert cursor position (char index) to byte index let byte_index = self.char_index_to_byte_index(self.cursor_position); self.buffer.insert(byte_index, c); self.cursor_position += 1; } /// Delete the character before the cursor (backspace) pub fn backspace(&mut self) { if self.cursor_position > 0 { self.break_coalescing(); self.save_undo_state(); self.cursor_position -= 1; let byte_index = self.char_index_to_byte_index(self.cursor_position); self.buffer.remove(byte_index); } } /// Delete the character at the cursor position (delete key) pub fn delete(&mut self) { let len = self.buffer.chars().count(); if self.cursor_position < len { self.break_coalescing(); self.save_undo_state(); let byte_index = self.char_index_to_byte_index(self.cursor_position); self.buffer.remove(byte_index); } } /// Delete the word before the cursor (Ctrl/Alt + Backspace) pub fn delete_word_backward(&mut self) { if self.cursor_position == 0 { return; } self.break_coalescing(); self.save_undo_state(); let old_position = self.cursor_position; self.move_word_left(); let new_position = self.cursor_position; // Delete from new position to old position let start_byte = self.char_index_to_byte_index(new_position); let end_byte = self.char_index_to_byte_index(old_position); self.buffer.drain(start_byte..end_byte); } /// Delete the word after the cursor (Ctrl/Alt + Delete) pub fn delete_word_forward(&mut self) { let chars: Vec<char> = self.buffer.chars().collect(); let len = chars.len(); if self.cursor_position >= len { return; } self.break_coalescing(); self.save_undo_state(); let start_position = self.cursor_position; let mut end_position = start_position; // Skip the current word while end_position < len && !chars[end_position].is_whitespace() { end_position += 1; } // Skip any whitespace after the word while end_position < len && chars[end_position].is_whitespace() { end_position += 1; } // Delete from start to end position let start_byte = self.char_index_to_byte_index(start_position); let end_byte = self.char_index_to_byte_index(end_position); self.buffer.drain(start_byte..end_byte); } /// Move cursor one position to the left pub fn move_left(&mut self) { if self.cursor_position > 0 { self.break_coalescing(); self.cursor_position -= 1; } } /// Move cursor one position to the right pub fn move_right(&mut self) { let len = self.buffer.chars().count(); if self.cursor_position < len { self.break_coalescing(); self.cursor_position += 1; } } /// Move cursor to the start of the text (Ctrl-A / Home) pub fn move_to_start(&mut self) { self.break_coalescing(); self.cursor_position = 0; } /// Move cursor to the end of the text (Ctrl-E / End) pub fn move_to_end(&mut self) { self.break_coalescing(); self.cursor_position = self.buffer.chars().count(); } /// Move cursor to the start of the previous word (Ctrl/Alt + Left) pub fn move_word_left(&mut self) { if self.cursor_position == 0 { return; } self.break_coalescing(); let chars: Vec<char> = self.buffer.chars().collect(); let mut pos = self.cursor_position; // Skip any whitespace immediately to the left while pos > 0 && chars[pos - 1].is_whitespace() { pos -= 1; } // Skip the word characters while pos > 0 && !chars[pos - 1].is_whitespace() { pos -= 1; } self.cursor_position = pos; } /// Move cursor to the start of the next word (Ctrl/Alt + Right) pub fn move_word_right(&mut self) { let chars: Vec<char> = self.buffer.chars().collect(); let len = chars.len(); if self.cursor_position >= len { return; } self.break_coalescing(); let mut pos = self.cursor_position; // Skip the current word while pos < len && !chars[pos].is_whitespace() { pos += 1; } // Skip any whitespace while pos < len && chars[pos].is_whitespace() { pos += 1; } self.cursor_position = pos; } /// Handle a key event and return the appropriate action /// This is the main entry point for key handling pub fn handle_key(&mut self, key: KeyWithModifier) -> InputAction { // Check for Ctrl modifiers if key.has_modifiers(&[KeyModifier::Ctrl]) { match key.bare_key { BareKey::Char('a') => { self.move_to_start(); return InputAction::Continue; }, BareKey::Char('e') => { self.move_to_end(); return InputAction::Continue; }, BareKey::Char('c') => { // Ctrl-C clears the prompt return InputAction::Cancel; }, BareKey::Char('z') => { // Ctrl-Z: Undo self.undo(); return InputAction::Continue; }, BareKey::Char('y') => { // Ctrl-Y: Redo self.redo(); return InputAction::Continue; }, BareKey::Left => { self.move_word_left(); return InputAction::Continue; }, BareKey::Right => { self.move_word_right(); return InputAction::Continue; }, BareKey::Backspace => { self.delete_word_backward(); return InputAction::Continue; }, BareKey::Delete => { self.delete_word_forward(); return InputAction::Continue; }, _ => {}, } } // Check for Ctrl+Shift modifiers (alternative redo: Ctrl+Shift+Z) if key.has_modifiers(&[KeyModifier::Ctrl, KeyModifier::Shift]) { match key.bare_key { BareKey::Char('Z') => { // Ctrl-Shift-Z: Redo (alternative) self.redo(); return InputAction::Continue; }, _ => {}, } } // Check for Alt modifiers if key.has_modifiers(&[KeyModifier::Alt]) { match key.bare_key { BareKey::Left => { self.move_word_left(); return InputAction::Continue; }, BareKey::Right => { self.move_word_right(); return InputAction::Continue; }, BareKey::Backspace => { self.delete_word_backward(); return InputAction::Continue; }, BareKey::Delete => { self.delete_word_forward(); return InputAction::Continue; }, _ => {}, } } // Handle bare keys (no modifiers) match key.bare_key { BareKey::Enter => InputAction::Submit, BareKey::Esc => InputAction::Cancel, BareKey::Tab => InputAction::Complete, BareKey::Backspace => { self.backspace(); InputAction::Continue }, BareKey::Delete => { self.delete(); InputAction::Continue }, BareKey::Left => { self.move_left(); InputAction::Continue }, BareKey::Right => { self.move_right(); InputAction::Continue }, BareKey::Home => { self.move_to_start(); InputAction::Continue }, BareKey::End => { self.move_to_end(); InputAction::Continue }, BareKey::Char(c) => { self.insert_char(c); InputAction::Continue }, _ => InputAction::NoAction, } } /// Helper: Convert character index to byte index fn char_index_to_byte_index(&self, char_index: usize) -> usize { self.buffer .char_indices() .nth(char_index) .map(|(byte_idx, _)| byte_idx) .unwrap_or(self.buffer.len()) } /// Save current state to undo stack before making changes fn save_undo_state(&mut self) { if self.undo_stack.len() >= MAX_UNDO_STACK_SIZE { self.undo_stack.remove(0); } self.undo_stack .push((self.buffer.clone(), self.cursor_position)); self.redo_stack.clear(); } /// Save state only if not coalescing with previous insert fn save_undo_state_unless_coalescing(&mut self) { if !self.last_edit_was_insert { self.save_undo_state(); } self.last_edit_was_insert = true; } /// Mark that a non-insert edit occurred (breaks coalescing) fn break_coalescing(&mut self) { self.last_edit_was_insert = false; } /// Undo last change pub fn undo(&mut self) -> bool { if let Some((buffer, cursor)) = self.undo_stack.pop() { self.redo_stack .push((self.buffer.clone(), self.cursor_position)); self.buffer = buffer; self.cursor_position = cursor; self.break_coalescing(); true } else { false } } /// Redo last undone change pub fn redo(&mut self) -> bool { if let Some((buffer, cursor)) = self.redo_stack.pop() { self.undo_stack .push((self.buffer.clone(), self.cursor_position)); self.buffer = buffer; self.cursor_position = cursor; self.break_coalescing(); true } else { false } } /// Check if undo is available pub fn can_undo(&self) -> bool { !self.undo_stack.is_empty() } /// Check if redo is available pub fn can_redo(&self) -> bool { !self.redo_stack.is_empty() } pub fn drain_text(&mut self) -> String { self.cursor_position = 0; self.buffer.drain(..).collect() } } // run with: // cargo test --lib --target x86_64-unknown-linux-gnu // #[cfg(test)] mod tests { use super::*; #[test] fn test_new_and_empty() { let input = TextInput::new("hello".to_string()); assert_eq!(input.get_text(), "hello"); assert_eq!(input.get_cursor_position(), 5); let empty = TextInput::empty(); assert_eq!(empty.get_text(), ""); assert_eq!(empty.get_cursor_position(), 0); } #[test] fn test_insert_char() { let mut input = TextInput::new("helo".to_string()); input.cursor_position = 3; // Position after "hel" input.insert_char('l'); assert_eq!(input.get_text(), "hello"); assert_eq!(input.get_cursor_position(), 4); } #[test] fn test_backspace() { let mut input = TextInput::new("hello".to_string()); input.backspace(); assert_eq!(input.get_text(), "hell"); assert_eq!(input.get_cursor_position(), 4); // Backspace at start does nothing input.cursor_position = 0; input.backspace(); assert_eq!(input.get_text(), "hell"); assert_eq!(input.get_cursor_position(), 0); } #[test] fn test_delete() { let mut input = TextInput::new("hello".to_string()); input.cursor_position = 0; input.delete(); assert_eq!(input.get_text(), "ello"); assert_eq!(input.get_cursor_position(), 0); // Delete at end does nothing input.move_to_end(); input.delete(); assert_eq!(input.get_text(), "ello"); } #[test] fn test_cursor_movement() { let mut input = TextInput::new("hello".to_string()); assert_eq!(input.get_cursor_position(), 5); input.move_left(); assert_eq!(input.get_cursor_position(), 4); input.move_right(); assert_eq!(input.get_cursor_position(), 5); input.move_to_start(); assert_eq!(input.get_cursor_position(), 0); input.move_to_end(); assert_eq!(input.get_cursor_position(), 5); } #[test] fn test_unicode_support() { let mut input = TextInput::new("hello 🦀 world".to_string()); assert_eq!(input.get_cursor_position(), 13); // 13 characters input.cursor_position = 6; // After "hello " input.insert_char('🐱'); assert_eq!(input.get_text(), "hello 🐱🦀 world"); } #[test] fn test_word_jump_right() { let mut input = TextInput::new("hello world foo bar".to_string()); input.cursor_position = 0; // Jump from start to "world" input.move_word_right(); assert_eq!(input.get_cursor_position(), 6); // After "hello " // Jump to "foo" input.move_word_right(); assert_eq!(input.get_cursor_position(), 12); // After "world " // Jump to "bar" input.move_word_right(); assert_eq!(input.get_cursor_position(), 16); // After "foo " // Jump to end input.move_word_right(); assert_eq!(input.get_cursor_position(), 19); // At end } #[test] fn test_word_jump_left() { let mut input = TextInput::new("hello world foo bar".to_string()); input.move_to_end(); assert_eq!(input.get_cursor_position(), 19); // Jump back to "bar" input.move_word_left(); assert_eq!(input.get_cursor_position(), 16); // Start of "bar" // Jump back to "foo" input.move_word_left(); assert_eq!(input.get_cursor_position(), 12); // Start of "foo" // Jump back to "world" input.move_word_left(); assert_eq!(input.get_cursor_position(), 6); // Start of "world" // Jump back to "hello" input.move_word_left(); assert_eq!(input.get_cursor_position(), 0); // Start of "hello" } #[test] fn test_word_jump_with_multiple_spaces() { let mut input = TextInput::new("hello world".to_string()); input.cursor_position = 0; // Jump over multiple spaces input.move_word_right(); assert_eq!(input.get_cursor_position(), 8); // After "hello ", at start of "world" // Jump back should skip spaces input.move_word_left(); assert_eq!(input.get_cursor_position(), 0); // Back to start of "hello" } #[test] fn test_word_jump_boundaries() { let mut input = TextInput::new("test".to_string()); // At start - word left does nothing input.cursor_position = 0; input.move_word_left(); assert_eq!(input.get_cursor_position(), 0); // At end - word right does nothing input.move_to_end(); let end_pos = input.get_cursor_position(); input.move_word_right(); assert_eq!(input.get_cursor_position(), end_pos); } #[test] fn test_up_down_arrows() { let mut input = TextInput::new("hello world".to_string()); // Start in the middle input.cursor_position = 5; assert_eq!(input.get_cursor_position(), 5); // Up arrow should go to start input.move_to_start(); assert_eq!(input.get_cursor_position(), 0); // Move back to middle input.cursor_position = 5; // Down arrow should go to end input.move_to_end(); assert_eq!(input.get_cursor_position(), 11); } #[test] fn test_delete_word_backward() { let mut input = TextInput::new("hello world foo".to_string()); // Delete "foo" from end input.move_to_end(); input.delete_word_backward(); assert_eq!(input.get_text(), "hello world "); assert_eq!(input.get_cursor_position(), 12); // Delete "world " input.delete_word_backward(); assert_eq!(input.get_text(), "hello "); assert_eq!(input.get_cursor_position(), 6); // Delete "hello " input.delete_word_backward(); assert_eq!(input.get_text(), ""); assert_eq!(input.get_cursor_position(), 0); // Delete on empty buffer does nothing input.delete_word_backward(); assert_eq!(input.get_text(), ""); assert_eq!(input.get_cursor_position(), 0); } #[test] fn test_delete_word_backward_middle() { let mut input = TextInput::new("hello world foo".to_string()); // Position in middle of "world" input.cursor_position = 8; // After "hello wo" input.delete_word_backward(); assert_eq!(input.get_text(), "hello rld foo"); assert_eq!(input.get_cursor_position(), 6); // After "hello " } #[test] fn test_delete_word_forward() { let mut input = TextInput::new("hello world foo".to_string()); // Delete "hello " from start input.cursor_position = 0; input.delete_word_forward(); assert_eq!(input.get_text(), "world foo"); assert_eq!(input.get_cursor_position(), 0); // Delete "world " input.delete_word_forward(); assert_eq!(input.get_text(), "foo"); assert_eq!(input.get_cursor_position(), 0); // Delete "foo" input.delete_word_forward(); assert_eq!(input.get_text(), ""); assert_eq!(input.get_cursor_position(), 0); // Delete on empty buffer does nothing input.delete_word_forward(); assert_eq!(input.get_text(), ""); assert_eq!(input.get_cursor_position(), 0); } #[test] fn test_delete_word_forward_middle() { let mut input = TextInput::new("hello world foo".to_string()); // Position in middle of "world" input.cursor_position = 8; // After "hello wo" input.delete_word_forward(); assert_eq!(input.get_text(), "hello wofoo"); assert_eq!(input.get_cursor_position(), 8); // Same position, text deleted forward } #[test] fn test_delete_word_with_multiple_spaces() { let mut input = TextInput::new("hello world".to_string()); // Delete forward includes trailing spaces input.cursor_position = 0; input.delete_word_forward(); assert_eq!(input.get_text(), "world"); assert_eq!(input.get_cursor_position(), 0); } #[test] fn test_undo_redo_basic() { let mut input = TextInput::empty(); // Type "hello" input.insert_char('h'); input.insert_char('e'); input.insert_char('l'); input.insert_char('l'); input.insert_char('o'); assert_eq!(input.get_text(), "hello"); // Undo should remove all characters (coalesced into one undo entry) assert!(input.can_undo()); assert!(input.undo()); assert_eq!(input.get_text(), ""); assert_eq!(input.get_cursor_position(), 0); // Redo should restore "hello" assert!(input.can_redo()); assert!(input.redo()); assert_eq!(input.get_text(), "hello"); assert_eq!(input.get_cursor_position(), 5); } #[test] fn test_undo_coalescing_breaks_on_cursor_move() { let mut input = TextInput::empty(); // Type "he" input.insert_char('h'); input.insert_char('e'); // Move cursor to start (breaks coalescing) input.move_to_start(); // Type "llo" input.insert_char('l'); input.insert_char('l'); input.insert_char('o'); assert_eq!(input.get_text(), "llohe"); // First undo removes "llo" (second coalesced group) input.undo(); assert_eq!(input.get_text(), "he"); // Second undo removes "he" (first coalesced group) input.undo(); assert_eq!(input.get_text(), ""); } #[test] fn test_undo_backspace() { let mut input = TextInput::new("hello".to_string()); // Backspace once input.backspace(); assert_eq!(input.get_text(), "hell"); // Undo should restore "hello" input.undo(); assert_eq!(input.get_text(), "hello"); assert_eq!(input.get_cursor_position(), 5); } #[test] fn test_undo_delete() { let mut input = TextInput::new("hello".to_string()); input.cursor_position = 0; // Delete first character input.delete(); assert_eq!(input.get_text(), "ello"); // Undo should restore "hello" input.undo(); assert_eq!(input.get_text(), "hello"); assert_eq!(input.get_cursor_position(), 0); } #[test] fn test_undo_word_delete() { let mut input = TextInput::new("hello world".to_string()); // Delete "world" backward input.delete_word_backward(); assert_eq!(input.get_text(), "hello "); // Undo should restore "hello world" input.undo(); assert_eq!(input.get_text(), "hello world"); assert_eq!(input.get_cursor_position(), 11); } #[test] fn test_undo_clear() { let mut input = TextInput::new("hello world".to_string()); // Clear the buffer input.clear(); assert_eq!(input.get_text(), ""); // Undo should restore the text input.undo(); assert_eq!(input.get_text(), "hello world"); } #[test] fn test_undo_set_text() { let mut input = TextInput::new("hello".to_string()); // Replace with new text input.set_text("goodbye".to_string()); assert_eq!(input.get_text(), "goodbye"); // Undo should restore "hello" input.undo(); assert_eq!(input.get_text(), "hello"); } #[test] fn test_redo_clears_on_new_edit() { let mut input = TextInput::empty(); // Type "hello" input.insert_char('h'); input.insert_char('e'); input.insert_char('l'); input.insert_char('l'); input.insert_char('o'); // Undo input.undo(); assert_eq!(input.get_text(), ""); assert!(input.can_redo()); // Make a new edit (should clear redo stack) input.insert_char('x'); assert!(!input.can_redo()); } #[test] fn test_multiple_undo_redo() { let mut input = TextInput::empty(); // First edit: type "hello" for c in "hello".chars() { input.insert_char(c); } // Break coalescing input.move_left(); // Second edit: type "world" for c in "world".chars() { input.insert_char(c); } assert_eq!(input.get_text(), "hellworldo"); // Undo "world" input.undo(); assert_eq!(input.get_text(), "hello"); // Undo "hello" input.undo(); assert_eq!(input.get_text(), ""); // Redo "hello" input.redo(); assert_eq!(input.get_text(), "hello"); // Redo "world" input.redo(); assert_eq!(input.get_text(), "hellworldo"); } #[test] fn test_undo_stack_limit() { let mut input = TextInput::empty(); // Perform 102 separate edits (breaking coalescing each time) for _i in 0..102 { input.backspace(); // Break coalescing input.insert_char('x'); } // Should have at most 100 undo entries let mut undo_count = 0; while input.undo() { undo_count += 1; } // We should have 100 undo entries (the stack limit) // Plus the final state change from the last coalescing break assert!( undo_count <= 100, "Undo count should be at most 100, got {}", undo_count ); } #[test] fn test_undo_redo_empty_stack() { let mut input = TextInput::empty(); // Undo on empty stack should return false assert!(!input.can_undo()); assert!(!input.undo()); // Redo on empty stack should return false assert!(!input.can_redo()); assert!(!input.redo()); } #[test] fn test_undo_restores_cursor_position() { let mut input = TextInput::new("hello world".to_string()); // Move cursor to position 5 (before "world") input.cursor_position = 5; // Insert a space input.insert_char(' '); assert_eq!(input.get_text(), "hello world"); assert_eq!(input.get_cursor_position(), 6); // Undo should restore both text and cursor position input.undo(); assert_eq!(input.get_text(), "hello world"); assert_eq!(input.get_cursor_position(), 5); } #[test] fn test_coalescing_consecutive_inserts() { let mut input = TextInput::empty(); // Type several characters input.insert_char('a'); input.insert_char('b'); input.insert_char('c'); assert_eq!(input.get_text(), "abc"); // Single undo should remove all three (they were coalesced) input.undo(); assert_eq!(input.get_text(), ""); // No more undo available assert!(!input.can_undo()); } #[test] fn test_backspace_breaks_coalescing() { let mut input = TextInput::empty(); // Type "ab" input.insert_char('a'); input.insert_char('b'); // Backspace input.backspace(); assert_eq!(input.get_text(), "a"); // Type "c" input.insert_char('c'); assert_eq!(input.get_text(), "ac"); // Undo should remove just "c" input.undo(); assert_eq!(input.get_text(), "a"); // Undo should remove backspace operation input.undo(); assert_eq!(input.get_text(), "ab"); // Undo should remove "ab" input.undo(); assert_eq!(input.get_text(), ""); } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/sequence/src/ui/fuzzy_complete.rs
default-plugins/sequence/src/ui/fuzzy_complete.rs
use fuzzy_matcher::skim::SkimMatcherV2; use fuzzy_matcher::FuzzyMatcher; use std::collections::BTreeMap; use std::fs; use std::path::PathBuf; /// Result of a fuzzy completion operation #[derive(Debug, Clone)] pub struct CompletionResult { pub completed_text: String, pub score: i64, pub is_directory: bool, pub is_prefix_completion: bool, // True if this is the shortest of multiple matches with same prefix } /// Type of completion #[derive(Debug, Clone, Copy, PartialEq)] pub enum CompletionType { Command, Path, } /// Fuzzy match executables (commands) from the PATH pub fn fuzzy_complete_command( query: &str, executables: &BTreeMap<String, PathBuf>, ) -> Option<CompletionResult> { if query.is_empty() { return None; } let matcher = SkimMatcherV2::default().ignore_case(); // Collect all matches with their scores let mut all_matches: Vec<(String, i64)> = Vec::new(); for executable_name in executables.keys() { if let Some((score, _indices)) = matcher.fuzzy_indices(executable_name, query) { all_matches.push((executable_name.clone(), score)); } } if all_matches.is_empty() { return None; } // Sort by score (highest first), then by length (shortest first) all_matches.sort_by(|a, b| b.1.cmp(&a.1).then_with(|| a.0.len().cmp(&b.0.len()))); // Get the best match let (best_name, best_score) = &all_matches[0]; // Check if there are multiple matches that share the same prefix let is_prefix_completion = check_for_prefix_matches(&all_matches, best_name); Some(CompletionResult { completed_text: best_name.clone(), score: *best_score, is_directory: false, // Commands are not directories is_prefix_completion, }) } /// Fuzzy match paths (directories and files) relative to the current working directory pub fn fuzzy_complete_path(query: &str, cwd: Option<&PathBuf>) -> Option<CompletionResult> { if query.is_empty() { return None; } let matcher = SkimMatcherV2::default().ignore_case(); // Determine the base directory to search in let (base_dir, search_prefix) = if query.starts_with('/') { // Absolute path let parts: Vec<&str> = query.rsplitn(2, '/').collect(); if parts.len() == 2 { // Has a directory component let dir = parts[1]; let prefix = parts[0]; (PathBuf::from(dir), prefix.to_string()) } else { // Just "/" (PathBuf::from("/"), String::new()) } } else if query.starts_with("~/") { // Home directory let home_dir = std::env::var("HOME").ok()?; let query_without_tilde = &query[2..]; let parts: Vec<&str> = query_without_tilde.rsplitn(2, '/').collect(); if parts.len() == 2 { let dir = PathBuf::from(&home_dir).join(parts[1]); let prefix = parts[0]; (dir, prefix.to_string()) } else { // Just "~/" or "~/something" (PathBuf::from(home_dir), query_without_tilde.to_string()) } } else { // Relative path let base = cwd?; let parts: Vec<&str> = query.rsplitn(2, '/').collect(); if parts.len() == 2 { // Has a directory component let dir = base.join(parts[1]); let prefix = parts[0]; (dir, prefix.to_string()) } else { // Just a name in the current directory (base.clone(), query.to_string()) } }; // Convert to host path let host_base_dir = PathBuf::from("/host").join(base_dir.strip_prefix("/").unwrap_or(&base_dir)); // Read directory entries let entries = fs::read_dir(&host_base_dir).ok()?; // Collect all matches with their scores and directory status let mut all_matches: Vec<(String, i64, bool)> = Vec::new(); for entry in entries.flatten() { if let Some(name) = entry.file_name().to_str() { // Skip hidden files unless the query starts with a dot if name.starts_with('.') && !search_prefix.starts_with('.') { continue; } if let Some((score, _indices)) = matcher.fuzzy_indices(name, &search_prefix) { // Check if this entry is a directory let is_dir = entry.metadata().ok().map(|m| m.is_dir()).unwrap_or(false); all_matches.push((name.to_string(), score, is_dir)); } } } if all_matches.is_empty() { return None; } // Sort by score (highest first), then by length (shortest first) all_matches.sort_by(|a, b| b.1.cmp(&a.1).then_with(|| a.0.len().cmp(&b.0.len()))); // Get the best match let (name, score, is_directory) = all_matches[0].clone(); // Check if there are multiple matches that share the same prefix let name_matches: Vec<(String, i64)> = all_matches .iter() .map(|(n, s, _)| (n.clone(), *s)) .collect(); let is_prefix_completion = check_for_prefix_matches(&name_matches, &name); // Reconstruct the full path let completed_text = if query.starts_with('/') { // Absolute path let dir_part = query.rsplitn(2, '/').nth(1).unwrap_or(""); if dir_part.is_empty() { format!("/{}", name) } else { format!("{}/{}", dir_part, name) } } else if query.starts_with("~/") { // Home directory let query_without_tilde = &query[2..]; let dir_part = query_without_tilde.rsplitn(2, '/').nth(1).unwrap_or(""); if dir_part.is_empty() { format!("~/{}", name) } else { format!("~/{}/{}", dir_part, name) } } else { // Relative path let dir_part = query.rsplitn(2, '/').nth(1).unwrap_or(""); if dir_part.is_empty() { name.clone() } else { format!("{}/{}", dir_part, name) } }; Some(CompletionResult { completed_text, score, is_directory, is_prefix_completion, }) } /// Check if there are multiple matches that start with the best match's name /// Returns true if the best match is a prefix of other matches fn check_for_prefix_matches(matches: &[(String, i64)], best_match: &str) -> bool { if matches.len() <= 1 { return false; } // Count how many matches start with the best_match text let prefix_count = matches .iter() .filter(|(name, _)| name.starts_with(best_match) && name != best_match) .count(); prefix_count > 0 } /// Perform fuzzy completion on the current input /// Returns the best completion (command or path) with appropriate suffix (space or slash) pub fn fuzzy_complete( input: &str, executables: &BTreeMap<String, PathBuf>, cwd: Option<&PathBuf>, ) -> Option<String> { // If input is empty, nothing to complete if input.is_empty() { return None; } // Extract the last word/token to complete // For simplicity, we'll complete the entire input if it's a single word, // or the last space-separated token if there are multiple words let tokens: Vec<&str> = input.split_whitespace().collect(); let to_complete = if tokens.is_empty() { input } else { tokens.last().unwrap() }; // Get both command and path completions let command_result = fuzzy_complete_command(to_complete, executables); let path_result = fuzzy_complete_path(to_complete, cwd); // Choose the better match based on score match (command_result, path_result) { (Some(cmd), Some(path)) => { // Both matched, use the higher score let (result, completion_type, is_dir, is_prefix) = if cmd.score >= path.score { ( cmd.completed_text, CompletionType::Command, false, cmd.is_prefix_completion, ) } else { ( path.completed_text, CompletionType::Path, path.is_directory, path.is_prefix_completion, ) }; // Add appropriate suffix (skip slash if this is a prefix completion) let suffix = get_completion_suffix(completion_type, is_dir, is_prefix); let completed = format!("{}{}", result, suffix); // Replace the last token with the completion Some(replace_last_token(input, &completed)) }, (Some(cmd), None) => { // Command completion - add space (unless it's a prefix completion) let suffix = if cmd.is_prefix_completion { "" } else { " " }; let completed = format!("{}{}", cmd.completed_text, suffix); Some(replace_last_token(input, &completed)) }, (None, Some(path)) => { // Path completion - add slash if directory (unless it's a prefix completion) let suffix = if path.is_prefix_completion { "" } else if path.is_directory { "/" } else { "" }; let completed = format!("{}{}", path.completed_text, suffix); Some(replace_last_token(input, &completed)) }, (None, None) => None, } } /// Get the appropriate suffix for a completion fn get_completion_suffix( completion_type: CompletionType, is_directory: bool, is_prefix_completion: bool, ) -> &'static str { // No suffix for prefix completions if is_prefix_completion { return ""; } match completion_type { CompletionType::Command => " ", CompletionType::Path => { if is_directory { "/" } else { "" } }, } } /// Replace the last whitespace-separated token in the input with the completion fn replace_last_token(input: &str, completion: &str) -> String { let tokens: Vec<&str> = input.split_whitespace().collect(); if tokens.is_empty() { return completion.to_string(); } if tokens.len() == 1 { completion.to_string() } else { // Join all tokens except the last one, then add the completion let mut result = tokens[..tokens.len() - 1].join(" "); result.push(' '); result.push_str(completion); result } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/sequence/src/ui/layout_calculations.rs
default-plugins/sequence/src/ui/layout_calculations.rs
pub fn calculate_viewport( total_commands: usize, max_visible: usize, selected_index: Option<usize>, running_index: usize, ) -> (usize, usize, usize, usize) { if total_commands <= max_visible { return (0, total_commands, 0, 0); } let focus_index = selected_index.unwrap_or(running_index); let half_visible = max_visible / 2; let offset = if focus_index < half_visible { 0 } else if focus_index >= total_commands.saturating_sub(half_visible) { total_commands.saturating_sub(max_visible) } else { focus_index.saturating_sub(half_visible) }; let offset = offset.min(total_commands.saturating_sub(max_visible)); let visible_count = max_visible.min(total_commands.saturating_sub(offset)); let hidden_above = offset; let hidden_below = total_commands.saturating_sub(offset + visible_count); (offset, visible_count, hidden_above, hidden_below) }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/sequence/src/ui/mod.rs
default-plugins/sequence/src/ui/mod.rs
pub mod components; pub mod fuzzy_complete; pub mod layout_calculations; pub mod text_input; pub mod truncation;
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/sequence/src/ui/components.rs
default-plugins/sequence/src/ui/components.rs
use crate::path_formatting::format_cwd; use crate::state::{CommandStatus, State}; use crate::ui::truncation::{calculate_available_cmd_width, truncate_middle}; use std::path::PathBuf; use zellij_tile::prelude::*; const SPINNER_FRAMES: &[&str] = &["⠋", "⠙", "⠹", "⠸", "⠼", "⠴", "⠦", "⠧"]; const HELP_RUNNING_WITH_SELECTION: &str = "<Ctrl c> - interrupt, <Enter> - run from selected"; const HELP_RUNNING_NO_SELECTION: &str = "<Ctrl c> - interrupt, <↓↑> - navigate"; const HELP_STOPPED_WITH_SELECTION: &str = "<Ctrl w> - close all, <Enter> - run from selected, <e> - edit"; const HELP_STOPPED_NO_SELECTION: &str = "<Ctrl w> - close all, <↓↑> - navigate, <Enter> - run from first"; const HELP_ONE_PENDING_COMMAND: &str = "<Enter> - run, <Ctrl Enter> - add command"; const HELP_ALL_COMMANDS_PENDING: &str = "<Enter> - run, <Ctrl Enter> - add command, <↓↑> - navigate"; const HELP_ALL_COMMANDS_PENDING_WITH_SELECTION: &str = "<Enter> - run, <Ctrl Enter> - add command, <↓↑> - navigate, <e> - edit selected"; const HELP_EDITING_FIRST_LINE: &str = "<Enter> - accept, <Ctrl Enter> - add command, <↓↑> - navigate"; fn select_help_text(sequence: &State) -> &'static str { let is_running = sequence .execution .all_commands .iter() .any(|command| matches!(command.get_status(), CommandStatus::Running(_))); let all_pending = sequence .execution .all_commands .iter() .all(|command| matches!(command.get_status(), CommandStatus::Pending)); let has_selection = sequence.selection.current_selected_command_index.is_some(); let is_editing = sequence.editing.editing_input.is_some(); if all_pending && !sequence.execution.is_running { if sequence.execution.all_commands.len() == 1 { HELP_ONE_PENDING_COMMAND } else if has_selection && !is_editing { HELP_ALL_COMMANDS_PENDING_WITH_SELECTION } else if is_editing { HELP_EDITING_FIRST_LINE } else { HELP_ALL_COMMANDS_PENDING } } else if is_running { if has_selection { HELP_RUNNING_WITH_SELECTION } else { HELP_RUNNING_NO_SELECTION } } else { if has_selection && !is_editing { HELP_STOPPED_WITH_SELECTION } else if is_editing { HELP_EDITING_FIRST_LINE } else { HELP_STOPPED_NO_SELECTION } } } fn style_help_text(text: &str) -> Text { Text::new(text) .color_substring(3, "<Ctrl c>") .color_substring(3, "<Ctrl w>") .color_substring(3, "<↓↑>") .color_substring(3, "<Esc>") .color_substring(3, "<Enter>") .color_substring(3, "<Ctrl Enter>") .color_substring(3, "<e>") } pub fn render_help_lines( sequence: &State, max_width: Option<usize>, ) -> (Text, usize, Option<(Text, usize)>) { let is_editing = sequence.editing.editing_input.is_some(); if is_editing { let help_text = select_help_text(sequence); let (truncated_text, help_len) = if let Some(width) = max_width { truncate_help_line(help_text, width) } else { (help_text.to_string(), help_text.chars().count()) }; let first_line = (style_help_text(&truncated_text).unbold_all(), help_len); let editing_help_text = "Navigate with cd, chain with ||, &&, ; "; let editing_help = Text::new(editing_help_text) .color_substring(3, "cd") .color_substring(3, "||") .color_substring(3, "&&") .color_substring(3, ";") .unbold_all(); let editing_len = editing_help_text.len(); return ( first_line.0, first_line.1, Some((editing_help, editing_len)), ); } let help_text = select_help_text(sequence); if let Some(width) = max_width { if let Some((first_line, second_line)) = split_help_line(help_text, width) { let first_len = first_line.chars().count(); let first_styled = style_help_text(&first_line).unbold_all(); let second_line_width = second_line.chars().count(); let (second_styled, second_len) = if second_line_width > width { let (truncated, len) = truncate_help_line(&second_line, width); (style_help_text(&truncated).unbold_all(), len) } else { ( style_help_text(&second_line).unbold_all(), second_line_width, ) }; return (first_styled, first_len, Some((second_styled, second_len))); } let (truncated_text, help_len) = truncate_help_line(help_text, width); ( style_help_text(&truncated_text).unbold_all(), help_len, None, ) } else { let help_len = help_text.chars().count(); (style_help_text(help_text).unbold_all(), help_len, None) } } fn split_help_line(help_text: &str, max_width: usize) -> Option<(String, String)> { let text_width = help_text.chars().count(); if text_width <= max_width { return None; } let keybindings = parse_help_keybindings(help_text); if keybindings.is_empty() { return None; } for split_point in 1..keybindings.len() { let first_part: Vec<String> = keybindings .iter() .take(split_point) .map(|(key, _, action)| format!("{} - {}", key, action)) .collect(); let first_line = first_part.join(", "); let second_part: Vec<String> = keybindings .iter() .skip(split_point) .map(|(key, _, action)| format!("{} - {}", key, action)) .collect(); let second_line = second_part.join(", "); if first_line.chars().count() <= max_width && second_line.chars().count() <= max_width { return Some((first_line, second_line)); } } None } fn folder_cell(cwd: &Option<PathBuf>) -> (Text, String) { let cwd_display = if let Some(cwd) = cwd { format_cwd(cwd) } else { "~".to_string() }; let folder_display = format!("{} >", cwd_display); let text = Text::new(&folder_display).color_range(0, 0..cwd_display.len()); (text, folder_display) } fn command_cell(text: &str) -> Text { let text = if text.is_empty() { " " } else { text }; Text::new(text) } fn chain_cell(chain_type: &crate::state::ChainType) -> Text { let text = chain_type_to_str(chain_type); if !text.is_empty() { Text::new(text).color_range(1, 0..text.len()) } else { Text::new(text) } } fn get_spinner_frame(frame: usize) -> &'static str { SPINNER_FRAMES[frame % SPINNER_FRAMES.len()] } pub fn format_status_text(status: &CommandStatus, spinner_frame: usize) -> String { match status { CommandStatus::Running(_) => { let spinner = get_spinner_frame(spinner_frame); format!("[RUNNING] {}", spinner) }, CommandStatus::Interrupted(_) => "[INTERRUPTED]".to_string(), CommandStatus::Exited(Some(code), _) => format!("[EXIT CODE: {}]", code), CommandStatus::Exited(None, _) => "[EXITED]".to_string(), CommandStatus::Pending => " ".to_string(), } } fn apply_status_color(text: Text, status: &CommandStatus, status_str: &str) -> Text { match status { CommandStatus::Running(_) => text.color_range(3, 0..status_str.len()), CommandStatus::Interrupted(_) => text.error_color_range(0..status_str.len()), CommandStatus::Exited(Some(0), _) => { let number_start = 12; let number_end = status_str.len() - 1; text.success_color_range(number_start..number_end) }, CommandStatus::Exited(Some(_), _) => { let number_start = 12; let number_end = status_str.len() - 1; text.error_color_range(number_start..number_end) }, _ => text, } } fn status_cell(status: &CommandStatus, spinner_frame: usize) -> Text { let status_str = format_status_text(status, spinner_frame); let text = Text::new(&status_str); apply_status_color(text, status, &status_str) } fn apply_row_styles(cells: Vec<Text>, is_selected: bool) -> Vec<Text> { let mut styled_cells = cells; if is_selected { styled_cells = styled_cells.into_iter().map(|c| c.selected()).collect(); } styled_cells } fn overflow_cell(indicator: &str, is_selected: bool) -> Text { let mut text = Text::new(indicator).dim_all(); if is_selected { text = text.selected(); } text } fn command_sequence_row( state: &State, cwd: &Option<PathBuf>, truncated_cmd_text: &str, chain_type: &crate::state::ChainType, status: &crate::state::CommandStatus, is_selected: bool, overflow_indicator: Option<String>, ) -> (Vec<Text>, usize) { let (folder_text, folder_display) = folder_cell(cwd); let mut cells = vec![ folder_text, command_cell(truncated_cmd_text), chain_cell(chain_type), status_cell(status, state.layout.spinner_frame), ]; cells = apply_row_styles(cells, is_selected); let truncated_cmd_text = if truncated_cmd_text.is_empty() { " " } else { truncated_cmd_text }; let mut row_length = folder_display.chars().count() + truncated_cmd_text.chars().count() + chain_type_to_str(chain_type).chars().count() + format_status_text(status, state.layout.spinner_frame) .chars() .count() + 3; if let Some(indicator) = overflow_indicator { let indicator_len = indicator.chars().count(); cells.push(overflow_cell(&indicator, is_selected)); row_length += indicator_len + 1; } (cells, row_length) } pub fn build_table_header(has_overflow: bool) -> Table { let header_cols = if has_overflow { vec![" ", " ", " ", " ", " "] } else { vec![" ", " ", " ", " "] }; Table::new().add_row(header_cols) } fn calculate_overflow_indicator( visible_idx: usize, visible_count: usize, hidden_above: usize, hidden_below: usize, ) -> Option<String> { if visible_idx == 0 && hidden_above > 0 { Some(format!("[+{}]", hidden_above)) } else if visible_idx == visible_count.saturating_sub(1) && hidden_below > 0 { Some(format!("[+{}]", hidden_below)) } else if hidden_above > 0 || hidden_below > 0 { Some(format!(" ")) } else { None } } pub fn calculate_row_layout_info( index: usize, offset: usize, visible_count: usize, hidden_above: usize, hidden_below: usize, cols: usize, folder_width: usize, max_chain_width: usize, max_status_width: usize, ) -> Option<(Option<String>, usize)> { if index < offset || index >= offset + visible_count { return None; } let visible_idx = index.saturating_sub(offset); let overflow_indicator = calculate_overflow_indicator(visible_idx, visible_count, hidden_above, hidden_below); let available_cmd_width = calculate_available_cmd_width( cols, folder_width, overflow_indicator.as_ref(), max_chain_width, max_status_width, ); Some((overflow_indicator, available_cmd_width)) } pub fn add_command_row( table: Table, state: &State, index: usize, offset: usize, visible_count: usize, hidden_above: usize, hidden_below: usize, ) -> Table { let command = match state.execution.all_commands.get(index) { Some(cmd) => cmd, None => return table, }; let cmd_text = &command.get_text(); let chain_type = &command.get_chain_type(); let status = &command.get_status(); let command_cwd = command.get_cwd().or_else(|| state.cwd.clone()); let cols = state.own_columns.unwrap_or(80); let longest_cwd_display = state.execution.longest_cwd_display(&state.cwd); let folder_display = format!("{} >", longest_cwd_display); let folder_width = folder_display.chars().count(); let (max_chain_width, max_status_width) = calculate_max_widths(&state.execution.all_commands, state.layout.spinner_frame); let Some((overflow_indicator, available_cmd_width)) = calculate_row_layout_info( index, offset, visible_count, hidden_above, hidden_below, cols, folder_width, max_chain_width, max_status_width, ) else { return table; }; if let Some(text_input) = &state.editing.editing_input { if state.selection.current_selected_command_index == Some(index) { let (truncated_editing_text, _) = truncate_middle( text_input.get_text(), available_cmd_width, Some(text_input.cursor_position()), ); let (row, _) = command_sequence_row( state, &command_cwd, &truncated_editing_text, chain_type, &CommandStatus::Pending, false, overflow_indicator, ); return table.add_styled_row(row); } } let truncated_cmd_text = truncate_middle(cmd_text, available_cmd_width, None).0; let (row, _) = command_sequence_row( state, &command_cwd, &truncated_cmd_text, chain_type, status, state.selection.current_selected_command_index == Some(index), overflow_indicator, ); table.add_styled_row(row) } pub fn calculate_max_widths( commands: &[crate::state::CommandEntry], spinner_frame: usize, ) -> (usize, usize) { let max_chain_width = commands .iter() .map(|cmd| chain_type_to_str(&cmd.get_chain_type()).chars().count()) .max() .unwrap_or(0); let max_status_width = commands .iter() .map(|cmd| { format_status_text(&cmd.get_status(), spinner_frame) .chars() .count() }) .max() .unwrap_or(0); (max_chain_width, max_status_width) } pub fn calculate_longest_line( longest_cwd_display: &str, longest_command: usize, max_chain_width: usize, max_status_width: usize, ) -> usize { let folder_display = format!("{} >", longest_cwd_display); let cell_padding = 3; folder_display.chars().count() + longest_command + max_chain_width + max_status_width + cell_padding } pub fn chain_type_to_str(chain_type: &crate::state::ChainType) -> &'static str { match chain_type { crate::state::ChainType::And => "AND", crate::state::ChainType::Or => "OR", crate::state::ChainType::Then => "THEN", crate::state::ChainType::None => " ", } } pub fn truncate_help_line(help_text: &str, max_width: usize) -> (String, usize) { let text_width = help_text.chars().count(); if text_width <= max_width { return (help_text.to_string(), text_width); } let keybindings = parse_help_keybindings(help_text); if let Some(result) = try_shortened_help(&keybindings, max_width) { return result; } if let Some(result) = try_keys_spaced(&keybindings, max_width) { return result; } if let Some(result) = try_keys_tight(&keybindings, max_width) { return result; } truncate_with_ellipsis( &keybindings .iter() .map(|(k, _, _)| *k) .collect::<Vec<_>>() .join("/"), max_width, ) } fn parse_help_keybindings(help_text: &str) -> Vec<(&str, &str, &str)> { help_text .split(", ") .filter_map(|part| { part.find(" - ").map(|dash_pos| { let key = &part[..dash_pos]; let action = &part[dash_pos + 3..]; let first_word = action.split_whitespace().next().unwrap_or(action); (key, first_word, action) }) }) .collect() } fn try_shortened_help( keybindings: &[(&str, &str, &str)], max_width: usize, ) -> Option<(String, usize)> { let shortened: Vec<String> = keybindings .iter() .map(|(key, first_word, _)| format!("{} - {}", key, first_word)) .collect(); let shortened_text = shortened.join(", "); let shortened_width = shortened_text.chars().count(); if shortened_width <= max_width { Some((shortened_text, shortened_width)) } else { None } } fn try_keys_spaced( keybindings: &[(&str, &str, &str)], max_width: usize, ) -> Option<(String, usize)> { let keys_spaced_text = keybindings .iter() .map(|(key, _, _)| *key) .collect::<Vec<_>>() .join(" / "); let keys_spaced_width = keys_spaced_text.chars().count(); if keys_spaced_width <= max_width { Some((keys_spaced_text, keys_spaced_width)) } else { None } } fn try_keys_tight(keybindings: &[(&str, &str, &str)], max_width: usize) -> Option<(String, usize)> { let keys_tight_text = keybindings .iter() .map(|(key, _, _)| *key) .collect::<Vec<_>>() .join("/"); let keys_tight_width = keys_tight_text.chars().count(); if keys_tight_width <= max_width { Some((keys_tight_text, keys_tight_width)) } else { None } } fn truncate_with_ellipsis(text: &str, max_width: usize) -> (String, usize) { let available = max_width.saturating_sub(3); let mut truncated = String::new(); let mut current_width = 0; for ch in text.chars() { if current_width + 1 <= available { truncated.push(ch); current_width += 1; } else { break; } } truncated.push_str("..."); let truncated_width = truncated.chars().count(); (truncated, truncated_width) }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/strider/src/search_view.rs
default-plugins/strider/src/search_view.rs
use crate::shared::{calculate_list_bounds, render_list_tip}; use fuzzy_matcher::skim::SkimMatcherV2; use fuzzy_matcher::FuzzyMatcher; use pretty_bytes::converter::convert as pretty_bytes; use unicode_width::UnicodeWidthStr; use zellij_tile::prelude::*; use crate::file_list_view::FsEntry; #[derive(Default, Debug)] pub struct SearchView { pub search_results: Vec<SearchResult>, pub selected_search_result: usize, } impl SearchView { pub fn search_result_count(&self) -> usize { self.search_results.len() } pub fn update_search_results(&mut self, search_term: &str, files: &Vec<FsEntry>) { self.selected_search_result = 0; if search_term.is_empty() { self.search_results.clear(); } else { let mut matches = vec![]; let matcher = SkimMatcherV2::default().use_cache(true); for file in files { let name = file.name(); if let Some((score, indices)) = matcher.fuzzy_indices(&name, search_term) { matches.push(SearchResult::new(file.clone(), score, indices)); } } matches.sort_by(|a, b| b.score.cmp(&a.score)); self.search_results = matches; } } pub fn clear_and_reset_selection(&mut self) { self.search_results.clear(); self.selected_search_result = 0; } pub fn move_selection_up(&mut self) { self.selected_search_result = self.selected_search_result.saturating_sub(1); } pub fn move_selection_down(&mut self) { if self.selected_search_result + 1 < self.search_results.len() { self.selected_search_result += 1; } } pub fn get_selected_entry(&self) -> Option<FsEntry> { self.search_results .get(self.selected_search_result) .map(|s| s.entry.clone()) } pub fn render(&mut self, rows: usize, cols: usize) { let (start_index, selected_index_in_range, end_index) = calculate_list_bounds( self.search_results.len(), rows.saturating_sub(1), Some(self.selected_search_result), ); render_list_tip(3, cols); for i in start_index..end_index { if let Some(search_result) = self.search_results.get(i) { let is_selected = Some(i) == selected_index_in_range; let mut search_result_text = search_result.name(); let size = search_result .size() .map(|s| pretty_bytes(s as f64)) .unwrap_or("".to_owned()); if search_result.is_folder() { search_result_text.push('/'); } let search_result_text_width = search_result_text.width(); let size_width = size.width(); let text = if search_result_text_width + size_width < cols { let padding = " ".repeat( cols.saturating_sub(search_result_text_width) .saturating_sub(size_width), ); format!("{}{}{}", search_result_text, padding, size) } else { // drop the size, no room for it let padding = " ".repeat(cols.saturating_sub(search_result_text_width)); format!("{}{}", search_result_text, padding) }; let mut text_element = if is_selected { Text::new(text).selected() } else { Text::new(text) }; if search_result.is_folder() { text_element = text_element.color_range(0, ..); } text_element = text_element.color_indices(3, search_result.indices()); print_text_with_coordinates( text_element, 0, i.saturating_sub(start_index) + 4, Some(cols), None, ); } } } } #[derive(Debug)] pub struct SearchResult { pub entry: FsEntry, pub score: i64, pub indices: Vec<usize>, } impl SearchResult { pub fn new(entry: FsEntry, score: i64, indices: Vec<usize>) -> Self { SearchResult { entry, score, indices, } } pub fn name(&self) -> String { self.entry.name() } pub fn size(&self) -> Option<u64> { self.entry.size() } pub fn indices(&self) -> Vec<usize> { self.indices.clone() } pub fn is_folder(&self) -> bool { self.entry.is_folder() } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/strider/src/state.rs
default-plugins/strider/src/state.rs
use crate::file_list_view::{FileListView, FsEntry}; use crate::search_view::SearchView; use crate::shared::calculate_list_bounds; use std::{ collections::BTreeMap, path::{Path, PathBuf}, }; use zellij_tile::prelude::*; #[derive(Default)] pub struct State { pub file_list_view: FileListView, pub search_view: SearchView, pub hide_hidden_files: bool, pub current_rows: Option<usize>, pub handling_filepick_request_from: Option<(PipeSource, BTreeMap<String, String>)>, pub initial_cwd: PathBuf, pub is_searching: bool, pub search_term: String, pub close_on_selection: bool, } impl State { pub fn update_search_term(&mut self, character: char) { self.search_term.push(character); if &self.search_term == ".." { self.descend_to_previous_path(); } else if &self.search_term == "/" { self.descend_to_root_path(); } else { self.is_searching = true; self.search_view .update_search_results(&self.search_term, &self.file_list_view.files); } } pub fn handle_backspace(&mut self) { if self.search_term.is_empty() { self.descend_to_previous_path(); } else { self.search_term.pop(); if self.search_term.is_empty() { self.is_searching = false; } self.search_view .update_search_results(&self.search_term, &self.file_list_view.files); } } pub fn clear_search_term(&mut self) { self.search_term.clear(); self.search_view .update_search_results(&self.search_term, &self.file_list_view.files); self.is_searching = false; } pub fn clear_search_term_or_descend(&mut self) { if self.search_term.is_empty() { self.descend_to_previous_path(); } else { self.search_term.clear(); self.search_view .update_search_results(&self.search_term, &self.file_list_view.files); self.is_searching = false; } } pub fn move_selection_up(&mut self) { if self.is_searching { self.search_view.move_selection_up(); } else { self.file_list_view.move_selection_up(); } } pub fn move_selection_down(&mut self) { if self.is_searching { self.search_view.move_selection_down(); } else { self.file_list_view.move_selection_down(); } } pub fn handle_left_click(&mut self, line: isize) { if let Some(current_rows) = self.current_rows { let rows_for_list = current_rows.saturating_sub(5); if self.is_searching { let (start_index, _selected_index_in_range, _end_index) = calculate_list_bounds( self.search_view.search_result_count(), rows_for_list, Some(self.search_view.selected_search_result), ); let prev_selected = self.search_view.selected_search_result; self.search_view.selected_search_result = (line as usize).saturating_sub(4) + start_index; if prev_selected == self.search_view.selected_search_result { self.traverse_dir(); } } else { let (start_index, _selected_index_in_range, _end_index) = calculate_list_bounds( self.file_list_view.files.len(), rows_for_list, self.file_list_view.selected(), ); let prev_selected = self.file_list_view.selected(); *self.file_list_view.selected_mut() = (line as usize).saturating_sub(4) + start_index; if prev_selected == self.file_list_view.selected() { self.traverse_dir(); } } } } pub fn handle_mouse_hover(&mut self, line: isize) { if let Some(current_rows) = self.current_rows { let rows_for_list = current_rows.saturating_sub(5); if self.is_searching { let (start_index, _selected_index_in_range, _end_index) = calculate_list_bounds( self.search_view.search_result_count(), rows_for_list, Some(self.search_view.selected_search_result), ); self.search_view.selected_search_result = (line as usize).saturating_sub(4) + start_index; } else { let (start_index, _selected_index_in_range, _end_index) = calculate_list_bounds( self.file_list_view.files.len(), rows_for_list, self.file_list_view.selected(), ); *self.file_list_view.selected_mut() = (line as usize).saturating_sub(4) + start_index; } } } pub fn descend_to_previous_path(&mut self) { self.search_term.clear(); self.search_view.clear_and_reset_selection(); self.file_list_view.descend_to_previous_path(); } pub fn descend_to_root_path(&mut self) { self.search_term.clear(); self.search_view.clear_and_reset_selection(); self.file_list_view.descend_to_root_path(&self.initial_cwd); refresh_directory(&self.file_list_view.path); } pub fn toggle_hidden_files(&mut self) { self.hide_hidden_files = !self.hide_hidden_files; } pub fn traverse_dir(&mut self) { let entry = if self.is_searching { self.search_view.get_selected_entry() } else { self.file_list_view.get_selected_entry() }; if let Some(entry) = entry { match &entry { FsEntry::Dir(_p) => { self.file_list_view.enter_dir(&entry); self.search_view.clear_and_reset_selection(); refresh_directory(&self.file_list_view.path); }, FsEntry::File(_p, _) => { self.file_list_view.enter_dir(&entry); self.search_view.clear_and_reset_selection(); if self.handling_filepick_request_from.is_some() { self.send_filepick_response(); } else { self.open_selected_path(); } }, } } else if self.handling_filepick_request_from.is_some() { self.send_filepick_response(); } else { self.open_selected_path(); } self.is_searching = false; self.search_term.clear(); self.search_view.clear_and_reset_selection(); } pub fn update_files(&mut self, paths: Vec<(PathBuf, Option<FileMetadata>)>) { self.file_list_view .update_files(paths, self.hide_hidden_files); } pub fn open_selected_path(&mut self) { if self.file_list_view.path_is_dir { if self.close_on_selection { open_terminal_in_place_of_plugin(&self.file_list_view.path, true); } else { open_terminal(&self.file_list_view.path); } } else { if let Some(parent_folder) = self.file_list_view.path.parent() { if self.close_on_selection { open_file_in_place_of_plugin( FileToOpen::new(&self.file_list_view.path).with_cwd(parent_folder.into()), true, BTreeMap::new(), ); } else { open_file( FileToOpen::new(&self.file_list_view.path).with_cwd(parent_folder.into()), BTreeMap::new(), ); } } else { if self.close_on_selection { open_file_in_place_of_plugin( FileToOpen::new(&self.file_list_view.path), true, BTreeMap::new(), ); } else { open_file(FileToOpen::new(&self.file_list_view.path), BTreeMap::new()); } } } } pub fn send_filepick_response(&mut self) { let selected_path = &self.file_list_view.path; match &self.handling_filepick_request_from { Some((PipeSource::Plugin(plugin_id), args)) => { pipe_message_to_plugin( MessageToPlugin::new("filepicker_result") .with_destination_plugin_id(*plugin_id) .with_args(args.clone()) .with_payload(selected_path.display().to_string()), ); #[cfg(target_family = "wasm")] close_self(); }, Some((PipeSource::Cli(pipe_id), _args)) => { #[cfg(target_family = "wasm")] cli_pipe_output(pipe_id, &selected_path.display().to_string()); #[cfg(target_family = "wasm")] unblock_cli_pipe_input(pipe_id); #[cfg(target_family = "wasm")] close_self(); }, _ => {}, } } } pub(crate) fn refresh_directory(full_path: &Path) { change_host_folder(PathBuf::from(full_path)); }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/strider/src/file_list_view.rs
default-plugins/strider/src/file_list_view.rs
use crate::shared::{calculate_list_bounds, render_list_tip}; use crate::state::refresh_directory; use pretty_bytes::converter::convert as pretty_bytes; use std::collections::HashMap; use std::path::PathBuf; use unicode_width::UnicodeWidthStr; use zellij_tile::prelude::*; #[derive(Debug, Clone)] pub struct FileListView { pub path: PathBuf, pub path_is_dir: bool, pub files: Vec<FsEntry>, pub cursor_hist: HashMap<PathBuf, usize>, } impl Default for FileListView { fn default() -> Self { FileListView { path_is_dir: true, path: PathBuf::new(), files: Default::default(), cursor_hist: Default::default(), } } } impl FileListView { pub fn descend_to_previous_path(&mut self) { if let Some(parent) = self.path.parent() { self.path = parent.to_path_buf(); } else { self.path = PathBuf::new(); } self.path_is_dir = true; self.files.clear(); self.clear_selected(); refresh_directory(&self.path); } pub fn descend_to_root_path(&mut self, initial_cwd: &PathBuf) { self.path = initial_cwd.clone(); self.path_is_dir = true; self.files.clear(); self.clear_selected(); } pub fn enter_dir(&mut self, entry: &FsEntry) { let is_dir = entry.is_folder(); let path = entry.get_full_pathbuf(); self.path = path; self.path_is_dir = is_dir; self.files.clear(); self.clear_selected(); } pub fn clear_selected(&mut self) { self.cursor_hist.remove(&self.path); } pub fn update_files( &mut self, paths: Vec<(PathBuf, Option<FileMetadata>)>, hide_hidden_files: bool, ) { let mut files = vec![]; for (entry, entry_metadata) in paths { let entry = self .path .join(entry.strip_prefix("/host").unwrap_or(&entry)); if entry_metadata.map(|e| e.is_symlink).unwrap_or(false) { continue; } let entry = if entry_metadata.map(|e| e.is_dir).unwrap_or(false) { FsEntry::Dir(entry) } else { let size = entry_metadata.map(|e| e.len).unwrap_or(0); FsEntry::File(entry, size) }; if !entry.is_hidden_file() || !hide_hidden_files { files.push(entry); } } self.files = files; self.files.sort_unstable(); } pub fn get_selected_entry(&self) -> Option<FsEntry> { self.selected().and_then(|f| self.files.get(f).cloned()) } pub fn selected_mut(&mut self) -> &mut usize { self.cursor_hist.entry(self.path.clone()).or_default() } pub fn selected(&self) -> Option<usize> { self.cursor_hist.get(&self.path).copied() } pub fn move_selection_up(&mut self) { if let Some(selected) = self.selected() { *self.selected_mut() = selected.saturating_sub(1); } } pub fn move_selection_down(&mut self) { if let Some(selected) = self.selected() { let next = selected.saturating_add(1); *self.selected_mut() = std::cmp::min(self.files.len().saturating_sub(1), next); } else { *self.selected_mut() = 0; } } pub fn render(&mut self, rows: usize, cols: usize) { let (start_index, selected_index_in_range, end_index) = calculate_list_bounds(self.files.len(), rows.saturating_sub(1), self.selected()); render_list_tip(3, cols); for i in start_index..end_index { if let Some(entry) = self.files.get(i) { let is_selected = Some(i) == selected_index_in_range; let mut file_or_folder_name = entry.name(); let size = entry .size() .map(|s| pretty_bytes(s as f64)) .unwrap_or("".to_owned()); if entry.is_folder() { file_or_folder_name.push('/'); } let file_or_folder_name_width = file_or_folder_name.width(); let size_width = size.width(); let text = if file_or_folder_name_width + size_width < cols { let padding = " ".repeat( cols.saturating_sub(file_or_folder_name_width) .saturating_sub(size_width), ); format!("{}{}{}", file_or_folder_name, padding, size) } else { let padding = " ".repeat(cols.saturating_sub(file_or_folder_name_width)); format!("{}{}", file_or_folder_name, padding) }; let mut text_element = if is_selected { Text::new(text).selected() } else { Text::new(text) }; if entry.is_folder() { text_element = text_element.color_range(0, ..); } print_text_with_coordinates( text_element, 0, 4 + i.saturating_sub(start_index), Some(cols), None, ); } } } } #[derive(PartialEq, Eq, PartialOrd, Ord, Clone, Debug)] pub enum FsEntry { Dir(PathBuf), File(PathBuf, u64), } impl FsEntry { pub fn name(&self) -> String { let path = match self { FsEntry::Dir(p) => p, FsEntry::File(p, _) => p, }; path.file_name().unwrap().to_string_lossy().into_owned() } pub fn size(&self) -> Option<u64> { match self { FsEntry::Dir(_p) => None, FsEntry::File(_, size) => Some(*size), } } pub fn get_full_pathbuf(&self) -> PathBuf { match self { FsEntry::Dir(p) => p.clone(), FsEntry::File(p, _) => p.clone(), } } pub fn is_hidden_file(&self) -> bool { self.name().starts_with('.') } pub fn is_folder(&self) -> bool { matches!(self, FsEntry::Dir(_)) } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/strider/src/main.rs
default-plugins/strider/src/main.rs
mod file_list_view; mod search_view; mod shared; mod state; use shared::{render_current_path, render_instruction_line, render_search_term}; use state::{refresh_directory, State}; use std::collections::BTreeMap; use std::path::PathBuf; use zellij_tile::prelude::*; register_plugin!(State); impl ZellijPlugin for State { fn load(&mut self, configuration: BTreeMap<String, String>) { let plugin_ids = get_plugin_ids(); self.initial_cwd = plugin_ids.initial_cwd; let show_hidden_files = configuration .get("show_hidden_files") .map(|v| v == "true") .unwrap_or(false); self.hide_hidden_files = !show_hidden_files; self.close_on_selection = configuration .get("close_on_selection") .map(|v| v == "true") .unwrap_or(false); subscribe(&[ EventType::Key, EventType::Mouse, EventType::CustomMessage, EventType::Timer, EventType::FileSystemUpdate, EventType::HostFolderChanged, EventType::PermissionRequestResult, ]); self.file_list_view.clear_selected(); match configuration.get("caller_cwd").map(|c| PathBuf::from(c)) { Some(caller_cwd) => { self.file_list_view.path = caller_cwd; }, None => { self.file_list_view.path = self.initial_cwd.clone(); }, } if self.initial_cwd != self.file_list_view.path { change_host_folder(self.file_list_view.path.clone()); } else { scan_host_folder(&"/host"); } } fn update(&mut self, event: Event) -> bool { let mut should_render = false; match event { Event::FileSystemUpdate(paths) => { self.update_files(paths); should_render = true; }, Event::HostFolderChanged(_new_host_folder) => { scan_host_folder(&"/host"); should_render = true; }, Event::Key(key) => match key.bare_key { BareKey::Char(character) if key.has_no_modifiers() => { self.update_search_term(character); should_render = true; }, BareKey::Backspace if key.has_no_modifiers() => { self.handle_backspace(); should_render = true; }, BareKey::Esc if key.has_no_modifiers() => { if self.is_searching { self.clear_search_term(); } else { self.file_list_view.clear_selected(); } should_render = true; }, BareKey::Char('c') if key.has_modifiers(&[KeyModifier::Ctrl]) => { self.clear_search_term_or_descend(); }, BareKey::Up if key.has_no_modifiers() => { self.move_selection_up(); should_render = true; }, BareKey::Down if key.has_no_modifiers() => { self.move_selection_down(); should_render = true; }, BareKey::Right | BareKey::Tab | BareKey::Enter if key.has_no_modifiers() => { self.traverse_dir(); should_render = true; }, BareKey::Right if key.has_no_modifiers() => { self.traverse_dir(); should_render = true; }, BareKey::Left if key.has_no_modifiers() => { self.descend_to_previous_path(); should_render = true; }, BareKey::Char('e') if key.has_modifiers(&[KeyModifier::Ctrl]) => { should_render = true; self.toggle_hidden_files(); refresh_directory(&self.file_list_view.path); }, _ => (), }, Event::Mouse(mouse_event) => match mouse_event { Mouse::ScrollDown(_) => { self.move_selection_down(); should_render = true; }, Mouse::ScrollUp(_) => { self.move_selection_up(); should_render = true; }, Mouse::LeftClick(line, _) => { self.handle_left_click(line); should_render = true; }, Mouse::Hover(line, _) => { if line >= 0 { self.handle_mouse_hover(line); should_render = true; } }, _ => {}, }, _ => { dbg!("Unknown event {:?}", event); }, }; should_render } fn pipe(&mut self, pipe_message: PipeMessage) -> bool { if pipe_message.is_private && pipe_message.name == "filepicker" { if let PipeSource::Cli(pipe_id) = &pipe_message.source { #[cfg(target_family = "wasm")] block_cli_pipe_input(pipe_id); } self.handling_filepick_request_from = Some((pipe_message.source, pipe_message.args)); true } else { false } } fn render(&mut self, rows: usize, cols: usize) { self.current_rows = Some(rows); let rows_for_list = rows.saturating_sub(6); render_search_term(&self.search_term); render_current_path( &self.file_list_view.path, self.file_list_view.path_is_dir, self.handling_filepick_request_from.is_some(), cols, ); if self.is_searching { self.search_view.render(rows_for_list, cols); } else { self.file_list_view.render(rows_for_list, cols); } render_instruction_line(rows, cols); } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/strider/src/shared.rs
default-plugins/strider/src/shared.rs
use std::path::PathBuf; use unicode_width::UnicodeWidthStr; use zellij_tile::prelude::*; pub fn render_instruction_line(y: usize, max_cols: usize) { if max_cols > 78 { let text = "Help: go back with <Ctrl c>, go to root with /, <Ctrl e> - toggle hidden files"; let text = Text::new(text) .color_range(3, 19..27) .color_range(3, 45..46) .color_range(3, 48..56); print_text_with_coordinates(text, 0, y, Some(max_cols), None); } else if max_cols > 56 { let text = "Help: <Ctrl c> - back, / - root, <Ctrl e> - hidden files"; let text = Text::new(text) .color_range(3, 6..14) .color_range(3, 23..24) .color_range(3, 33..41); print_text_with_coordinates(text, 0, y, Some(max_cols), None); } else if max_cols > 25 { let text = "<Ctrl c> - back, / - root"; let text = Text::new(text).color_range(3, ..8).color_range(3, 17..18); print_text_with_coordinates(text, 0, y, Some(max_cols), None); } } pub fn render_list_tip(y: usize, max_cols: usize) { let tip = Text::new(format!("(<↓↑> - Navigate, <TAB> - Select)")) .color_range(3, 1..5) .color_range(3, 18..23); print_text_with_coordinates(tip, 0, y, Some(max_cols), None); } pub fn calculate_list_bounds( result_count: usize, max_result_count: usize, selected_index_in_all_results: Option<usize>, ) -> (usize, Option<usize>, usize) { match selected_index_in_all_results { Some(selected_index_in_all_results) => { let mut room_in_list = max_result_count; let mut start_index = selected_index_in_all_results; let mut end_index = selected_index_in_all_results + 1; let mut alternate = false; loop { if room_in_list == 0 { break; } if !alternate && start_index > 0 { start_index = start_index.saturating_sub(1); room_in_list = room_in_list.saturating_sub(1); } else if alternate && end_index < result_count { end_index += 1; room_in_list = room_in_list.saturating_sub(1); } else if start_index > 0 { start_index = start_index.saturating_sub(1); room_in_list = room_in_list.saturating_sub(1); } else if end_index < result_count { end_index += 1; room_in_list = room_in_list.saturating_sub(1); } else { break; } alternate = !alternate; } (start_index, Some(selected_index_in_all_results), end_index) }, None => (0, None, max_result_count + 1), } } pub fn render_search_term(search_term: &str) { let prompt = "FIND: "; let text = Text::new(format!("{}{}_", prompt, search_term)) .color_range(2, 0..prompt.len()) .color_range(3, prompt.len()..); print_text(text); println!("") } pub fn render_current_path( full_path: &PathBuf, path_is_dir: bool, handling_filepick: bool, max_cols: usize, ) { let prompt = "PATH: "; let current_path = full_path.display().to_string(); let prompt_len = prompt.width(); let current_path_len = current_path.width(); let enter_tip = if handling_filepick { "Select" } else if path_is_dir { "Open terminal here" } else { "Open in editor" }; if max_cols > prompt_len + current_path_len + enter_tip.width() + 13 { let path_end = prompt_len + current_path_len; let current_path = Text::new(format!( "{}{} (<ENTER> - {})", prompt, current_path, enter_tip )) .color_range(2, 0..prompt_len) .color_range(0, prompt_len..path_end) .color_range(3, path_end + 2..path_end + 9); print_text(current_path); } else { let max_path_len = max_cols .saturating_sub(prompt_len) .saturating_sub(8) .saturating_sub(prompt_len); let current_path = if current_path_len <= max_path_len { current_path } else { truncate_path( full_path.clone(), current_path_len.saturating_sub(max_path_len), ) }; let current_path_len = current_path.width(); let path_end = prompt_len + current_path_len; let current_path = Text::new(format!("{}{} <ENTER>", prompt, current_path)) .color_range(2, 0..prompt_len) .color_range(0, prompt_len..path_end) .color_range(3, path_end + 1..path_end + 9); print_text(current_path); } println!(); println!(); } fn truncate_path(path: PathBuf, mut char_count_to_remove: usize) -> String { let mut truncated = String::new(); let component_count = path.iter().count(); for (i, component) in path.iter().enumerate() { let mut component_str = component.to_string_lossy().to_string(); if char_count_to_remove > 0 { truncated.push(component_str.remove(0)); if i != 0 && i + 1 != component_count { truncated.push('/'); } char_count_to_remove = char_count_to_remove.saturating_sub(component_str.width() + 1); } else { truncated.push_str(&component_str); if i != 0 && i + 1 != component_count { truncated.push('/'); } } } truncated }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/compact-bar/src/tooltip.rs
default-plugins/compact-bar/src/tooltip.rs
use crate::keybind_utils::KeybindProcessor; use zellij_tile::prelude::*; pub struct TooltipRenderer<'a> { mode_info: &'a ModeInfo, } impl<'a> TooltipRenderer<'a> { pub fn new(mode_info: &'a ModeInfo) -> Self { Self { mode_info } } pub fn render(&self, rows: usize, cols: usize) { let current_mode = self.mode_info.mode; if current_mode == InputMode::Normal { let (text_components, tooltip_rows, tooltip_columns) = self.normal_mode_tooltip(current_mode); let base_x = cols.saturating_sub(tooltip_columns) / 2; let base_y = rows.saturating_sub(tooltip_rows) / 2; for (text, ribbon, x, y) in text_components { let text_width = text.content().chars().count(); let ribbon_content_width = ribbon.content().chars().count(); let ribbon_total_width = ribbon_content_width + 4; let total_element_width = text_width + ribbon_total_width + 1; // Check if this element would exceed the available columns and render an ellipses // if it does if base_x + x + total_element_width > cols { let remaining_space = cols.saturating_sub(base_x + x); let ellipsis = Text::new("...").opaque(); print_text_with_coordinates( ellipsis, base_x + x, base_y + y, Some(remaining_space), None, ); break; } print_text_with_coordinates(text, base_x + x, base_y + y, None, None); print_ribbon_with_coordinates( ribbon, base_x + x + text_width, base_y + y, None, None, ); } } else { let (table, tooltip_rows, tooltip_columns) = self.other_mode_tooltip(current_mode); let base_x = cols.saturating_sub(tooltip_columns) / 2; let base_y = rows.saturating_sub(tooltip_rows) / 2; print_table_with_coordinates(table, base_x, base_y, None, None); } } pub fn calculate_dimensions(&self, current_mode: InputMode) -> (usize, usize) { match current_mode { InputMode::Normal => { let (_, tooltip_rows, tooltip_cols) = self.normal_mode_tooltip(current_mode); (tooltip_rows, tooltip_cols) }, _ => { let (_, tooltip_rows, tooltip_cols) = self.other_mode_tooltip(current_mode); (tooltip_rows + 1, tooltip_cols) // + 1 for the invisible table title }, } } fn normal_mode_tooltip( &self, current_mode: InputMode, ) -> (Vec<(Text, Text, usize, usize)>, usize, usize) { let actions = KeybindProcessor::get_predetermined_actions(self.mode_info, current_mode); let y = 0; let mut running_x = 0; let mut components = Vec::new(); let mut max_columns = 0; let mut is_first = true; for (key, description) in actions { let text = if is_first { Text::new(format!("{} ", &key)).color_all(3).opaque() } else { Text::new(format!(" {} ", &key)).color_all(3).opaque() }; let ribbon = Text::new(&description); let line_length = if is_first { key.chars().count() + description.chars().count() } else { key.chars().count() + 1 + description.chars().count() }; components.push((text, ribbon, running_x, y)); running_x += line_length + 5; max_columns = max_columns.max(running_x); is_first = false; } let total_rows = 1; (components, total_rows, max_columns) } fn other_mode_tooltip(&self, current_mode: InputMode) -> (Table, usize, usize) { let actions = KeybindProcessor::get_predetermined_actions(self.mode_info, current_mode); let actions_vec: Vec<_> = actions.into_iter().collect(); let mut table = Table::new().add_row(vec![" ".to_owned(); 2]); let mut row_count = 1; // Start with header row if actions_vec.is_empty() { let tooltip_text = match self.mode_info.mode { InputMode::EnterSearch => "Entering search term...".to_owned(), InputMode::RenameTab => "Renaming tab...".to_owned(), InputMode::RenamePane => "Renaming pane...".to_owned(), _ => { format!("{:?}", self.mode_info.mode) }, }; let total_width = tooltip_text.chars().count(); table = table.add_styled_row(vec![Text::new(tooltip_text).color_all(0)]); row_count += 1; (table, row_count, total_width) } else { let mut key_width = 0; let mut action_width = 0; for (key, description) in actions_vec.into_iter() { let description_formatted = format!("- {}", description); key_width = key_width.max(key.chars().count()); action_width = action_width.max(description_formatted.chars().count()); table = table.add_styled_row(vec![ Text::new(&key).color_all(3), Text::new(description_formatted), ]); row_count += 1; } let total_width = key_width + action_width + 1; // +1 for separator (table, row_count, total_width) } } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/compact-bar/src/action_types.rs
default-plugins/compact-bar/src/action_types.rs
use zellij_tile::prelude::actions::Action; use zellij_tile::prelude::*; #[derive(Debug, Clone, PartialEq, Eq, Hash)] pub enum ActionType { MoveFocus, MovePaneWithDirection, MovePaneWithoutDirection, ResizeIncrease, ResizeDecrease, ResizeAny, Search, NewPaneWithDirection, NewPaneWithoutDirection, BreakPaneLeftOrRight, GoToAdjacentTab, Scroll, PageScroll, HalfPageScroll, SessionManager, Configuration, PluginManager, About, SwitchToMode(InputMode), TogglePaneEmbedOrFloating, ToggleFocusFullscreen, ToggleFloatingPanes, CloseFocus, CloseTab, ToggleActiveSyncTab, ToggleTab, BreakPane, EditScrollback, NewTab, Detach, Quit, NewStackedPane, Other(String), // Fallback for unhandled actions } impl ActionType { pub fn description(&self) -> String { match self { ActionType::MoveFocus => "Move focus".to_string(), ActionType::MovePaneWithDirection => "Move pane".to_string(), ActionType::MovePaneWithoutDirection => "Move pane".to_string(), ActionType::ResizeIncrease => "Increase size in direction".to_string(), ActionType::ResizeDecrease => "Decrease size in direction".to_string(), ActionType::ResizeAny => "Increase or decrease size".to_string(), ActionType::Search => "Search".to_string(), ActionType::NewPaneWithDirection => "Split right/down".to_string(), ActionType::NewPaneWithoutDirection => "New pane".to_string(), ActionType::BreakPaneLeftOrRight => "Break pane to adjacent tab".to_string(), ActionType::GoToAdjacentTab => "Move tab focus".to_string(), ActionType::Scroll => "Scroll".to_string(), ActionType::PageScroll => "Scroll page".to_string(), ActionType::HalfPageScroll => "Scroll half Page".to_string(), ActionType::SessionManager => "Session manager".to_string(), ActionType::PluginManager => "Plugin manager".to_string(), ActionType::Configuration => "Configuration".to_string(), ActionType::About => "About Zellij".to_string(), ActionType::SwitchToMode(input_mode) if input_mode == &InputMode::RenamePane => { "Rename pane".to_string() }, ActionType::SwitchToMode(input_mode) if input_mode == &InputMode::RenameTab => { "Rename tab".to_string() }, ActionType::SwitchToMode(input_mode) if input_mode == &InputMode::EnterSearch => { "Search".to_string() }, ActionType::SwitchToMode(input_mode) if input_mode == &InputMode::Locked => { "Lock".to_string() }, ActionType::SwitchToMode(input_mode) if input_mode == &InputMode::Normal => { "Unlock".to_string() }, ActionType::SwitchToMode(input_mode) => format!("{:?}", input_mode), ActionType::TogglePaneEmbedOrFloating => "Float or embed".to_string(), ActionType::NewStackedPane => "New stacked pane".to_string(), ActionType::ToggleFocusFullscreen => "Toggle fullscreen".to_string(), ActionType::ToggleFloatingPanes => "Show/hide floating panes".to_string(), ActionType::CloseFocus => "Close pane".to_string(), ActionType::CloseTab => "Close tab".to_string(), ActionType::ToggleActiveSyncTab => "Sync panes in tab".to_string(), ActionType::ToggleTab => "Circle tab focus".to_string(), ActionType::BreakPane => "Break pane to new tab".to_string(), ActionType::EditScrollback => "Open pane scrollback in editor".to_string(), ActionType::NewTab => "New tab".to_string(), ActionType::Detach => "Detach".to_string(), ActionType::Quit => "Quit".to_string(), ActionType::Other(_) => "Other action".to_string(), } } pub fn from_action(action: &Action) -> Self { match action { Action::MoveFocus { .. } => ActionType::MoveFocus, Action::MovePane { direction: Some(_) } => ActionType::MovePaneWithDirection, Action::MovePane { direction: None } => ActionType::MovePaneWithoutDirection, Action::Resize { resize: Resize::Increase, direction: Some(_), } => ActionType::ResizeIncrease, Action::Resize { resize: Resize::Decrease, direction: Some(_), } => ActionType::ResizeDecrease, Action::Resize { resize: _, direction: None, } => ActionType::ResizeAny, Action::Search { .. } => ActionType::Search, Action::NewPane { direction: Some(_), .. } => ActionType::NewPaneWithDirection, Action::NewPane { direction: None, .. } => ActionType::NewPaneWithoutDirection, Action::NewStackedPane { .. } => ActionType::NewStackedPane, Action::BreakPaneLeft | Action::BreakPaneRight => ActionType::BreakPaneLeftOrRight, Action::GoToPreviousTab | Action::GoToNextTab => ActionType::GoToAdjacentTab, Action::ScrollUp | Action::ScrollDown => ActionType::Scroll, Action::PageScrollUp | Action::PageScrollDown => ActionType::PageScroll, Action::HalfPageScrollUp | Action::HalfPageScrollDown => ActionType::HalfPageScroll, Action::SwitchToMode { input_mode } => ActionType::SwitchToMode(*input_mode), Action::TogglePaneEmbedOrFloating => ActionType::TogglePaneEmbedOrFloating, Action::ToggleFocusFullscreen => ActionType::ToggleFocusFullscreen, Action::ToggleFloatingPanes => ActionType::ToggleFloatingPanes, Action::CloseFocus => ActionType::CloseFocus, Action::CloseTab => ActionType::CloseTab, Action::ToggleActiveSyncTab => ActionType::ToggleActiveSyncTab, Action::ToggleTab => ActionType::ToggleTab, Action::BreakPane => ActionType::BreakPane, Action::EditScrollback => ActionType::EditScrollback, Action::Detach => ActionType::Detach, Action::Quit => ActionType::Quit, action if action.launches_plugin("session-manager") => ActionType::SessionManager, action if action.launches_plugin("configuration") => ActionType::Configuration, action if action.launches_plugin("plugin-manager") => ActionType::PluginManager, action if action.launches_plugin("zellij:about") => ActionType::About, action if matches!(action, Action::NewTab { .. }) => ActionType::NewTab, _ => ActionType::Other(format!("{:?}", action)), } } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/compact-bar/src/keybind_utils.rs
default-plugins/compact-bar/src/keybind_utils.rs
use crate::action_types::ActionType; use std::collections::HashSet; use zellij_tile::prelude::actions::Action; use zellij_tile::prelude::*; pub struct KeybindProcessor; impl KeybindProcessor { /// Find predetermined actions based on predicates while maintaining order pub fn find_predetermined_actions<F>( mode_info: &ModeInfo, mode: InputMode, predicates: Vec<F>, ) -> Vec<(String, String)> where F: Fn(&Action) -> bool, { let mut result = Vec::new(); let keybinds = mode_info.get_keybinds_for_mode(mode); let mut processed_action_types = HashSet::new(); // Iterate through predicates in order to maintain the desired sequence for predicate in predicates { // Find the first matching action for this predicate let mut found_match = false; for (_key, actions) in &keybinds { if let Some(first_action) = actions.first() { if predicate(first_action) { let action_type = ActionType::from_action(first_action); // Skip if we've already processed this action type if processed_action_types.contains(&action_type) { found_match = true; break; } let mut matching_keys = Vec::new(); // Find all keys that match this action type (including different directions) for (inner_key, inner_actions) in &keybinds { if let Some(inner_first_action) = inner_actions.first() { if ActionType::from_action(inner_first_action) == action_type { matching_keys.push(format!("{}", inner_key)); } } } if !matching_keys.is_empty() { let description = action_type.description(); let should_add_brackets_to_keys = mode != InputMode::Normal; // Check if this is switching to normal mode let is_switching_to_locked = matches!( first_action, Action::SwitchToMode { input_mode: InputMode::Locked } ); let grouped_keys = Self::group_key_sets( &matching_keys, should_add_brackets_to_keys, is_switching_to_locked, ); result.push((grouped_keys, description)); processed_action_types.insert(action_type); } found_match = true; break; } } } // If we found a match for this predicate, we've processed it if found_match { continue; } } result } /// Group keys into sets and separate different key types with '|' fn group_key_sets( keys: &[String], should_add_brackets_to_keys: bool, is_switching_to_locked: bool, ) -> String { if keys.is_empty() { return String::new(); } // Filter out Esc and Enter keys when switching to normal mode, but only if other keys exist let filtered_keys: Vec<String> = if is_switching_to_locked { let non_esc_enter_keys: Vec<String> = keys .iter() .filter(|k| k.as_str() != "ESC" && k.as_str() != "ENTER") .cloned() .collect(); if non_esc_enter_keys.is_empty() { // If no other keys exist, keep the original keys keys.to_vec() } else { // Use filtered keys (without Esc/Enter) non_esc_enter_keys } } else { keys.to_vec() }; if filtered_keys.len() == 1 { return if should_add_brackets_to_keys { format!("<{}>", filtered_keys[0]) } else { filtered_keys[0].clone() }; } // Group keys by type let mut arrow_keys = Vec::new(); let mut hjkl_lower = Vec::new(); let mut hjkl_upper = Vec::new(); let mut square_bracket_keys = Vec::new(); let mut plus_minus_keys = Vec::new(); let mut pgup_pgdown = Vec::new(); let mut other_keys = Vec::new(); for key in &filtered_keys { match key.as_str() { "Left" | "←" => arrow_keys.push("←"), "Down" | "↓" => arrow_keys.push("↓"), "Up" | "↑" => arrow_keys.push("↑"), "Right" | "→" => arrow_keys.push("→"), "h" => hjkl_lower.push("h"), "j" => hjkl_lower.push("j"), "k" => hjkl_lower.push("k"), "l" => hjkl_lower.push("l"), "H" => hjkl_upper.push("H"), "J" => hjkl_upper.push("J"), "K" => hjkl_upper.push("K"), "L" => hjkl_upper.push("L"), "[" => square_bracket_keys.push("["), "]" => square_bracket_keys.push("]"), "+" => plus_minus_keys.push("+"), "-" => plus_minus_keys.push("-"), "=" => plus_minus_keys.push("="), "PgUp" => pgup_pgdown.push("PgUp"), "PgDn" => pgup_pgdown.push("PgDn"), _ => { if should_add_brackets_to_keys { other_keys.push(format!("<{}>", key)); } else { other_keys.push(key.clone()); } }, } } let mut groups = Vec::new(); // Add hjkl group if present (prioritize hjkl over arrows) if !hjkl_lower.is_empty() { Self::sort_hjkl(&mut hjkl_lower); groups.push(Self::format_key_group( &hjkl_lower, should_add_brackets_to_keys, false, )); } // Add HJKL group if present if !hjkl_upper.is_empty() { Self::sort_hjkl_upper(&mut hjkl_upper); groups.push(Self::format_key_group( &hjkl_upper, should_add_brackets_to_keys, false, )); } // Add arrow keys group if present if !arrow_keys.is_empty() { Self::sort_arrows(&mut arrow_keys); groups.push(Self::format_key_group( &arrow_keys, should_add_brackets_to_keys, false, )); } if !square_bracket_keys.is_empty() { Self::sort_square_brackets(&mut square_bracket_keys); groups.push(Self::format_key_group( &square_bracket_keys, should_add_brackets_to_keys, false, )); } if !plus_minus_keys.is_empty() { Self::sort_plus_minus(&mut plus_minus_keys); groups.push(Self::format_key_group( &plus_minus_keys, should_add_brackets_to_keys, false, )); } if !pgup_pgdown.is_empty() { Self::sort_pgup_pgdown(&mut pgup_pgdown); groups.push(Self::format_key_group( &pgup_pgdown, should_add_brackets_to_keys, true, )); } // Add other keys with / separator if !other_keys.is_empty() { groups.push(other_keys.join("/")); } groups.join("/") } fn sort_hjkl(keys: &mut Vec<&str>) { keys.sort_by(|a, b| { let order = ["h", "j", "k", "l"]; let pos_a = order.iter().position(|&x| &x == a).unwrap_or(usize::MAX); let pos_b = order.iter().position(|&x| &x == b).unwrap_or(usize::MAX); pos_a.cmp(&pos_b) }); } fn sort_hjkl_upper(keys: &mut Vec<&str>) { keys.sort_by(|a, b| { let order = ["H", "J", "K", "L"]; let pos_a = order.iter().position(|&x| &x == a).unwrap_or(usize::MAX); let pos_b = order.iter().position(|&x| &x == b).unwrap_or(usize::MAX); pos_a.cmp(&pos_b) }); } fn sort_arrows(keys: &mut Vec<&str>) { keys.sort(); keys.dedup(); keys.sort_by(|a, b| { let order = ["←", "↓", "↑", "→"]; let pos_a = order.iter().position(|&x| &x == a).unwrap_or(usize::MAX); let pos_b = order.iter().position(|&x| &x == b).unwrap_or(usize::MAX); pos_a.cmp(&pos_b) }); } fn sort_square_brackets(keys: &mut Vec<&str>) { keys.sort_by(|a, b| { let order = ["[", "]"]; let pos_a = order.iter().position(|&x| &x == a).unwrap_or(usize::MAX); let pos_b = order.iter().position(|&x| &x == b).unwrap_or(usize::MAX); pos_a.cmp(&pos_b) }); } fn sort_plus_minus(keys: &mut Vec<&str>) { keys.sort_by(|a, b| { let order = ["+", "-"]; let pos_a = order.iter().position(|&x| &x == a).unwrap_or(usize::MAX); let pos_b = order.iter().position(|&x| &x == b).unwrap_or(usize::MAX); pos_a.cmp(&pos_b) }); // Remove "=" if both "+" and "=" are present if keys.contains(&"+") && keys.contains(&"=") { keys.retain(|k| k != &"="); } } fn sort_pgup_pgdown(keys: &mut Vec<&str>) { keys.sort_by(|a, b| { let order = ["PgUp", "PgDn"]; let pos_a = order.iter().position(|&x| &x == a).unwrap_or(usize::MAX); let pos_b = order.iter().position(|&x| &x == b).unwrap_or(usize::MAX); pos_a.cmp(&pos_b) }); } fn format_key_group( keys: &[&str], should_add_brackets: bool, use_pipe_separator: bool, ) -> String { let separator = if use_pipe_separator { "|" } else { "" }; let joined = keys.join(separator); if should_add_brackets { format!("<{}>", joined) } else { joined } } /// Get predetermined actions for a specific mode pub fn get_predetermined_actions( mode_info: &ModeInfo, mode: InputMode, ) -> Vec<(String, String)> { match mode { InputMode::Locked => { let ordered_predicates = vec![|action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::Normal } ) }]; Self::find_predetermined_actions(mode_info, mode, ordered_predicates) }, InputMode::Normal => { let ordered_predicates = vec![ |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::Locked } ) }, |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::Pane } ) }, |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::Tab } ) }, |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::Resize } ) }, |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::Move } ) }, |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::Scroll } ) }, |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::Session } ) }, |action: &Action| matches!(action, Action::Quit), ]; Self::find_predetermined_actions(mode_info, mode, ordered_predicates) }, InputMode::Pane => { let ordered_predicates = vec![ |action: &Action| { matches!( action, Action::NewPane { direction: None, pane_name: None, start_suppressed: false } ) }, |action: &Action| { matches!( action, Action::MoveFocus { direction: Direction::Left } ) }, |action: &Action| { matches!( action, Action::MoveFocus { direction: Direction::Down } ) }, |action: &Action| { matches!( action, Action::MoveFocus { direction: Direction::Up } ) }, |action: &Action| { matches!( action, Action::MoveFocus { direction: Direction::Right } ) }, |action: &Action| matches!(action, Action::CloseFocus), |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::RenamePane } ) }, |action: &Action| matches!(action, Action::ToggleFocusFullscreen), |action: &Action| matches!(action, Action::ToggleFloatingPanes), |action: &Action| matches!(action, Action::TogglePaneEmbedOrFloating), |action: &Action| { matches!( action, Action::NewStackedPane { command: None, pane_name: None, near_current_pane: false, } ) }, |action: &Action| { matches!( action, Action::NewPane { direction: Some(Direction::Right), pane_name: None, start_suppressed: false } ) }, |action: &Action| { matches!( action, Action::NewPane { direction: Some(Direction::Down), pane_name: None, start_suppressed: false } ) }, ]; Self::find_predetermined_actions(mode_info, mode, ordered_predicates) }, InputMode::Tab => { let ordered_predicates = vec![ |action: &Action| matches!(action, Action::GoToPreviousTab), |action: &Action| matches!(action, Action::GoToNextTab), |action: &Action| { matches!( action, Action::NewTab { tiled_layout: None, floating_layouts: _, swap_tiled_layouts: None, swap_floating_layouts: None, tab_name: None, should_change_focus_to_new_tab: true, cwd: None, initial_panes: _, first_pane_unblock_condition: _, } ) }, |action: &Action| matches!(action, Action::CloseTab), |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::RenameTab } ) }, |action: &Action| matches!(action, Action::TabNameInput { .. }), |action: &Action| matches!(action, Action::ToggleActiveSyncTab), |action: &Action| matches!(action, Action::BreakPane), |action: &Action| matches!(action, Action::BreakPaneLeft), |action: &Action| matches!(action, Action::BreakPaneRight), |action: &Action| matches!(action, Action::ToggleTab), ]; Self::find_predetermined_actions(mode_info, mode, ordered_predicates) }, InputMode::Resize => { let ordered_predicates = vec![ |action: &Action| { matches!( action, Action::Resize { resize: Resize::Increase, direction: None } ) }, |action: &Action| { matches!( action, Action::Resize { resize: Resize::Decrease, direction: None } ) }, |action: &Action| { matches!( action, Action::Resize { resize: Resize::Increase, direction: Some(Direction::Left) } ) }, |action: &Action| { matches!( action, Action::Resize { resize: Resize::Increase, direction: Some(Direction::Down) } ) }, |action: &Action| { matches!( action, Action::Resize { resize: Resize::Increase, direction: Some(Direction::Up) } ) }, |action: &Action| { matches!( action, Action::Resize { resize: Resize::Increase, direction: Some(Direction::Right) } ) }, |action: &Action| { matches!( action, Action::Resize { resize: Resize::Decrease, direction: Some(Direction::Left) } ) }, |action: &Action| { matches!( action, Action::Resize { resize: Resize::Decrease, direction: Some(Direction::Down) } ) }, |action: &Action| { matches!( action, Action::Resize { resize: Resize::Decrease, direction: Some(Direction::Up) } ) }, |action: &Action| { matches!( action, Action::Resize { resize: Resize::Decrease, direction: Some(Direction::Right) } ) }, ]; Self::find_predetermined_actions(mode_info, mode, ordered_predicates) }, InputMode::Move => { let ordered_predicates = vec![ |action: &Action| { matches!( action, Action::MovePane { direction: Some(Direction::Left) } ) }, |action: &Action| { matches!( action, Action::MovePane { direction: Some(Direction::Down) } ) }, |action: &Action| { matches!( action, Action::MovePane { direction: Some(Direction::Up) } ) }, |action: &Action| { matches!( action, Action::MovePane { direction: Some(Direction::Right) } ) }, ]; Self::find_predetermined_actions(mode_info, mode, ordered_predicates) }, InputMode::Scroll => { let ordered_predicates = vec![ |action: &Action| matches!(action, Action::ScrollDown), |action: &Action| matches!(action, Action::ScrollUp), |action: &Action| matches!(action, Action::HalfPageScrollDown), |action: &Action| matches!(action, Action::HalfPageScrollUp), |action: &Action| matches!(action, Action::PageScrollDown), |action: &Action| matches!(action, Action::PageScrollUp), |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::EnterSearch } ) }, |action: &Action| matches!(action, Action::EditScrollback), ]; Self::find_predetermined_actions(mode_info, mode, ordered_predicates) }, InputMode::Search => { let ordered_predicates = vec![ |action: &Action| { matches!( action, Action::SwitchToMode { input_mode: InputMode::EnterSearch } ) }, |action: &Action| matches!(action, Action::SearchInput { .. }), |action: &Action| matches!(action, Action::ScrollDown), |action: &Action| matches!(action, Action::ScrollUp), |action: &Action| matches!(action, Action::PageScrollDown), |action: &Action| matches!(action, Action::PageScrollUp), |action: &Action| matches!(action, Action::HalfPageScrollDown), |action: &Action| matches!(action, Action::HalfPageScrollUp), |action: &Action| { matches!( action, Action::Search { direction: actions::SearchDirection::Down } ) }, |action: &Action| { matches!( action, Action::Search { direction: actions::SearchDirection::Up } ) }, |action: &Action| { matches!( action, Action::SearchToggleOption { option: actions::SearchOption::CaseSensitivity } ) }, |action: &Action| { matches!( action, Action::SearchToggleOption { option: actions::SearchOption::Wrap } ) }, |action: &Action| { matches!( action, Action::SearchToggleOption { option: actions::SearchOption::WholeWord } ) }, ]; Self::find_predetermined_actions(mode_info, mode, ordered_predicates) }, InputMode::Session => { let ordered_predicates = vec![ |action: &Action| matches!(action, Action::Detach), |action: &Action| action.launches_plugin("session-manager"), |action: &Action| action.launches_plugin("plugin-manager"), |action: &Action| action.launches_plugin("configuration"), |action: &Action| action.launches_plugin("zellij:about"), ]; Self::find_predetermined_actions(mode_info, mode, ordered_predicates) }, InputMode::EnterSearch | InputMode::RenameTab | InputMode::RenamePane | InputMode::Prompt | InputMode::Tmux => Vec::new(), } } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/compact-bar/src/clipboard_utils.rs
default-plugins/compact-bar/src/clipboard_utils.rs
use crate::LinePart; use zellij_tile::prelude::*; pub fn text_copied_hint(copy_destination: CopyDestination) -> LinePart { let hint = match copy_destination { CopyDestination::Command => "Text piped to external command", #[cfg(not(target_os = "macos"))] CopyDestination::Primary => "Text copied to system primary selection", #[cfg(target_os = "macos")] // primary selection does not exist on macos CopyDestination::Primary => "Text copied to system clipboard", CopyDestination::System => "Text copied to system clipboard", }; LinePart { part: serialize_text(&Text::new(&hint).color_range(2, ..).opaque()), len: hint.len(), tab_index: None, } } pub fn system_clipboard_error() -> LinePart { let hint = " Error using the system clipboard."; LinePart { part: serialize_text(&Text::new(&hint).color_range(2, ..).opaque()), len: hint.len(), tab_index: None, } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/compact-bar/src/main.rs
default-plugins/compact-bar/src/main.rs
mod action_types; mod clipboard_utils; mod keybind_utils; mod line; mod tab; mod tooltip; use std::cmp::{max, min}; use std::collections::BTreeMap; use std::convert::TryInto; use tab::get_tab_to_focus; use zellij_tile::prelude::*; use crate::clipboard_utils::{system_clipboard_error, text_copied_hint}; use crate::line::tab_line; use crate::tab::tab_style; use crate::tooltip::TooltipRenderer; static ARROW_SEPARATOR: &str = ""; const CONFIG_IS_TOOLTIP: &str = "is_tooltip"; const CONFIG_TOGGLE_TOOLTIP_KEY: &str = "tooltip"; const MSG_TOGGLE_TOOLTIP: &str = "toggle_tooltip"; const MSG_TOGGLE_PERSISTED_TOOLTIP: &str = "toggle_persisted_tooltip"; const MSG_LAUNCH_TOOLTIP: &str = "launch_tooltip_if_not_launched"; #[derive(Debug, Default)] pub struct LinePart { part: String, len: usize, tab_index: Option<usize>, } #[derive(Default)] struct State { // Tab state tabs: Vec<TabInfo>, active_tab_idx: usize, // Display state mode_info: ModeInfo, tab_line: Vec<LinePart>, display_area_rows: usize, display_area_cols: usize, // Clipboard state text_copy_destination: Option<CopyDestination>, display_system_clipboard_failure: bool, // Plugin configuration config: BTreeMap<String, String>, own_plugin_id: Option<u32>, toggle_tooltip_key: Option<String>, // Tooltip state is_tooltip: bool, tooltip_is_active: bool, persist: bool, is_first_run: bool, own_tab_index: Option<usize>, own_client_id: u16, } struct TabRenderData { tabs: Vec<LinePart>, active_tab_index: usize, active_swap_layout_name: Option<String>, is_swap_layout_dirty: bool, } register_plugin!(State); impl ZellijPlugin for State { fn load(&mut self, configuration: BTreeMap<String, String>) { let plugin_ids = get_plugin_ids(); self.own_plugin_id = Some(plugin_ids.plugin_id); self.own_client_id = plugin_ids.client_id; self.initialize_configuration(configuration); self.setup_subscriptions(); self.configure_keybinds(); } fn update(&mut self, event: Event) -> bool { self.is_first_run = false; match event { Event::ModeUpdate(mode_info) => self.handle_mode_update(mode_info), Event::TabUpdate(tabs) => self.handle_tab_update(tabs), Event::PaneUpdate(pane_manifest) => self.handle_pane_update(pane_manifest), Event::Mouse(mouse_event) => { self.handle_mouse_event(mouse_event); false }, Event::CopyToClipboard(copy_destination) => { self.handle_clipboard_copy(copy_destination) }, Event::SystemClipboardFailure => self.handle_clipboard_failure(), Event::InputReceived => self.handle_input_received(), _ => false, } } fn pipe(&mut self, message: PipeMessage) -> bool { if self.is_tooltip && message.is_private { self.handle_tooltip_pipe(message); } else if message.name == MSG_TOGGLE_TOOLTIP && message.is_private && self.toggle_tooltip_key.is_some() // only launch once per plugin instance && self.own_tab_index == Some(self.active_tab_idx.saturating_sub(1)) // only launch once per client of plugin instance && Some(format!("{}", self.own_client_id)) == message.payload { self.toggle_persisted_tooltip(self.mode_info.mode); } false } fn render(&mut self, rows: usize, cols: usize) { if self.is_tooltip { self.render_tooltip(rows, cols); } else { self.render_tab_line(cols); } } } impl State { fn initialize_configuration(&mut self, configuration: BTreeMap<String, String>) { self.config = configuration.clone(); self.is_tooltip = self.parse_bool_config(CONFIG_IS_TOOLTIP, false); if !self.is_tooltip { if let Some(tooltip_toggle_key) = configuration.get(CONFIG_TOGGLE_TOOLTIP_KEY) { self.toggle_tooltip_key = Some(tooltip_toggle_key.clone()); } } if self.is_tooltip { self.is_first_run = true; } } fn setup_subscriptions(&self) { set_selectable(false); let events = if self.is_tooltip { vec![EventType::ModeUpdate, EventType::TabUpdate] } else { vec![ EventType::TabUpdate, EventType::PaneUpdate, EventType::ModeUpdate, EventType::Mouse, EventType::CopyToClipboard, EventType::InputReceived, EventType::SystemClipboardFailure, ] }; subscribe(&events); } fn configure_keybinds(&self) { if !self.is_tooltip && self.toggle_tooltip_key.is_some() { if let Some(toggle_key) = &self.toggle_tooltip_key { reconfigure( bind_toggle_key_config(toggle_key, self.own_client_id), false, ); } } } fn parse_bool_config(&self, key: &str, default: bool) -> bool { self.config .get(key) .and_then(|v| v.parse().ok()) .unwrap_or(default) } // Event handlers fn handle_mode_update(&mut self, mode_info: ModeInfo) -> bool { let should_render = self.mode_info != mode_info; let old_mode = self.mode_info.mode; let new_mode = mode_info.mode; let base_mode = mode_info.base_mode.unwrap_or(InputMode::Normal); self.mode_info = mode_info; if self.is_tooltip { self.handle_tooltip_mode_update(old_mode, new_mode, base_mode); } else { self.handle_main_mode_update(new_mode, base_mode); } should_render } fn handle_main_mode_update(&self, new_mode: InputMode, base_mode: InputMode) { if self.toggle_tooltip_key.is_some() && new_mode != base_mode && !self.is_restricted_mode(new_mode) { self.launch_tooltip_if_not_launched(new_mode); } } fn handle_tooltip_mode_update( &mut self, old_mode: InputMode, new_mode: InputMode, base_mode: InputMode, ) { if !self.persist && (new_mode == base_mode || self.is_restricted_mode(new_mode)) { close_self(); } else if new_mode != old_mode || self.persist { self.update_tooltip_for_mode_change(new_mode); } } fn handle_tab_update(&mut self, tabs: Vec<TabInfo>) -> bool { self.update_display_area(&tabs); if let Some(active_tab_index) = tabs.iter().position(|t| t.active) { let active_tab_idx = active_tab_index + 1; // Convert to 1-based indexing let should_render = self.active_tab_idx != active_tab_idx || self.tabs != tabs; if self.is_tooltip && self.active_tab_idx != active_tab_idx { self.move_tooltip_to_new_tab(active_tab_idx); } self.active_tab_idx = active_tab_idx; self.tabs = tabs; should_render } else { false } } fn handle_pane_update(&mut self, pane_manifest: PaneManifest) -> bool { if self.toggle_tooltip_key.is_some() { let previous_tooltip_state = self.tooltip_is_active; self.tooltip_is_active = self.detect_tooltip_presence(&pane_manifest); self.own_tab_index = self.find_own_tab_index(&pane_manifest); previous_tooltip_state != self.tooltip_is_active } else { false } } fn handle_mouse_event(&mut self, mouse_event: Mouse) { if self.is_tooltip { return; } match mouse_event { Mouse::LeftClick(_, col) => self.handle_tab_click(col), Mouse::ScrollUp(_) => self.scroll_tab_up(), Mouse::ScrollDown(_) => self.scroll_tab_down(), _ => {}, } } fn handle_clipboard_copy(&mut self, copy_destination: CopyDestination) -> bool { if self.is_tooltip { return false; } let should_render = match self.text_copy_destination { Some(current) => current != copy_destination, None => true, }; self.text_copy_destination = Some(copy_destination); should_render } fn handle_clipboard_failure(&mut self) -> bool { if self.is_tooltip { return false; } self.display_system_clipboard_failure = true; true } fn handle_input_received(&mut self) -> bool { if self.is_tooltip { return false; } let should_render = self.text_copy_destination.is_some() || self.display_system_clipboard_failure; self.clear_clipboard_state(); should_render } fn handle_tooltip_pipe(&mut self, message: PipeMessage) { if message.name == MSG_TOGGLE_PERSISTED_TOOLTIP { if self.is_first_run { self.persist = true; } else { #[cfg(target_family = "wasm")] close_self(); } } } // Helper methods fn update_display_area(&mut self, tabs: &[TabInfo]) { for tab in tabs { if tab.active { self.display_area_rows = tab.display_area_rows; self.display_area_cols = tab.display_area_columns; break; } } } fn detect_tooltip_presence(&self, pane_manifest: &PaneManifest) -> bool { for (_tab_index, panes) in &pane_manifest.panes { for pane in panes { if pane.plugin_url == Some("zellij:compact-bar".to_owned()) && pane.pane_x != pane.pane_content_x { return true; } } } false } fn find_own_tab_index(&self, pane_manifest: &PaneManifest) -> Option<usize> { for (tab_index, panes) in &pane_manifest.panes { for pane in panes { if pane.is_plugin && Some(pane.id) == self.own_plugin_id { return Some(*tab_index); } } } None } fn handle_tab_click(&self, col: usize) { if let Some(tab_idx) = get_tab_to_focus(&self.tab_line, self.active_tab_idx, col) { switch_tab_to(tab_idx.try_into().unwrap()); } } fn scroll_tab_up(&self) { let next_tab = min(self.active_tab_idx + 1, self.tabs.len()); switch_tab_to(next_tab as u32); } fn scroll_tab_down(&self) { let prev_tab = max(self.active_tab_idx.saturating_sub(1), 1); switch_tab_to(prev_tab as u32); } fn clear_clipboard_state(&mut self) { self.text_copy_destination = None; self.display_system_clipboard_failure = false; } fn is_restricted_mode(&self, mode: InputMode) -> bool { matches!( mode, InputMode::Locked | InputMode::EnterSearch | InputMode::RenameTab | InputMode::RenamePane | InputMode::Prompt | InputMode::Tmux ) } // Tooltip operations fn toggle_persisted_tooltip(&self, new_mode: InputMode) { #[allow(unused_variables)] let message = self .create_tooltip_message(MSG_TOGGLE_PERSISTED_TOOLTIP, new_mode) .with_args(self.create_persist_args()); #[cfg(target_family = "wasm")] pipe_message_to_plugin(message); } fn launch_tooltip_if_not_launched(&self, new_mode: InputMode) { let message = self.create_tooltip_message(MSG_LAUNCH_TOOLTIP, new_mode); pipe_message_to_plugin(message); } fn create_tooltip_message(&self, name: &str, mode: InputMode) -> MessageToPlugin { let mut tooltip_config = self.config.clone(); tooltip_config.insert(CONFIG_IS_TOOLTIP.to_string(), "true".to_string()); MessageToPlugin::new(name) .with_plugin_url("zellij:OWN_URL") .with_plugin_config(tooltip_config) .with_floating_pane_coordinates(self.calculate_tooltip_coordinates()) .new_plugin_instance_should_have_pane_title(format!("{:?}", mode)) } fn create_persist_args(&self) -> BTreeMap<String, String> { let mut args = BTreeMap::new(); args.insert("persist".to_string(), String::new()); args } fn update_tooltip_for_mode_change(&self, new_mode: InputMode) { if let Some(plugin_id) = self.own_plugin_id { let coordinates = self.calculate_tooltip_coordinates(); change_floating_panes_coordinates(vec![(PaneId::Plugin(plugin_id), coordinates)]); rename_plugin_pane(plugin_id, format!("{:?}", new_mode)); } } fn move_tooltip_to_new_tab(&self, new_tab_index: usize) { if let Some(plugin_id) = self.own_plugin_id { break_panes_to_tab_with_index( &[PaneId::Plugin(plugin_id)], new_tab_index.saturating_sub(1), // Convert to 0-based indexing false, ); } } fn calculate_tooltip_coordinates(&self) -> FloatingPaneCoordinates { let tooltip_renderer = TooltipRenderer::new(&self.mode_info); let (tooltip_rows, tooltip_cols) = tooltip_renderer.calculate_dimensions(self.mode_info.mode); let width = tooltip_cols + 4; // 2 for borders, 2 for padding let height = tooltip_rows + 2; // 2 for borders let x_position = 2; let y_position = self.display_area_rows.saturating_sub(height + 2); FloatingPaneCoordinates::new( Some(x_position.to_string()), Some(y_position.to_string()), Some(width.to_string()), Some(height.to_string()), Some(true), ) .unwrap_or_default() } // Rendering fn render_tooltip(&self, rows: usize, cols: usize) { let tooltip_renderer = TooltipRenderer::new(&self.mode_info); tooltip_renderer.render(rows, cols); } fn render_tab_line(&mut self, cols: usize) { if let Some(copy_destination) = self.text_copy_destination { self.render_clipboard_hint(copy_destination); } else if self.display_system_clipboard_failure { self.render_clipboard_error(); } else { self.render_tabs(cols); } } fn render_clipboard_hint(&self, copy_destination: CopyDestination) { let hint = text_copied_hint(copy_destination).part; self.render_background_with_text(&hint); } fn render_clipboard_error(&self) { let hint = system_clipboard_error().part; self.render_background_with_text(&hint); } fn render_background_with_text(&self, text: &str) { let background = self.mode_info.style.colors.text_unselected.background; match background { PaletteColor::Rgb((r, g, b)) => { print!("{}\u{1b}[48;2;{};{};{}m\u{1b}[0K", text, r, g, b); }, PaletteColor::EightBit(color) => { print!("{}\u{1b}[48;5;{}m\u{1b}[0K", text, color); }, } } fn render_tabs(&mut self, cols: usize) { if self.tabs.is_empty() { return; } let tab_data = self.prepare_tab_data(); self.tab_line = tab_line( &self.mode_info, tab_data, cols, self.toggle_tooltip_key.clone(), self.tooltip_is_active, ); let output = self .tab_line .iter() .fold(String::new(), |acc, part| acc + &part.part); self.render_background_with_text(&output); } fn prepare_tab_data(&self) -> TabRenderData { let mut all_tabs = Vec::new(); let mut active_tab_index = 0; let mut active_swap_layout_name = None; let mut is_swap_layout_dirty = false; let mut is_alternate_tab = false; for tab in &self.tabs { let tab_name = self.get_tab_display_name(tab); if tab.active { active_tab_index = tab.position; if self.mode_info.mode != InputMode::RenameTab { is_swap_layout_dirty = tab.is_swap_layout_dirty; active_swap_layout_name = tab.active_swap_layout_name.clone(); } } let styled_tab = tab_style( tab_name, tab, is_alternate_tab, self.mode_info.style.colors, self.mode_info.capabilities, ); is_alternate_tab = !is_alternate_tab; all_tabs.push(styled_tab); } TabRenderData { tabs: all_tabs, active_tab_index, active_swap_layout_name, is_swap_layout_dirty, } } fn get_tab_display_name(&self, tab: &TabInfo) -> String { let mut tab_name = tab.name.clone(); if tab.active && self.mode_info.mode == InputMode::RenameTab && tab_name.is_empty() { tab_name = "Enter name...".to_string(); } tab_name } } fn bind_toggle_key_config(toggle_key: &str, client_id: u16) -> String { format!( r#" keybinds {{ shared {{ bind "{}" {{ MessagePlugin "compact-bar" {{ name "toggle_tooltip" tooltip "{}" payload "{}" }} }} }} }} "#, toggle_key, toggle_key, client_id ) }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/compact-bar/src/line.rs
default-plugins/compact-bar/src/line.rs
use ansi_term::ANSIStrings; use unicode_width::UnicodeWidthStr; use crate::{LinePart, TabRenderData, ARROW_SEPARATOR}; use zellij_tile::prelude::*; use zellij_tile_utils::style; pub fn tab_line( mode_info: &ModeInfo, tab_data: TabRenderData, cols: usize, toggle_tooltip_key: Option<String>, tooltip_is_active: bool, ) -> Vec<LinePart> { let config = TabLineConfig { session_name: mode_info.session_name.to_owned(), hide_session_name: mode_info.style.hide_session_name, mode: mode_info.mode, active_swap_layout_name: tab_data.active_swap_layout_name, is_swap_layout_dirty: tab_data.is_swap_layout_dirty, toggle_tooltip_key, tooltip_is_active, }; let builder = TabLineBuilder::new(config, mode_info.style.colors, mode_info.capabilities, cols); builder.build(tab_data.tabs, tab_data.active_tab_index) } #[derive(Debug, Clone)] pub struct TabLineConfig { pub session_name: Option<String>, pub hide_session_name: bool, pub mode: InputMode, pub active_swap_layout_name: Option<String>, pub is_swap_layout_dirty: bool, pub toggle_tooltip_key: Option<String>, pub tooltip_is_active: bool, } fn calculate_total_length(parts: &[LinePart]) -> usize { parts.iter().map(|p| p.len).sum() } struct TabLinePopulator { cols: usize, palette: Styling, capabilities: PluginCapabilities, } impl TabLinePopulator { fn new(cols: usize, palette: Styling, capabilities: PluginCapabilities) -> Self { Self { cols, palette, capabilities, } } fn populate_tabs( &self, tabs_before_active: &mut Vec<LinePart>, tabs_after_active: &mut Vec<LinePart>, tabs_to_render: &mut Vec<LinePart>, ) { let mut middle_size = calculate_total_length(tabs_to_render); let mut total_left = 0; let mut total_right = 0; loop { let left_count = tabs_before_active.len(); let right_count = tabs_after_active.len(); let collapsed_indicators = self.create_collapsed_indicators(left_count, right_count, tabs_to_render.len()); let total_size = collapsed_indicators.left.len + middle_size + collapsed_indicators.right.len; if total_size > self.cols { break; } let tab_sizes = TabSizes { left: tabs_before_active.last().map_or(usize::MAX, |tab| tab.len), right: tabs_after_active.get(0).map_or(usize::MAX, |tab| tab.len), }; let fit_analysis = self.analyze_tab_fit( &tab_sizes, total_size, left_count, right_count, &collapsed_indicators, ); match self.decide_next_action(&fit_analysis, total_left, total_right) { TabAction::AddLeft => { if let Some(tab) = tabs_before_active.pop() { middle_size += tab.len; total_left += tab.len; tabs_to_render.insert(0, tab); } }, TabAction::AddRight => { if !tabs_after_active.is_empty() { let tab = tabs_after_active.remove(0); middle_size += tab.len; total_right += tab.len; tabs_to_render.push(tab); } }, TabAction::Finish => { tabs_to_render.insert(0, collapsed_indicators.left); tabs_to_render.push(collapsed_indicators.right); break; }, } } } fn create_collapsed_indicators( &self, left_count: usize, right_count: usize, rendered_count: usize, ) -> CollapsedIndicators { let left_more_tab_index = left_count.saturating_sub(1); let right_more_tab_index = left_count + rendered_count; CollapsedIndicators { left: self.create_left_indicator(left_count, left_more_tab_index), right: self.create_right_indicator(right_count, right_more_tab_index), } } fn analyze_tab_fit( &self, tab_sizes: &TabSizes, total_size: usize, left_count: usize, right_count: usize, collapsed_indicators: &CollapsedIndicators, ) -> TabFitAnalysis { let size_by_adding_left = tab_sizes .left .saturating_add(total_size) .saturating_sub(if left_count == 1 { collapsed_indicators.left.len } else { 0 }); let size_by_adding_right = tab_sizes .right .saturating_add(total_size) .saturating_sub(if right_count == 1 { collapsed_indicators.right.len } else { 0 }); TabFitAnalysis { left_fits: size_by_adding_left <= self.cols, right_fits: size_by_adding_right <= self.cols, } } fn decide_next_action( &self, fit_analysis: &TabFitAnalysis, total_left: usize, total_right: usize, ) -> TabAction { if (total_left <= total_right || !fit_analysis.right_fits) && fit_analysis.left_fits { TabAction::AddLeft } else if fit_analysis.right_fits { TabAction::AddRight } else { TabAction::Finish } } fn create_left_indicator(&self, tab_count: usize, tab_index: usize) -> LinePart { if tab_count == 0 { return LinePart::default(); } let more_text = self.format_count_text(tab_count, "← +{}", " ← +many "); self.create_styled_indicator(more_text, tab_index) } fn create_right_indicator(&self, tab_count: usize, tab_index: usize) -> LinePart { if tab_count == 0 { return LinePart::default(); } let more_text = self.format_count_text(tab_count, "+{} →", " +many → "); self.create_styled_indicator(more_text, tab_index) } fn format_count_text(&self, count: usize, format_str: &str, fallback: &str) -> String { if count < 10000 { format!(" {} ", format_str.replace("{}", &count.to_string())) } else { fallback.to_string() } } fn create_styled_indicator(&self, text: String, tab_index: usize) -> LinePart { let separator = tab_separator(self.capabilities); let text_len = text.width() + 2 * separator.width(); let colors = IndicatorColors { text: self.palette.ribbon_unselected.base, separator: self.palette.text_unselected.background, background: self.palette.text_selected.emphasis_0, }; let styled_parts = [ style!(colors.separator, colors.background).paint(separator), style!(colors.text, colors.background).bold().paint(text), style!(colors.background, colors.separator).paint(separator), ]; LinePart { part: ANSIStrings(&styled_parts).to_string(), len: text_len, tab_index: Some(tab_index), } } } #[derive(Debug)] struct CollapsedIndicators { left: LinePart, right: LinePart, } #[derive(Debug)] struct TabSizes { left: usize, right: usize, } #[derive(Debug)] struct TabFitAnalysis { left_fits: bool, right_fits: bool, } #[derive(Debug)] struct IndicatorColors { text: PaletteColor, separator: PaletteColor, background: PaletteColor, } #[derive(Debug)] enum TabAction { AddLeft, AddRight, Finish, } struct TabLinePrefixBuilder { palette: Styling, cols: usize, } impl TabLinePrefixBuilder { fn new(palette: Styling, cols: usize) -> Self { Self { palette, cols } } fn build(&self, session_name: Option<&str>, mode: InputMode) -> Vec<LinePart> { let mut parts = vec![self.create_zellij_part()]; let mut used_len = parts.get(0).map_or(0, |p| p.len); if let Some(name) = session_name { if let Some(name_part) = self.create_session_name_part(name, used_len) { used_len += name_part.len; parts.push(name_part); } } if let Some(mode_part) = self.create_mode_part(mode, used_len) { parts.push(mode_part); } parts } fn create_zellij_part(&self) -> LinePart { let prefix_text = " Zellij "; let colors = self.get_text_colors(); LinePart { part: style!(colors.text, colors.background) .bold() .paint(prefix_text) .to_string(), len: prefix_text.chars().count(), tab_index: None, } } fn create_session_name_part(&self, name: &str, used_len: usize) -> Option<LinePart> { let name_part = format!("({})", name); let name_part_len = name_part.width(); if self.cols.saturating_sub(used_len) >= name_part_len { let colors = self.get_text_colors(); Some(LinePart { part: style!(colors.text, colors.background) .bold() .paint(name_part) .to_string(), len: name_part_len, tab_index: None, }) } else { None } } fn create_mode_part(&self, mode: InputMode, used_len: usize) -> Option<LinePart> { let mode_text = format!(" {} ", format!("{:?}", mode).to_uppercase()); let mode_len = mode_text.width(); if self.cols.saturating_sub(used_len) >= mode_len { let colors = self.get_text_colors(); let style = match mode { InputMode::Locked => { style!(self.palette.text_unselected.emphasis_3, colors.background) }, InputMode::Normal => { style!(self.palette.text_unselected.emphasis_2, colors.background) }, _ => style!(self.palette.text_unselected.emphasis_0, colors.background), }; Some(LinePart { part: style.bold().paint(mode_text).to_string(), len: mode_len, tab_index: None, }) } else { None } } fn get_text_colors(&self) -> IndicatorColors { IndicatorColors { text: self.palette.text_unselected.base, background: self.palette.text_unselected.background, separator: self.palette.text_unselected.background, } } } struct RightSideElementsBuilder { palette: Styling, capabilities: PluginCapabilities, } impl RightSideElementsBuilder { fn new(palette: Styling, capabilities: PluginCapabilities) -> Self { Self { palette, capabilities, } } fn build(&self, config: &TabLineConfig, available_space: usize) -> Vec<LinePart> { let mut elements = Vec::new(); if let Some(ref tooltip_key) = config.toggle_tooltip_key { elements.push(self.create_tooltip_indicator(tooltip_key, config.tooltip_is_active)); } if let Some(swap_status) = self.create_swap_layout_status(config, available_space) { elements.push(swap_status); } elements } fn create_tooltip_indicator(&self, toggle_key: &str, is_active: bool) -> LinePart { let key_text = toggle_key; let key = Text::new(key_text).color_all(3).opaque(); let ribbon_text = "Tooltip"; let mut ribbon = Text::new(ribbon_text); if is_active { ribbon = ribbon.selected(); } LinePart { part: format!("{} {}", serialize_text(&key), serialize_ribbon(&ribbon)), len: key_text.chars().count() + ribbon_text.chars().count() + 6, tab_index: None, } } fn create_swap_layout_status( &self, config: &TabLineConfig, max_len: usize, ) -> Option<LinePart> { let swap_layout_name = config.active_swap_layout_name.as_ref()?; let mut layout_name = format!(" {} ", swap_layout_name); layout_name.make_ascii_uppercase(); let layout_name_len = layout_name.len() + 3; let colors = SwapLayoutColors { bg: self.palette.text_unselected.background, fg: self.palette.ribbon_unselected.background, green: self.palette.ribbon_selected.background, }; let separator = tab_separator(self.capabilities); let styled_parts = self.create_swap_layout_styled_parts( &layout_name, config.mode, config.is_swap_layout_dirty, &colors, separator, ); let indicator = format!("{}{}{}", styled_parts.0, styled_parts.1, styled_parts.2); let (part, full_len) = (indicator.clone(), layout_name_len); let short_len = layout_name_len + 1; if full_len <= max_len { Some(LinePart { part, len: full_len, tab_index: None, }) } else if short_len <= max_len && config.mode != InputMode::Locked { Some(LinePart { part: indicator, len: short_len, tab_index: None, }) } else { None } } fn create_swap_layout_styled_parts( &self, layout_name: &str, mode: InputMode, is_dirty: bool, colors: &SwapLayoutColors, separator: &str, ) -> (String, String, String) { match mode { InputMode::Locked => ( style!(colors.bg, colors.fg).paint(separator).to_string(), style!(colors.bg, colors.fg) .italic() .paint(layout_name) .to_string(), style!(colors.fg, colors.bg).paint(separator).to_string(), ), _ if is_dirty => ( style!(colors.bg, colors.fg).paint(separator).to_string(), style!(colors.bg, colors.fg) .bold() .paint(layout_name) .to_string(), style!(colors.fg, colors.bg).paint(separator).to_string(), ), _ => ( style!(colors.bg, colors.green).paint(separator).to_string(), style!(colors.bg, colors.green) .bold() .paint(layout_name) .to_string(), style!(colors.green, colors.bg).paint(separator).to_string(), ), } } } #[derive(Debug)] struct SwapLayoutColors { bg: PaletteColor, fg: PaletteColor, green: PaletteColor, } pub struct TabLineBuilder { config: TabLineConfig, palette: Styling, capabilities: PluginCapabilities, cols: usize, } impl TabLineBuilder { pub fn new( config: TabLineConfig, palette: Styling, capabilities: PluginCapabilities, cols: usize, ) -> Self { Self { config, palette, capabilities, cols, } } pub fn build(self, all_tabs: Vec<LinePart>, active_tab_index: usize) -> Vec<LinePart> { let (tabs_before_active, active_tab, tabs_after_active) = self.split_tabs(all_tabs, active_tab_index); let prefix_builder = TabLinePrefixBuilder::new(self.palette, self.cols); let session_name = if self.config.hide_session_name { None } else { self.config.session_name.as_deref() }; let mut prefix = prefix_builder.build(session_name, self.config.mode); let prefix_len = calculate_total_length(&prefix); if prefix_len + active_tab.len > self.cols { return prefix; } let mut tabs_to_render = vec![active_tab]; let populator = TabLinePopulator::new( self.cols.saturating_sub(prefix_len), self.palette, self.capabilities, ); let mut tabs_before = tabs_before_active; let mut tabs_after = tabs_after_active; populator.populate_tabs(&mut tabs_before, &mut tabs_after, &mut tabs_to_render); prefix.append(&mut tabs_to_render); self.add_right_side_elements(&mut prefix); prefix } fn split_tabs( &self, mut all_tabs: Vec<LinePart>, active_tab_index: usize, ) -> (Vec<LinePart>, LinePart, Vec<LinePart>) { let mut tabs_after_active = all_tabs.split_off(active_tab_index); let mut tabs_before_active = all_tabs; let active_tab = if !tabs_after_active.is_empty() { tabs_after_active.remove(0) } else { tabs_before_active.pop().unwrap_or_default() }; (tabs_before_active, active_tab, tabs_after_active) } fn add_right_side_elements(&self, prefix: &mut Vec<LinePart>) { let current_len = calculate_total_length(prefix); if current_len < self.cols { let right_builder = RightSideElementsBuilder::new(self.palette, self.capabilities); let available_space = self.cols.saturating_sub(current_len); let mut right_elements = right_builder.build(&self.config, available_space); let right_len = calculate_total_length(&right_elements); if current_len + right_len <= self.cols { let remaining_space = self .cols .saturating_sub(current_len) .saturating_sub(right_len); if remaining_space > 0 { prefix.push(self.create_spacer(remaining_space)); } prefix.append(&mut right_elements); } } } fn create_spacer(&self, space: usize) -> LinePart { let bg = self.palette.text_unselected.background; let buffer = (0..space) .map(|_| style!(bg, bg).paint(" ").to_string()) .collect::<String>(); LinePart { part: buffer, len: space, tab_index: None, } } } pub fn tab_separator(capabilities: PluginCapabilities) -> &'static str { if !capabilities.arrow_fonts { ARROW_SEPARATOR } else { "" } }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
zellij-org/zellij
https://github.com/zellij-org/zellij/blob/3fe48a972c55537502128779116d38d8f8aedb7e/default-plugins/compact-bar/src/tab.rs
default-plugins/compact-bar/src/tab.rs
use crate::{line::tab_separator, LinePart}; use ansi_term::{ANSIString, ANSIStrings}; use unicode_width::UnicodeWidthStr; use zellij_tile::prelude::*; use zellij_tile_utils::style; fn cursors<'a>( focused_clients: &'a [ClientId], colors: MultiplayerColors, ) -> (Vec<ANSIString<'a>>, usize) { // cursor section, text length let mut len = 0; let mut cursors = vec![]; for client_id in focused_clients.iter() { if let Some(color) = client_id_to_colors(*client_id, colors) { cursors.push(style!(color.1, color.0).paint(" ")); len += 1; } } len += 2; // 2 for the brackets: [ and ] (cursors, len) } pub fn render_tab( text: String, tab: &TabInfo, is_alternate_tab: bool, palette: Styling, separator: &str, ) -> LinePart { let focused_clients = tab.other_focused_clients.as_slice(); let separator_width = separator.width(); let alternate_tab_color = if is_alternate_tab { palette.ribbon_unselected.emphasis_1 } else { palette.ribbon_unselected.background }; let background_color = if tab.active { palette.ribbon_selected.background } else if is_alternate_tab { alternate_tab_color } else { palette.ribbon_unselected.background }; let foreground_color = if tab.active { palette.ribbon_selected.base } else { palette.ribbon_unselected.base }; let separator_fill_color = palette.text_unselected.background; let left_separator = style!(separator_fill_color, background_color).paint(separator); let mut tab_text_len = text.width() + (separator_width * 2) + 2; // + 2 for padding let tab_styled_text = style!(foreground_color, background_color) .bold() .paint(format!(" {} ", text)); let right_separator = style!(background_color, separator_fill_color).paint(separator); let tab_styled_text = if !focused_clients.is_empty() { let (cursor_section, extra_length) = cursors(focused_clients, palette.multiplayer_user_colors); tab_text_len += extra_length; let mut s = String::new(); let cursor_beginning = style!(foreground_color, background_color) .bold() .paint("[") .to_string(); let cursor_section = ANSIStrings(&cursor_section).to_string(); let cursor_end = style!(foreground_color, background_color) .bold() .paint("]") .to_string(); s.push_str(&left_separator.to_string()); s.push_str(&tab_styled_text.to_string()); s.push_str(&cursor_beginning); s.push_str(&cursor_section); s.push_str(&cursor_end); s.push_str(&right_separator.to_string()); s } else { ANSIStrings(&[left_separator, tab_styled_text, right_separator]).to_string() }; LinePart { part: tab_styled_text, len: tab_text_len, tab_index: Some(tab.position), } } pub fn tab_style( mut tabname: String, tab: &TabInfo, mut is_alternate_tab: bool, palette: Styling, capabilities: PluginCapabilities, ) -> LinePart { let separator = tab_separator(capabilities); if tab.is_fullscreen_active { tabname.push_str(" (FULLSCREEN)"); } else if tab.is_sync_panes_active { tabname.push_str(" (SYNC)"); } // we only color alternate tabs differently if we can't use the arrow fonts to separate them if !capabilities.arrow_fonts { is_alternate_tab = false; } render_tab(tabname, tab, is_alternate_tab, palette, separator) } pub(crate) fn get_tab_to_focus( tab_line: &[LinePart], active_tab_idx: usize, mouse_click_col: usize, ) -> Option<usize> { let clicked_line_part = get_clicked_line_part(tab_line, mouse_click_col)?; let clicked_tab_idx = clicked_line_part.tab_index?; // tabs are indexed starting from 1 so we need to add 1 let clicked_tab_idx = clicked_tab_idx + 1; if clicked_tab_idx != active_tab_idx { return Some(clicked_tab_idx); } None } pub(crate) fn get_clicked_line_part( tab_line: &[LinePart], mouse_click_col: usize, ) -> Option<&LinePart> { let mut len = 0; for tab_line_part in tab_line { if mouse_click_col >= len && mouse_click_col < len + tab_line_part.len { return Some(tab_line_part); } len += tab_line_part.len; } None }
rust
MIT
3fe48a972c55537502128779116d38d8f8aedb7e
2026-01-04T15:35:12.838106Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/settings.rs
src/settings.rs
use std::borrow::Cow; use std::{env, io}; use api::grpc::transport_channel_pool::{ DEFAULT_CONNECT_TIMEOUT, DEFAULT_GRPC_TIMEOUT, DEFAULT_POOL_SIZE, }; use collection::operations::validation; use collection::shards::shard::PeerId; use common::flags::FeatureFlags; use config::{Config, ConfigError, Environment, File, FileFormat, Source}; use serde::Deserialize; use storage::types::StorageConfig; use validator::{Validate, ValidationError}; use crate::common::debugger::DebuggerConfig; use crate::common::inference::config::InferenceConfig; use crate::tracing; const MAX_PEER_ID: u64 = (1 << 53) - 1; const DEFAULT_CONFIG: &str = include_str!("../config/config.yaml"); #[derive(Debug, Deserialize, Validate, Clone)] pub struct ServiceConfig { #[validate(length(min = 1))] pub host: String, pub http_port: u16, pub grpc_port: Option<u16>, // None means that gRPC is disabled pub max_request_size_mb: usize, pub max_workers: Option<usize>, #[serde(default = "default_cors")] pub enable_cors: bool, #[serde(default)] pub enable_tls: bool, #[serde(default)] pub verify_https_client_certificate: bool, pub api_key: Option<String>, pub read_only_api_key: Option<String>, #[serde(default)] pub jwt_rbac: Option<bool>, #[serde(default)] pub hide_jwt_dashboard: Option<bool>, /// Directory where static files are served from. /// For example, the Web-UI should be placed here. #[serde(default)] pub static_content_dir: Option<String>, /// If serving of the static content is enabled. /// This includes the Web-UI. True by default. #[serde(default)] pub enable_static_content: Option<bool>, /// How much time is considered too long for a query to execute. pub slow_query_secs: Option<f32>, /// Whether to enable reporting of measured hardware utilization in API responses. #[serde(default)] pub hardware_reporting: Option<bool>, /// Global prefix for metrics. #[serde(default)] #[validate(custom(function = validate_metrics_prefix))] pub metrics_prefix: Option<String>, } impl ServiceConfig { pub fn hardware_reporting(&self) -> bool { self.hardware_reporting.unwrap_or_default() } } #[derive(Debug, Deserialize, Clone, Default, Validate)] pub struct ClusterConfig { pub enabled: bool, // disabled by default #[serde(default)] #[validate(range(min = 1, max = MAX_PEER_ID))] pub peer_id: Option<PeerId>, #[serde(default = "default_timeout_ms")] #[validate(range(min = 1))] pub grpc_timeout_ms: u64, #[serde(default = "default_connection_timeout_ms")] #[validate(range(min = 1))] pub connection_timeout_ms: u64, #[serde(default)] #[validate(nested)] pub p2p: P2pConfig, #[serde(default)] #[validate(nested)] pub consensus: ConsensusConfig, #[serde(default)] pub resharding_enabled: bool, // disabled by default } #[derive(Debug, Deserialize, Clone, Validate)] pub struct P2pConfig { #[serde(default)] pub port: Option<u16>, #[serde(default = "default_connection_pool_size")] #[validate(range(min = 1))] pub connection_pool_size: usize, #[serde(default)] pub enable_tls: bool, } impl Default for P2pConfig { fn default() -> Self { P2pConfig { port: None, connection_pool_size: default_connection_pool_size(), enable_tls: false, } } } #[derive(Debug, Deserialize, Clone, Validate)] pub struct ConsensusConfig { #[serde(default = "default_max_message_queue_size")] pub max_message_queue_size: usize, // controls the back-pressure at the Raft level #[serde(default = "default_tick_period_ms")] #[validate(range(min = 1))] pub tick_period_ms: u64, #[serde(default = "default_bootstrap_timeout_sec")] #[validate(range(min = 1))] pub bootstrap_timeout_sec: u64, #[validate(range(min = 1))] #[serde(default = "default_message_timeout_tics")] pub message_timeout_ticks: u64, /// Compact WAL when it grows to enough applied entries #[serde(default = "default_compact_wal_entries")] pub compact_wal_entries: u64, } impl Default for ConsensusConfig { fn default() -> Self { ConsensusConfig { max_message_queue_size: default_max_message_queue_size(), tick_period_ms: default_tick_period_ms(), bootstrap_timeout_sec: default_bootstrap_timeout_sec(), message_timeout_ticks: default_message_timeout_tics(), compact_wal_entries: default_compact_wal_entries(), } } } #[derive(Debug, Deserialize, Clone, Validate)] pub struct TlsConfig { pub cert: String, pub key: String, pub ca_cert: Option<String>, #[serde(default = "default_tls_cert_ttl")] #[validate(range(min = 1))] pub cert_ttl: Option<u64>, } #[allow(dead_code)] #[derive(Clone, Debug, Deserialize, Validate)] pub struct GpuConfig { /// Enable GPU indexing. #[serde(default)] pub indexing: bool, /// Force half precision for `f32` values while indexing. /// `f16` conversion will take place only inside GPU memory and won't affect storage type. #[serde(default)] pub force_half_precision: bool, /// Used vulkan "groups" of GPU. In other words, how many parallel points can be indexed by GPU. /// Optimal value might depend on the GPU model. /// Proportional, but doesn't necessary equal to the physical number of warps. /// Do not change this value unless you know what you are doing. /// Default: 512 #[serde(default)] #[validate(range(min = 1))] pub groups_count: Option<usize>, /// Filter for GPU devices by hardware name. Case insensitive. /// Comma-separated list of substrings to match against the gpu device name. /// Example: "nvidia" /// Default: "" - all devices are accepted. #[serde(default)] pub device_filter: String, /// List of explicit GPU devices to use. /// If host has multiple GPUs, this option allows to select specific devices /// by their index in the list of found devices. /// If `device_filter` is set, indexes are applied after filtering. /// By default, all devices are accepted. #[serde(default)] pub devices: Option<Vec<usize>>, /// How many parallel indexing processes are allowed to run. /// Default: 1 #[serde(default)] pub parallel_indexes: Option<usize>, /// Allow to use integrated GPUs. /// Default: false #[serde(default)] pub allow_integrated: bool, /// Allow to use emulated GPUs like LLVMpipe. Useful for CI. /// Default: false #[serde(default)] pub allow_emulated: bool, } #[derive(Debug, Deserialize, Clone, Validate)] pub struct Settings { #[serde(default)] pub log_level: Option<String>, #[serde(default)] pub logger: tracing::LoggerConfig, #[validate(nested)] pub storage: StorageConfig, #[validate(nested)] pub service: ServiceConfig, #[serde(default)] #[validate(nested)] pub cluster: ClusterConfig, #[serde(default = "default_telemetry_disabled")] pub telemetry_disabled: bool, #[validate(nested)] pub tls: Option<TlsConfig>, #[serde(default)] pub debugger: DebuggerConfig, /// A list of messages for errors that happened during loading the configuration. We collect /// them and store them here while loading because then our logger is not configured yet. /// We therefore need to log these messages later, after the logger is ready. #[serde(default, skip)] pub load_errors: Vec<LogMsg>, #[serde(default)] pub inference: Option<InferenceConfig>, #[serde(default)] #[validate(nested)] pub gpu: Option<GpuConfig>, #[serde(default)] pub feature_flags: FeatureFlags, } impl Settings { pub fn new(custom_config_path: Option<String>) -> Result<Self, ConfigError> { let mut load_errors = vec![]; let config_exists = |path| File::with_name(path).collect().is_ok(); // Check if custom config file exists, report error if not if let Some(path) = &custom_config_path && !config_exists(path) { load_errors.push(LogMsg::Error(format!( "Config file via --config-path is not found: {path}" ))); } let env = env::var("RUN_MODE").unwrap_or_else(|_| "development".into()); let config_path_env = format!("config/{env}"); // Report error if main or env config files exist, report warning if not // Check if main and env configuration file load_errors.extend( ["config/config", &config_path_env] .into_iter() .filter(|path| !config_exists(path)) .map(|path| LogMsg::Warn(format!("Config file not found: {path}"))), ); // Configuration builder: define different levels of configuration files let mut config = Config::builder() // Start with compile-time base config .add_source(File::from_str(DEFAULT_CONFIG, FileFormat::Yaml)) // Merge main config: config/config .add_source(File::with_name("config/config").required(false)) // Merge env config: config/{env} // Uses RUN_MODE, defaults to 'development' .add_source(File::with_name(&config_path_env).required(false)) // Merge local config, not tracked in git: config/local .add_source(File::with_name("config/local").required(false)); #[cfg(feature = "deb")] { // Read config, installed with deb package config = config.add_source(File::with_name("/etc/qdrant/config").required(false)); } // Merge user provided config with --config-path if let Some(path) = custom_config_path { config = config.add_source(File::with_name(&path).required(false)); } // Merge environment settings // E.g.: `QDRANT_DEBUG=1 ./target/app` would set `debug=true` config = config.add_source(Environment::with_prefix("QDRANT").separator("__")); // Build and merge config and deserialize into Settings, attach any load errors we had let mut settings: Settings = config.build()?.try_deserialize()?; settings.load_errors.extend(load_errors); Ok(settings) } pub fn tls(&self) -> io::Result<&TlsConfig> { self.tls .as_ref() .ok_or_else(Self::tls_config_is_undefined_error) } pub fn tls_config_is_undefined_error() -> io::Error { io::Error::other("TLS config is not defined in the Qdrant config file") } pub fn validate_and_warn(&self) { // // JWT RBAC // // Using HMAC-SHA256, recommended secret size is 32 bytes const JWT_RECOMMENDED_SECRET_LENGTH: usize = 256 / 8; // Log if JWT RBAC is enabled but no API key is set if self.service.jwt_rbac.unwrap_or_default() { if self.service.api_key.clone().unwrap_or_default().is_empty() { log::warn!("JWT RBAC configured but no API key set, JWT RBAC is not enabled") // Log if JWT RAC is enabled, API key is set but smaller than recommended size for JWT secret } else if self.service.api_key.clone().unwrap_or_default().len() < JWT_RECOMMENDED_SECRET_LENGTH { log::warn!( "It is highly recommended to use an API key of {JWT_RECOMMENDED_SECRET_LENGTH} bytes when JWT RBAC is enabled", ) } } // Print any load error messages we had self.load_errors.iter().for_each(LogMsg::log); if let Err(ref errs) = self.validate() { validation::warn_validation_errors("Settings configuration file", errs); } } } /// Returns the number of maximum actix workers. pub fn max_web_workers(settings: &Settings) -> usize { match settings.service.max_workers { Some(0) => { let num_cpu = common::cpu::get_num_cpus(); std::cmp::max(1, num_cpu - 1) } Some(max_workers) => max_workers, None => settings.storage.performance.max_search_threads, } } #[derive(Clone, Debug)] pub enum LogMsg { Warn(String), Error(String), } impl LogMsg { fn log(&self) { match self { Self::Warn(msg) => log::warn!("{msg}"), Self::Error(msg) => log::error!("{msg}"), } } } const fn default_telemetry_disabled() -> bool { false } const fn default_cors() -> bool { true } const fn default_timeout_ms() -> u64 { DEFAULT_GRPC_TIMEOUT.as_millis() as u64 } const fn default_connection_timeout_ms() -> u64 { DEFAULT_CONNECT_TIMEOUT.as_millis() as u64 } const fn default_tick_period_ms() -> u64 { 100 } // Should not be less than `DEFAULT_META_OP_WAIT` as bootstrapping perform sync. consensus meta operations. const fn default_bootstrap_timeout_sec() -> u64 { 15 } const fn default_max_message_queue_size() -> usize { 100 } const fn default_connection_pool_size() -> usize { DEFAULT_POOL_SIZE } const fn default_message_timeout_tics() -> u64 { 10 } const fn default_compact_wal_entries() -> u64 { 128 } #[allow(clippy::unnecessary_wraps)] // Used as serde default const fn default_tls_cert_ttl() -> Option<u64> { // Default one hour Some(3600) } /// Custom validation function for metrics prefixes. fn validate_metrics_prefix(prefix: &str) -> Result<(), ValidationError> { // Prefix is not required if prefix.is_empty() { return Ok(()); } // Only allow alphanumeric characters or '_' if !prefix .chars() .all(|c| c.is_ascii_alphanumeric() || c == '_') { return Err( ValidationError::new("invalid_metrics_prefix").with_message(Cow::Borrowed( "Metrics prefix must be of all alphanumeric characters, with an exception for '_'", )), ); } Ok(()) } #[cfg(test)] mod tests { use std::io::Write; use fs_err as fs; use sealed_test::prelude::*; use super::*; /// Ensure we can successfully deserialize into [`Settings`] with just the default configuration. #[test] fn test_default_config() { Config::builder() .add_source(File::from_str(DEFAULT_CONFIG, FileFormat::Yaml)) .build() .expect("failed to build default config") .try_deserialize::<Settings>() .expect("failed to deserialize default config") .validate() .expect("failed to validate default config"); } #[expect( clippy::disallowed_methods, reason = "#[sealed_test] uses std::fs::copy" )] #[expect(clippy::disallowed_types, reason = "#[sealed_test] uses std::fs::File")] #[sealed_test(files = ["config/config.yaml", "config/development.yaml"])] fn test_runtime_development_config() { unsafe { env::set_var("RUN_MODE", "development") }; // `sealed_test` copies files into the same directory as the test runs in. // We need them in a subdirectory. fs::create_dir("config").expect("failed to create `config` subdirectory."); fs::copy("config.yaml", "config/config.yaml").expect("failed to copy `config.yaml`."); fs::copy("development.yaml", "config/development.yaml") .expect("failed to copy `development.yaml`."); // Read config let config = Settings::new(None).expect("failed to load development config at runtime"); // Validate config .validate() .expect("failed to validate development config at runtime"); assert!(config.load_errors.is_empty(), "must not have load errors") } #[expect(clippy::disallowed_types, reason = "#[sealed_test] uses std::fs::File")] #[sealed_test] fn test_no_config_files() { let non_existing_config_path = "config/non_existing_config".to_string(); // Read config let config = Settings::new(Some(non_existing_config_path)) .expect("failed to load with non-existing runtime config"); // Validate config .validate() .expect("failed to validate with non-existing runtime config"); assert!(!config.load_errors.is_empty(), "must have load errors") } #[expect(clippy::disallowed_types, reason = "#[sealed_test] uses std::fs::File")] #[sealed_test] fn test_custom_config() { let path = "config/custom.yaml"; // Create custom config file { fs::create_dir("config").unwrap(); let mut custom = fs::File::create(path).unwrap(); write!(&mut custom, "service:\n http_port: 9999").unwrap(); custom.flush().unwrap(); } // Load settings with custom config let config = Settings::new(Some(path.into())).unwrap(); // Ensure our custom config is the most important assert_eq!(config.service.http_port, 9999); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/wal_inspector.rs
src/wal_inspector.rs
use std::env; use std::path::Path; use collection::operations::OperationWithClockTag; use shard::wal::SerdeWal; use storage::content_manager::consensus::consensus_wal::ConsensusOpWal; use storage::content_manager::consensus_ops::ConsensusOperations; use wal::WalOptions; /// Executable to inspect the content of a write ahead log folder (collection OR consensus WAL). /// e.g: /// `cargo run --bin wal_inspector storage/collections/test-collection/0/wal/ collection` /// `cargo run --bin wal_inspector -- storage/node4/wal/ consensus` (expects `collections_meta_wal` folder as first child) fn main() { let args: Vec<String> = env::args().collect(); let wal_path = Path::new(&args[1]); let wal_type = args[2].as_str(); match wal_type { "collection" => print_collection_wal(wal_path), "consensus" => print_consensus_wal(wal_path), _ => eprintln!("Unknown wal type: {wal_type}"), } } fn print_consensus_wal(wal_path: &Path) { // must live within a folder named `collections_meta_wal` let wal = ConsensusOpWal::new(wal_path); println!("=========================="); let first_index = wal.first_entry().unwrap(); println!("First entry: {first_index:?}"); let last_index = wal.last_entry().unwrap(); println!("Last entry: {last_index:?}"); println!( "Offset of first entry: {:?}", wal.index_offset().unwrap().wal_to_raft_offset ); let entries = wal .entries( first_index.map(|f| f.index).unwrap_or(1), last_index.map(|f| f.index).unwrap_or(0) + 1, None, ) .unwrap(); for entry in entries { println!("=========================="); let command = ConsensusOperations::try_from(&entry); let data = match command { Ok(command) => format!("{command:?}"), Err(_) => format!("{:?}", entry.data), }; println!( "Entry ID:{}\nterm:{}\nentry_type:{}\ndata:{:?}", entry.index, entry.term, entry.entry_type, data ) } } fn print_collection_wal(wal_path: &Path) { let wal: Result<SerdeWal<OperationWithClockTag>, _> = SerdeWal::new(wal_path, WalOptions::default()); match wal { Err(error) => { eprintln!("Unable to open write ahead log in directory {wal_path:?}: {error}."); } Ok(wal) => { // print all entries let mut count = 0; for (idx, op) in wal.read_all(true) { println!("=========================="); println!( "Entry: {idx} Operation: {:?} Clock: {:?}", op.operation, op.clock_tag ); count += 1; } println!("=========================="); println!("End of WAL."); println!("Found {count} entries."); } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/issues_setup.rs
src/issues_setup.rs
use std::time::Duration; use collection::events::{CollectionDeletedEvent, IndexCreatedEvent, SlowQueryEvent}; use collection::problems::unindexed_field; use storage::issues_subscribers::UnindexedFieldSubscriber; use crate::settings::Settings; pub fn setup_subscribers(settings: &Settings) { settings .service .slow_query_secs .map(|secs| unindexed_field::SLOW_QUERY_THRESHOLD.set(Duration::from_secs_f32(secs))); let unindexed_subscriber = UnindexedFieldSubscriber; issues::broker::add_subscriber::<SlowQueryEvent>(Box::new(unindexed_subscriber)); issues::broker::add_subscriber::<IndexCreatedEvent>(Box::new(unindexed_subscriber)); issues::broker::add_subscriber::<CollectionDeletedEvent>(Box::new(unindexed_subscriber)); }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/greeting.rs
src/greeting.rs
use std::cmp::min; use std::env; use std::io::{IsTerminal, stdout}; use std::net::{IpAddr, Ipv4Addr, Ipv6Addr}; use api::rest::models::get_git_commit_id; use colored::{Color, ColoredString, Colorize}; use crate::settings::Settings; fn paint_red(text: &str, true_color: bool) -> ColoredString { if true_color { text.bold().truecolor(184, 20, 56) } else { text.bold().color(Color::Red) } } fn paint_green(text: &str, true_color: bool) -> ColoredString { if true_color { text.truecolor(134, 186, 144) } else { text.color(Color::Green) } } fn paint_blue(text: &str, true_color: bool) -> ColoredString { if true_color { text.bold().truecolor(82, 139, 183) } else { text.bold().color(Color::Blue) } } /// Check whether the given IP will be reachable from `localhost` /// /// This is a static analysis based on (very) common defaults and doesn't probe the current /// routing table. fn is_localhost_ip(host: &str) -> bool { let Ok(ip) = host.parse::<IpAddr>() else { return false; }; // Unspecified IPs bind to all interfaces, so `localhost` always points to it if ip == IpAddr::V4(Ipv4Addr::UNSPECIFIED) || ip == IpAddr::V6(Ipv6Addr::UNSPECIFIED) { return true; } // On all tested OSes IPv4 localhost points to `localhost` if ip == IpAddr::V4(Ipv4Addr::LOCALHOST) { return true; } // On macOS IPv6 localhost points to `localhost`, on Linux it is `ip6-localhost` if cfg!(target_os = "macos") && ip == IpAddr::V6(Ipv6Addr::LOCALHOST) { return true; } false } /// Prints welcome message pub fn welcome(settings: &Settings) { if !stdout().is_terminal() { colored::control::set_override(false); } let mut true_color = true; match env::var("COLORTERM") { Ok(val) => { if val != "24bit" && val != "truecolor" { true_color = false; } } Err(_) => true_color = false, } let title = [ r" _ _ ", r" __ _ __| |_ __ __ _ _ __ | |_ ", r" / _` |/ _` | '__/ _` | '_ \| __| ", r"| (_| | (_| | | | (_| | | | | |_ ", r" \__, |\__,_|_| \__,_|_| |_|\__| ", r" |_| ", ]; for line in title { println!("{}", paint_red(line, true_color)); } println!(); // Print current version and, if available, first 8 characters of the git commit hash let git_commit_info = get_git_commit_id() .map(|git_commit| { format!( ", {} {}", paint_green("build:", true_color), paint_blue(&git_commit[..min(8, git_commit.len())], true_color), ) }) .unwrap_or_default(); println!( "{} {}{}", paint_green("Version:", true_color), paint_blue(env!("CARGO_PKG_VERSION"), true_color), git_commit_info ); // Print link to web UI let ui_link = format!( "http{}://{}:{}/dashboard", if settings.service.enable_tls { "s" } else { "" }, if is_localhost_ip(&settings.service.host) { "localhost" } else { &settings.service.host }, settings.service.http_port ); println!( "{} {}", paint_green("Access web UI at", true_color), paint_blue(&ui_link, true_color).underline() ); println!(); } #[cfg(test)] mod tests { use super::*; #[test] fn test_welcome() { welcome(&Settings::new(None).unwrap()); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/segment_inspector.rs
src/segment_inspector.rs
use std::path::Path; use std::sync::atomic::AtomicBool; use clap::Parser; use common::counter::hardware_counter::HardwareCounterCell; use segment::entry::entry_point::SegmentEntry; use segment::segment_constructor::load_segment; use segment::types::PointIdType; #[derive(Parser, Debug)] #[command(version, about)] struct Args { /// Path to the segment folder. May be a list #[clap(short, long, num_args=1..)] path: Vec<String>, /// Print segment info #[clap(long)] info: bool, /// Point ID to inspect #[clap(long)] point_id_int: Option<u64>, /// Point ID to inspect (UUID) #[clap(long)] point_id_uuid: Option<String>, } fn main() { let args: Args = Args::parse(); for segment_path in args.path { let path = Path::new(&segment_path); if !path.exists() { eprintln!("Path does not exist: {segment_path}"); continue; } if !path.is_dir() { eprintln!("Path is not a directory: {segment_path}"); continue; } // Open segment let segment = load_segment(path, &AtomicBool::new(false)) .unwrap() .unwrap(); eprintln!( "path = {:#?}, size-points = {}", path, segment.available_point_count() ); if args.info { let info = segment.info(); eprintln!("info = {info:#?}"); } if let Some(point_id_int) = args.point_id_int { let point_id = PointIdType::NumId(point_id_int); let internal_id = segment.get_internal_id(point_id); if internal_id.is_some() { let version = segment.point_version(point_id); let payload = segment .payload(point_id, &HardwareCounterCell::disposable()) .unwrap(); // let vectors = segment.all_vectors(point_id).unwrap(); println!("Internal ID: {internal_id:?}"); println!("Version: {version:?}"); println!("Payload: {payload:?}"); // println!("Vectors: {vectors:?}"); } } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/snapshots.rs
src/snapshots.rs
use std::io::BufReader; use std::path::{Path, PathBuf}; use collection::collection::Collection; use collection::shards::shard::PeerId; use fs_err as fs; use fs_err::File; use log::info; use segment::common::validate_snapshot_archive::open_snapshot_archive_with_validation; use storage::content_manager::alias_mapping::AliasPersistence; use storage::content_manager::snapshots::SnapshotConfig; use storage::content_manager::toc::{ALIASES_PATH, COLLECTIONS_DIR}; /// Recover snapshots from the given arguments /// /// # Arguments /// /// * `mapping` - `[ <path>:<collection_name> ]` /// * `force` - if true, allow to overwrite collections from snapshots /// /// # Returns /// /// * `Vec<String>` - list of collections that were recovered pub fn recover_snapshots( mapping: &[String], force: bool, temp_dir: Option<&str>, storage_dir: &str, this_peer_id: PeerId, is_distributed: bool, ) -> Vec<String> { let collection_dir_path = Path::new(storage_dir).join(COLLECTIONS_DIR); let mut recovered_collections: Vec<String> = vec![]; for snapshot_params in mapping { let mut split = snapshot_params.split(':'); let path = split .next() .unwrap_or_else(|| panic!("Snapshot path is missing: {snapshot_params}")); let snapshot_path = Path::new(path); let collection_name = split .next() .unwrap_or_else(|| panic!("Collection name is missing: {snapshot_params}")); recovered_collections.push(collection_name.to_string()); assert!( split.next().is_none(), "Too many parts in snapshot mapping: {snapshot_params}" ); info!("Recovering snapshot {collection_name} from {path}"); // check if collection already exists // if it does, we need to check if we want to overwrite it // if not, we need to abort let collection_path = collection_dir_path.join(collection_name); info!("Collection path: {}", collection_path.display()); if collection_path.exists() { if !force { panic!( "Collection {collection_name} already exists. Use --force-snapshot to overwrite it." ); } info!("Overwriting collection {collection_name}"); } let collection_temp_path = temp_dir .map(PathBuf::from) .unwrap_or_else(|| collection_path.with_extension("tmp")); if let Err(err) = Collection::restore_snapshot( snapshot_path, &collection_temp_path, this_peer_id, is_distributed, ) { panic!("Failed to recover snapshot {collection_name}: {err}"); } // Remove collection_path directory if exists if collection_path.exists() && let Err(err) = fs::remove_dir_all(&collection_path) { panic!("Failed to remove collection {collection_name}: {err}"); } fs::rename(&collection_temp_path, &collection_path).unwrap(); } recovered_collections } pub fn recover_full_snapshot( temp_dir: Option<&str>, snapshot_path: &str, storage_dir: &str, force: bool, this_peer_id: PeerId, is_distributed: bool, ) -> Vec<String> { let snapshot_temp_path = temp_dir .map(PathBuf::from) .unwrap_or_else(|| Path::new(storage_dir).join("snapshots_recovery_tmp")); fs::create_dir_all(&snapshot_temp_path).unwrap(); // Un-tar snapshot into temporary directory let mut ar = open_snapshot_archive_with_validation(Path::new(snapshot_path)).unwrap(); ar.unpack(&snapshot_temp_path).unwrap(); // Read configuration file with snapshot-to-collection mapping let config_path = snapshot_temp_path.join("config.json"); let config_file = BufReader::new(File::open(config_path).unwrap()); let config_json: SnapshotConfig = serde_json::from_reader(config_file).unwrap(); // Create mapping from the configuration file let mapping: Vec<String> = config_json .collections_mapping .iter() .map(|(collection_name, snapshot_file)| { format!( "{}:{collection_name}", snapshot_temp_path.join(snapshot_file).to_str().unwrap(), ) }) .collect(); // Launch regular recovery of snapshots let recovered_collection = recover_snapshots( &mapping, force, temp_dir, storage_dir, this_peer_id, is_distributed, ); let alias_path = Path::new(storage_dir).join(ALIASES_PATH); let mut alias_persistence = AliasPersistence::open(&alias_path).expect("Can't open database by the provided config"); for (alias, collection_name) in config_json.collections_aliases { if alias_persistence.get(&alias).is_some() && !force { panic!("Alias {alias} already exists. Use --force-snapshot to overwrite it."); } alias_persistence.insert(alias, collection_name).unwrap(); } // Remove temporary directory fs::remove_dir_all(&snapshot_temp_path).unwrap(); recovered_collection }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/wal_pop.rs
src/wal_pop.rs
use std::env; use std::path::Path; use wal::Wal; fn main() { let args: Vec<String> = env::args().collect(); let wal_path = Path::new(&args[1]); let mut wal = Wal::open(wal_path).expect("Can't open consensus WAL"); let last_index = wal.last_index(); eprintln!("last_index = {last_index}"); wal.truncate(last_index).unwrap(); wal.flush_open_segment().unwrap(); }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/startup.rs
src/startup.rs
//! Contains a collection of functions that are called at the start of the program. use std::backtrace::Backtrace; use std::panic; use std::path::PathBuf; use fs_err as fs; use crate::common::error_reporting::ErrorReporter; const DEFAULT_INITIALIZED_FILE: &str = ".qdrant-initialized"; fn get_init_file_path() -> PathBuf { std::env::var("QDRANT_INIT_FILE_PATH") .map(PathBuf::from) .unwrap_or_else(|_| DEFAULT_INITIALIZED_FILE.into()) } pub fn setup_panic_hook(reporting_enabled: bool, reporting_id: String) { panic::set_hook(Box::new(move |panic_info| { let backtrace = Backtrace::force_capture().to_string(); let loc = if let Some(loc) = panic_info.location() { format!(" in file {} at line {}", loc.file(), loc.line()) } else { String::new() }; let message = if let Some(s) = panic_info.payload().downcast_ref::<&str>() { s } else if let Some(s) = panic_info.payload().downcast_ref::<String>() { s } else { "Payload not captured as it is not a string." }; log::error!("Panic backtrace: \n{backtrace}"); log::error!("Panic occurred{loc}: {message}"); if reporting_enabled { ErrorReporter::report(message, &reporting_id, Some(&loc)); } })); } /// Creates a file that indicates that the server has been started. /// This file is used to check if the server has been successfully started before potential kill. pub fn touch_started_file_indicator() { if let Err(err) = fs::write(get_init_file_path(), "") { log::warn!("Failed to create init file indicator: {err}"); } } /// Removes a file that indicates that the server has been started. /// Use before server initialization to avoid false positives. pub fn remove_started_file_indicator() { let path = get_init_file_path(); if path.exists() && let Err(err) = fs::remove_file(path) { log::warn!("Failed to remove init file indicator: {err}"); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/main.rs
src/main.rs
mod actix; mod common; mod consensus; mod greeting; mod issues_setup; mod migrations; mod settings; mod snapshots; mod startup; mod tonic; mod tracing; use std::io::Error; use std::path::Path; use std::sync::Arc; use std::thread; use std::thread::JoinHandle; use std::time::Duration; use ::common::budget::{ResourceBudget, get_io_budget}; use ::common::cpu::get_cpu_budget; use ::common::flags::{feature_flags, init_feature_flags}; use ::tonic::transport::Uri; use api::grpc::transport_channel_pool::TransportChannelPool; use clap::Parser; use collection::profiling::interface::init_requests_profile_collector; use collection::shards::channel_service::ChannelService; use consensus::Consensus; use fs_err as fs; use memory::checkfs::{check_fs_info, check_mmap_functionality}; use memory::mmap_ops::MULTI_MMAP_SUPPORT_CHECK_RESULT; use slog::Drain; use startup::setup_panic_hook; use storage::content_manager::consensus::operation_sender::OperationSender; use storage::content_manager::consensus::persistent::Persistent; use storage::content_manager::consensus_manager::{ConsensusManager, ConsensusStateRef}; use storage::content_manager::toc::TableOfContent; use storage::content_manager::toc::dispatcher::TocDispatcher; use storage::dispatcher::Dispatcher; use storage::rbac::Access; #[cfg(all( not(target_env = "msvc"), any(target_arch = "x86_64", target_arch = "aarch64") ))] use tikv_jemallocator::Jemalloc; use crate::common::helpers::{ create_general_purpose_runtime, create_search_runtime, create_update_runtime, load_tls_client_config, }; use crate::common::inference::service::InferenceService; use crate::common::telemetry::TelemetryCollector; use crate::common::telemetry_reporting::TelemetryReporter; use crate::greeting::welcome; use crate::migrations::single_to_cluster::handle_existing_collections; use crate::settings::Settings; use crate::snapshots::{recover_full_snapshot, recover_snapshots}; use crate::startup::{remove_started_file_indicator, touch_started_file_indicator}; #[cfg(all( not(target_env = "msvc"), any(target_arch = "x86_64", target_arch = "aarch64") ))] #[global_allocator] static GLOBAL: Jemalloc = Jemalloc; const FULL_ACCESS: Access = Access::full("For main"); /// Qdrant (read: quadrant ) is a vector similarity search engine. /// It provides a production-ready service with a convenient API to store, search, and manage points - vectors with an additional payload. /// /// This CLI starts a Qdrant peer/server. #[derive(Parser, Debug)] #[command(version, about)] struct Args { /// Uri of the peer to bootstrap from in case of multi-peer deployment. /// If not specified - this peer will be considered as a first in a new deployment. #[arg(long, value_parser, value_name = "URI", env = "QDRANT_BOOTSTRAP")] bootstrap: Option<Uri>, /// Uri of this peer. /// Other peers should be able to reach it by this uri. /// /// This value has to be supplied if this is the first peer in a new deployment. /// /// In case this is not the first peer and it bootstraps the value is optional. /// If not supplied then qdrant will take internal grpc port from config and derive the IP address of this peer on bootstrap peer (receiving side) #[arg(long, value_parser, value_name = "URI", env = "QDRANT_URI")] uri: Option<Uri>, /// Force snapshot re-creation /// If provided - existing collections will be replaced with snapshots. /// Default is to not recreate from snapshots. #[arg(short, long, action, default_value_t = false)] force_snapshot: bool, /// List of paths to snapshot files. /// Format: <snapshot_file_path>:<target_collection_name> /// /// WARN: Do not use this option if you are recovering collection in existing distributed cluster. /// Use `/collections/<collection-name>/snapshots/recover` API instead. #[arg(long, value_name = "PATH:NAME", alias = "collection-snapshot")] snapshot: Option<Vec<String>>, /// Path to snapshot of multiple collections. /// Format: <snapshot_file_path> /// /// WARN: Do not use this option if you are recovering collection in existing distributed cluster. /// Use `/collections/<collection-name>/snapshots/recover` API instead. #[arg(long, value_name = "PATH")] storage_snapshot: Option<String>, /// Path to an alternative configuration file. /// Format: <config_file_path> /// /// Default path: config/config.yaml #[arg(long, value_name = "PATH")] config_path: Option<String>, /// Disable telemetry sending to developers /// If provided - telemetry collection will be disabled. /// Read more: <https://qdrant.tech/documentation/guides/telemetry> #[arg(long, action, default_value_t = false)] disable_telemetry: bool, /// Run stacktrace collector. Used for debugging. #[arg(long, action, default_value_t = false)] stacktrace: bool, /// Reinit consensus state. /// When enabled, the service will assume the consensus should be reinitialized. /// The exact behavior depends on if this current node has bootstrap URI or not. /// If it has - it'll remove current consensus state and consensus WAL (while keeping peer ID) /// and will try to receive state from the bootstrap peer. /// If it doesn't have - it'll remove other peers from voters promote /// the current peer to the leader and the single member of the cluster. /// It'll also compact consensus WAL to force snapshot #[arg(long, action, default_value_t = false)] reinit: bool, } fn main() -> anyhow::Result<()> { let args = Args::parse(); // Run backtrace collector, expected to used by `rstack` crate if args.stacktrace { #[cfg(all(target_os = "linux", feature = "stacktrace"))] { let _ = rstack_self::child(); } return Ok(()); } let settings = Settings::new(args.config_path)?; // Set global feature flags, sourced from configuration init_feature_flags(settings.feature_flags); let reporting_enabled = !settings.telemetry_disabled && !args.disable_telemetry; let reporting_id = TelemetryCollector::generate_id(); // Setup logging (no logging before this point) let logger_handle = tracing::setup( settings .logger .with_top_level_directive(settings.log_level.clone()), )?; remove_started_file_indicator(); setup_panic_hook(reporting_enabled, reporting_id.to_string()); memory::madvise::set_global(settings.storage.mmap_advice); segment::vector_storage::common::set_async_scorer( settings .storage .performance .async_scorer .unwrap_or_default(), ); welcome(&settings); #[cfg(feature = "gpu")] if let Some(settings_gpu) = &settings.gpu { use segment::index::hnsw_index::gpu::*; // initialize GPU devices manager. if settings_gpu.indexing { set_gpu_force_half_precision(settings_gpu.force_half_precision); set_gpu_groups_count(settings_gpu.groups_count); let mut gpu_device_manager = GPU_DEVICES_MANAGER.write(); *gpu_device_manager = match gpu_devices_manager::GpuDevicesMaganer::new( &settings_gpu.device_filter, settings_gpu.devices.as_deref(), settings_gpu.allow_integrated, settings_gpu.allow_emulated, true, // Currently we always wait for the free gpu device. settings_gpu.parallel_indexes.unwrap_or(1), ) { Ok(gpu_device_manager) => Some(gpu_device_manager), Err(err) => { log::error!("Can't initialize GPU devices manager: {err}"); None } } } } if let Some(recovery_warning) = &settings.storage.recovery_mode { log::warn!("Qdrant is loaded in recovery mode: {recovery_warning}"); log::warn!( "Read more: https://qdrant.tech/documentation/guides/administration/#recovery-mode" ); } // Validate as soon as possible, but we must initialize logging first settings.validate_and_warn(); fs::create_dir_all(&settings.storage.storage_path)?; // Check if the filesystem is compatible with Qdrant let mmaps_working; match check_fs_info(&settings.storage.storage_path) { memory::checkfs::FsCheckResult::Good => { mmaps_working = true; } memory::checkfs::FsCheckResult::Unknown(details) => { match check_mmap_functionality(&settings.storage.storage_path) { Ok(true) => { log::warn!( "There is a potential issue with the filesystem for storage path {}. Details: {details}", settings.storage.storage_path, ); mmaps_working = true; } Ok(false) => { log::error!( "Filesystem check failed for storage path {}. Details: {details}", settings.storage.storage_path, ); mmaps_working = false; } Err(e) => { log::error!( "Unable to check mmap functionality for storage path {}. Details: {details}, error: {e}", settings.storage.storage_path, ); mmaps_working = false; } } } memory::checkfs::FsCheckResult::Bad(details) => { log::error!( "Filesystem check failed for storage path {}. Details: {details}", settings.storage.storage_path, ); mmaps_working = false; } } let _ = MULTI_MMAP_SUPPORT_CHECK_RESULT.set(mmaps_working); // Report feature flags that are enabled for easier debugging let flags = feature_flags(); if !flags.is_default() { log::debug!("Feature flags: {flags:?}"); } let bootstrap = if args.bootstrap == args.uri { if args.bootstrap.is_some() { log::warn!( "Bootstrap URI is the same as this peer URI. Consider this peer as a first in a new deployment.", ); } None } else { args.bootstrap }; // Saved state of the consensus. let persistent_consensus_state = Persistent::load_or_init( &settings.storage.storage_path, bootstrap.is_none(), args.reinit, settings.cluster.peer_id, )?; let is_distributed_deployment = settings.cluster.enabled; let temp_path = settings.storage.temp_path.as_deref(); let restored_collections = if let Some(full_snapshot) = args.storage_snapshot { recover_full_snapshot( temp_path, &full_snapshot, &settings.storage.storage_path, args.force_snapshot, persistent_consensus_state.this_peer_id(), is_distributed_deployment, ) } else if let Some(snapshots) = args.snapshot { // recover from snapshots recover_snapshots( &snapshots, args.force_snapshot, temp_path, &settings.storage.storage_path, persistent_consensus_state.this_peer_id(), is_distributed_deployment, ) } else { vec![] }; // Create and own search runtime out of the scope of async context to ensure correct // destruction of it let search_runtime = create_search_runtime(settings.storage.performance.max_search_threads) .expect("Can't search create runtime."); let update_runtime = create_update_runtime( settings .storage .performance .max_optimization_runtime_threads, ) .expect("Can't optimizer create runtime."); let general_runtime = create_general_purpose_runtime().expect("Can't optimizer general purpose runtime."); let runtime_handle = general_runtime.handle().clone(); // Use global CPU budget for optimizations based on settings let cpu_budget = get_cpu_budget(settings.storage.performance.optimizer_cpu_budget); let io_budget = get_io_budget(settings.storage.performance.optimizer_io_budget, cpu_budget); let optimizer_resource_budget = ResourceBudget::new(cpu_budget, io_budget); // Create a signal sender and receiver. It is used to communicate with the consensus thread. let (propose_sender, propose_receiver) = std::sync::mpsc::channel(); let propose_operation_sender = if settings.cluster.enabled { // High-level channel which could be used to send User-space consensus operations Some(OperationSender::new(propose_sender)) } else { // We don't need sender for the single-node mode None }; // Channel service is used to manage connections between peers. // It allocates required number of channels and manages proper reconnection handling let mut channel_service = ChannelService::new(settings.service.http_port, settings.service.api_key.clone()); if is_distributed_deployment { // We only need channel_service in case if cluster is enabled. // So we initialize it with real values here let p2p_grpc_timeout = Duration::from_millis(settings.cluster.grpc_timeout_ms); let connection_timeout = Duration::from_millis(settings.cluster.connection_timeout_ms); let tls_config = load_tls_client_config(&settings)?; channel_service.channel_pool = Arc::new(TransportChannelPool::new( p2p_grpc_timeout, connection_timeout, settings.cluster.p2p.connection_pool_size, tls_config, )); channel_service.id_to_address = persistent_consensus_state.peer_address_by_id.clone(); channel_service.id_to_metadata = persistent_consensus_state.peer_metadata_by_id.clone(); } // Table of content manages the list of collections. // It is a main entry point for the storage. let toc = TableOfContent::new( &settings.storage, search_runtime, update_runtime, general_runtime, optimizer_resource_budget, channel_service.clone(), persistent_consensus_state.this_peer_id(), propose_operation_sender.clone(), ); toc.clear_all_tmp_directories()?; // Here we load all stored collections. runtime_handle.block_on(async { for collection in toc.all_collections(&FULL_ACCESS).await { log::debug!("Loaded collection: {collection}"); } }); let toc_arc = Arc::new(toc); let storage_path = toc_arc.storage_path(); // Holder for all actively running threads of the service: web, gPRC, consensus, etc. let mut handles: Vec<JoinHandle<Result<(), Error>>> = vec![]; // Router for external queries. // It decides if query should go directly to the ToC or through the consensus. let mut dispatcher = Dispatcher::new(toc_arc.clone()); let (telemetry_collector, dispatcher_arc, health_checker) = if is_distributed_deployment { let consensus_state: ConsensusStateRef = ConsensusManager::new( persistent_consensus_state, toc_arc.clone(), propose_operation_sender.unwrap(), Path::new(storage_path), ) .expect("initialize consensus manager") .into(); let is_new_deployment = consensus_state.is_new_deployment(); dispatcher = dispatcher.with_consensus(consensus_state.clone(), settings.cluster.resharding_enabled); let toc_dispatcher = TocDispatcher::new(Arc::downgrade(&toc_arc), consensus_state.clone()); toc_arc.with_toc_dispatcher(toc_dispatcher); let dispatcher_arc = Arc::new(dispatcher); // Monitoring and telemetry. let telemetry_collector = TelemetryCollector::new(settings.clone(), dispatcher_arc.clone(), reporting_id); let tonic_telemetry_collector = telemetry_collector.tonic_telemetry_collector.clone(); // `raft` crate uses `slog` crate so it is needed to use `slog_stdlog::StdLog` to forward // logs from it to `log` crate let slog_logger = slog::Logger::root(slog_stdlog::StdLog.fuse(), slog::o!()); // Runs raft consensus in a separate thread. // Create a pipe `message_sender` to communicate with the consensus let health_checker = Arc::new(common::health::HealthChecker::spawn( toc_arc.clone(), consensus_state.clone(), &runtime_handle, // NOTE: `wait_for_bootstrap` should be calculated *before* starting `Consensus` thread consensus_state.is_new_deployment() && bootstrap.is_some(), )); let handle = Consensus::run( &slog_logger, consensus_state.clone(), bootstrap, args.uri.map(|uri| uri.to_string()), settings.clone(), channel_service, propose_receiver, tonic_telemetry_collector, toc_arc.clone(), runtime_handle.clone(), args.reinit, ) .expect("Can't initialize consensus"); handles.push(handle); let toc_arc_clone = toc_arc.clone(); let consensus_state_clone = consensus_state.clone(); let _cancel_transfer_handle = runtime_handle.spawn(async move { consensus_state_clone.is_leader_established.await_ready(); match toc_arc_clone .cancel_related_transfers("Source or target peer restarted") .await { Ok(_) => { log::debug!("All transfers if any cancelled"); } Err(err) => { log::error!("Can't cancel related transfers: {err}"); } } }); // TODO(resharding): Remove resharding driver? // // runtime_handle.block_on(async { // toc_arc.resume_resharding_tasks().await; // }); let collections_to_recover_in_consensus = if is_new_deployment { let existing_collections = runtime_handle.block_on(toc_arc.all_collections(&FULL_ACCESS)); existing_collections .into_iter() .map(|pass| pass.name().to_string()) .collect() } else { restored_collections }; if !collections_to_recover_in_consensus.is_empty() { runtime_handle.block_on(handle_existing_collections( toc_arc.clone(), consensus_state.clone(), dispatcher_arc.clone(), consensus_state.this_peer_id(), collections_to_recover_in_consensus, )); } (telemetry_collector, dispatcher_arc, Some(health_checker)) } else { log::info!("Distributed mode disabled"); let dispatcher_arc = Arc::new(dispatcher); // Monitoring and telemetry. let telemetry_collector = TelemetryCollector::new(settings.clone(), dispatcher_arc.clone(), reporting_id); (telemetry_collector, dispatcher_arc, None) }; let tonic_telemetry_collector = telemetry_collector.tonic_telemetry_collector.clone(); // // Telemetry reporting // let reporting_id = telemetry_collector.reporting_id(); let telemetry_collector = Arc::new(tokio::sync::Mutex::new(telemetry_collector)); if reporting_enabled { log::info!("Telemetry reporting enabled, id: {reporting_id}"); runtime_handle.spawn(TelemetryReporter::run(telemetry_collector.clone())); } else { log::info!("Telemetry reporting disabled"); } if settings.service.hardware_reporting == Some(true) { log::info!("Hardware reporting enabled"); } // Setup subscribers to listen for issue-able events issues_setup::setup_subscribers(&settings); init_requests_profile_collector(runtime_handle.clone()); // Helper to better log start errors let log_err_if_any = |server_name, result| match result { Err(err) => { log::error!("Error while starting {server_name} server: {err}"); Err(err) } ok => ok, }; // // Inference Service // if let Err(err) = InferenceService::init_global(settings.inference.clone()) { log::error!("Inference service init failed: {err}"); } // // REST API server // { let dispatcher_arc = dispatcher_arc.clone(); let settings = settings.clone(); let handle = thread::Builder::new() .name("web".to_string()) .spawn(move || { log_err_if_any( "REST", actix::init( dispatcher_arc.clone(), telemetry_collector, health_checker, settings, logger_handle, ), ) }) .unwrap(); handles.push(handle); } // // gRPC server // if let Some(grpc_port) = settings.service.grpc_port { let settings = settings.clone(); let handle = thread::Builder::new() .name("grpc".to_string()) .spawn(move || { log_err_if_any( "gRPC", tonic::init( dispatcher_arc, tonic_telemetry_collector, settings, grpc_port, runtime_handle, ), ) }) .unwrap(); handles.push(handle); } else { log::info!("gRPC endpoint disabled"); } #[cfg(feature = "service_debug")] { use std::fmt::Write; use parking_lot::deadlock; const DEADLOCK_CHECK_PERIOD: Duration = Duration::from_secs(10); thread::Builder::new() .name("deadlock_checker".to_string()) .spawn(move || { loop { thread::sleep(DEADLOCK_CHECK_PERIOD); let deadlocks = deadlock::check_deadlock(); if deadlocks.is_empty() { continue; } let mut error = format!("{} deadlocks detected\n", deadlocks.len()); for (i, threads) in deadlocks.iter().enumerate() { writeln!(error, "Deadlock #{i}").expect("fail to writeln!"); for t in threads { writeln!( error, "Thread Id {:#?}\n{:#?}", t.thread_id(), t.backtrace(), ) .expect("fail to writeln!"); } } log::error!("{error}"); } }) .unwrap(); } touch_started_file_indicator(); for handle in handles { log::debug!( "Waiting for thread {} to finish", handle.thread().name().unwrap() ); handle.join().expect("thread is not panicking")?; } drop(toc_arc); drop(settings); Ok(()) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/consensus.rs
src/consensus.rs
use std::collections::{HashMap, HashSet}; use std::str::FromStr; use std::sync::{Arc, mpsc}; use std::thread::JoinHandle; use std::time::{Duration, Instant}; use std::{cmp, fmt, thread}; use anyhow::{Context as _, anyhow}; use api::grpc::dynamic_channel_pool::make_grpc_channel; use api::grpc::qdrant::raft_client::RaftClient; use api::grpc::qdrant::{AllPeers, PeerId as GrpcPeerId, RaftMessage as GrpcRaftMessage}; use api::grpc::transport_channel_pool::TransportChannelPool; use collection::shards::channel_service::ChannelService; use collection::shards::shard::PeerId; #[cfg(target_os = "linux")] use common::cpu::linux_high_thread_priority; use raft::eraftpb::Message as RaftMessage; use raft::prelude::*; use raft::{INVALID_ID, SoftState, StateRole}; use storage::content_manager::consensus_manager::ConsensusStateRef; use storage::content_manager::consensus_ops::{ConsensusOperations, SnapshotStatus}; use storage::content_manager::toc::TableOfContent; use tokio::runtime::Handle; use tokio::sync::mpsc::{Receiver, Sender}; use tokio::sync::watch; use tokio::time::sleep; use tonic::transport::{ClientTlsConfig, Uri}; use crate::common::helpers; use crate::common::telemetry_ops::requests_telemetry::TonicTelemetryCollector; use crate::settings::{ConsensusConfig, Settings}; use crate::tonic::init_internal; type Node = RawNode<ConsensusStateRef>; const RECOVERY_RETRY_TIMEOUT: Duration = Duration::from_secs(1); const RECOVERY_MAX_RETRY_COUNT: usize = 3; pub enum Message { FromClient(ConsensusOperations), FromPeer(Box<RaftMessage>), } /// Aka Consensus Thread /// Manages proposed changes to consensus state, ensures that everything is ordered properly pub struct Consensus { /// Raft structure which handles raft-related state node: Node, /// Receives proposals from peers and client for applying in consensus receiver: Receiver<Message>, /// Runtime for async message sending runtime: Handle, /// Uri to some other known peer, used to join the consensus /// ToDo: Make if many config: ConsensusConfig, broker: RaftMessageBroker, raft_config: Config, } impl Consensus { /// Create and run consensus node #[allow(clippy::too_many_arguments)] pub fn run( logger: &slog::Logger, state_ref: ConsensusStateRef, bootstrap_peer: Option<Uri>, uri: Option<String>, settings: Settings, channel_service: ChannelService, propose_receiver: mpsc::Receiver<ConsensusOperations>, telemetry_collector: Arc<parking_lot::Mutex<TonicTelemetryCollector>>, toc: Arc<TableOfContent>, runtime: Handle, reinit: bool, ) -> anyhow::Result<JoinHandle<std::io::Result<()>>> { let tls_client_config = helpers::load_tls_client_config(&settings)?; let p2p_host = settings.service.host.clone(); let p2p_port = settings.cluster.p2p.port.expect("P2P port is not set"); let config = settings.cluster.consensus.clone(); let (mut consensus, message_sender) = Self::new( logger, state_ref.clone(), bootstrap_peer, uri, p2p_port, config, tls_client_config, channel_service, runtime.clone(), reinit, )?; let state_ref_clone = state_ref.clone(); thread::Builder::new() .name("consensus".to_string()) .spawn(move || { // On Linux, try to use high thread priority because consensus is important // Likely fails as we cannot set a higher priority by default due to permissions #[cfg(target_os = "linux")] if let Err(err) = linux_high_thread_priority() { log::debug!( "Failed to set high thread priority for consensus, ignoring: {err}" ); } if let Err(err) = consensus.start() { log::error!("Consensus stopped with error: {err:#}"); state_ref_clone.on_consensus_thread_err(err); } else { log::info!("Consensus stopped"); state_ref_clone.on_consensus_stopped(); } })?; let message_sender_moved = message_sender.clone(); thread::Builder::new() .name("forward-proposals".to_string()) .spawn(move || { // On Linux, try to use high thread priority because consensus is important // Likely fails as we cannot set a higher priority by default due to permissions #[cfg(target_os = "linux")] if let Err(err) = linux_high_thread_priority() { log::debug!( "Failed to set high thread priority for consensus, ignoring: {err}" ); } while let Ok(entry) = propose_receiver.recv() { if message_sender_moved .blocking_send(Message::FromClient(entry)) .is_err() { log::error!("Can not forward new entry to consensus as it was stopped."); break; } } })?; let server_tls = if settings.cluster.p2p.enable_tls { let tls_config = settings .tls .clone() .ok_or_else(Settings::tls_config_is_undefined_error)?; Some(helpers::load_tls_internal_server_config(&tls_config)?) } else { None }; let handle = thread::Builder::new() .name("grpc_internal".to_string()) .spawn(move || { init_internal( toc, state_ref, telemetry_collector, settings, p2p_host, p2p_port, server_tls, message_sender, runtime, ) }) .unwrap(); Ok(handle) } /// If `bootstrap_peer` peer is supplied, then either `uri` or `p2p_port` should be also supplied #[allow(clippy::too_many_arguments)] pub fn new( logger: &slog::Logger, state_ref: ConsensusStateRef, bootstrap_peer: Option<Uri>, uri: Option<String>, p2p_port: u16, config: ConsensusConfig, tls_config: Option<ClientTlsConfig>, channel_service: ChannelService, runtime: Handle, reinit: bool, ) -> anyhow::Result<(Self, Sender<Message>)> { // If we want to re-initialize consensus, we need to prevent other peers // from re-playing consensus WAL operations, as they should already have them applied. // Do ensure that we are forcing compacting WAL on the first re-initialized peer, // which should trigger snapshot transferring instead of replaying WAL. let force_compact_wal = reinit && bootstrap_peer.is_none(); // On the bootstrap-ed peers during reinit of the consensus // we want to make sure only the bootstrap peer will hold the true state // Therefore we clear the WAL on the bootstrap peer to force it to request a snapshot let clear_wal = reinit && bootstrap_peer.is_some(); if clear_wal { log::debug!("Clearing WAL on the bootstrap peer to force snapshot transfer"); state_ref.clear_wal()?; } // raft will not return entries to the application smaller or equal to `applied` let last_applied = state_ref.last_applied_entry().unwrap_or_default(); let raft_config = Config { id: state_ref.this_peer_id(), applied: last_applied, ..Default::default() }; raft_config.validate()?; // bounded channel for backpressure let (sender, receiver) = tokio::sync::mpsc::channel(config.max_message_queue_size); // State might be initialized but the node might be shutdown without actually syncing or committing anything. if state_ref.is_new_deployment() || reinit { let leader_established_in_ms = config.tick_period_ms * raft_config.max_election_tick() as u64; Self::init( &state_ref, bootstrap_peer.clone(), uri, p2p_port, &config, tls_config.clone(), &runtime, leader_established_in_ms, ) .map_err(|err| anyhow!("Failed to initialize Consensus for new Raft state: {err}"))?; } else { runtime .block_on(Self::recover( &state_ref, uri.clone(), p2p_port, &config, tls_config.clone(), )) .map_err(|err| { anyhow!("Failed to recover Consensus from existing Raft state: {err}") })?; if bootstrap_peer.is_some() || uri.is_some() { log::debug!("Local raft state found - bootstrap and uri cli arguments were ignored") } log::debug!("Local raft state found - skipping initialization"); }; let mut node = Node::new(&raft_config, state_ref.clone(), logger)?; node.set_batch_append(true); // Before consensus has started apply any unapplied committed entries // They might have not been applied due to unplanned Qdrant shutdown let _stop_consensus = state_ref.apply_entries(&mut node)?; if force_compact_wal { // Making sure that the WAL will be compacted on start state_ref.compact_wal(1)?; } else { state_ref.compact_wal(config.compact_wal_entries)?; } let broker = RaftMessageBroker::new( runtime.clone(), bootstrap_peer, tls_config, config.clone(), node.store().clone(), channel_service.channel_pool, ); let consensus = Self { node, receiver, runtime, config, broker, raft_config, }; if !state_ref.is_new_deployment() { state_ref.recover_first_voter()?; } Ok((consensus, sender)) } #[allow(clippy::too_many_arguments)] fn init( state_ref: &ConsensusStateRef, bootstrap_peer: Option<Uri>, uri: Option<String>, p2p_port: u16, config: &ConsensusConfig, tls_config: Option<ClientTlsConfig>, runtime: &Handle, leader_established_in_ms: u64, ) -> anyhow::Result<()> { if let Some(bootstrap_peer) = bootstrap_peer { log::debug!("Bootstrapping from peer with address: {bootstrap_peer}"); runtime.block_on(Self::bootstrap( state_ref, bootstrap_peer, uri, p2p_port, config, tls_config, ))?; Ok(()) } else { log::debug!( "Bootstrapping is disabled. Assuming this peer is the first in the network" ); let tick_period = config.tick_period_ms; log::info!( "With current tick period of {tick_period}ms, leader will be established in approximately {leader_established_in_ms}ms. To avoid rejected operations - add peers and submit operations only after this period.", ); // First peer needs to add its own address state_ref.add_peer( state_ref.this_peer_id(), uri.ok_or_else(|| anyhow::anyhow!("First peer should specify its uri."))? .parse()?, )?; Ok(()) } } async fn add_peer_to_known_for( this_peer_id: PeerId, cluster_uri: Uri, current_uri: Option<String>, p2p_port: u16, config: &ConsensusConfig, tls_config: Option<ClientTlsConfig>, ) -> anyhow::Result<AllPeers> { // Use dedicated transport channel for bootstrapping because of specific timeout let channel = make_grpc_channel( Duration::from_secs(config.bootstrap_timeout_sec), Duration::from_secs(config.bootstrap_timeout_sec), cluster_uri, tls_config, ) .await .map_err(|err| anyhow!("Failed to create timeout channel: {err}"))?; let mut client = RaftClient::new(channel); let all_peers = client .add_peer_to_known(tonic::Request::new( api::grpc::qdrant::AddPeerToKnownMessage { uri: current_uri, port: Some(u32::from(p2p_port)), id: this_peer_id, }, )) .await .map_err(|err| anyhow!("Failed to add peer to known: {err}"))? .into_inner(); Ok(all_peers) } // Re-attach peer to the consensus: // Notifies the cluster(any node) that this node changed its address async fn recover( state_ref: &ConsensusStateRef, uri: Option<String>, p2p_port: u16, config: &ConsensusConfig, tls_config: Option<ClientTlsConfig>, ) -> anyhow::Result<()> { let this_peer_id = state_ref.this_peer_id(); let mut peer_to_uri = state_ref .persistent .read() .peer_address_by_id .read() .clone(); let this_peer_url = peer_to_uri.remove(&this_peer_id); // Recover url if a different one is provided let do_recover = match (&this_peer_url, &uri) { (Some(this_peer_url), Some(uri)) => this_peer_url != &Uri::from_str(uri)?, _ => false, }; if do_recover { let mut tries = RECOVERY_MAX_RETRY_COUNT; while tries > 0 { // Try to inform any peer about the change of address for (peer_id, peer_uri) in &peer_to_uri { let res = Self::add_peer_to_known_for( this_peer_id, peer_uri.clone(), uri.clone(), p2p_port, config, tls_config.clone(), ) .await; if res.is_err() { log::warn!( "Failed to recover from peer with id {peer_id} at {peer_uri} with error {res:?}, trying others" ); } else { log::debug!( "Successfully recovered from peer with id {peer_id} at {peer_uri}" ); return Ok(()); } } tries -= 1; log::warn!( "Retrying recovering from known peers (retry {})", RECOVERY_MAX_RETRY_COUNT - tries ); let exp_timeout = RECOVERY_RETRY_TIMEOUT * (RECOVERY_MAX_RETRY_COUNT - tries) as u32; sleep(exp_timeout).await; } return Err(anyhow::anyhow!("Failed to recover from any known peers")); } Ok(()) } /// Add node sequence: /// /// 1. Add current node as a learner /// 2. Start applying entries from consensus /// 3. Eventually leader submits the promotion proposal /// 4. Learners become voters once they read about the promotion from consensus log async fn bootstrap( state_ref: &ConsensusStateRef, bootstrap_peer: Uri, uri: Option<String>, p2p_port: u16, config: &ConsensusConfig, tls_config: Option<ClientTlsConfig>, ) -> anyhow::Result<()> { let this_peer_id = state_ref.this_peer_id(); let all_peers = Self::add_peer_to_known_for( this_peer_id, bootstrap_peer, uri.clone(), p2p_port, config, tls_config, ) .await?; // Although peer addresses are synchronized with consensus, addresses need to be pre-fetched in the case of a new peer // or it will not know how to answer the Raft leader for peer in all_peers.all_peers { state_ref .add_peer( peer.id, peer.uri .parse() .context(format!("Failed to parse peer URI: {}", peer.uri))?, ) .map_err(|err| anyhow!("Failed to add peer: {err}"))? } // Only first peer has itself as a voter in the initial conf state. // This needs to be propagated manually to other peers as it is not contained in any log entry. // So we skip the learner phase for the first peer. state_ref.set_first_voter(all_peers.first_peer_id)?; state_ref.set_conf_state(ConfState::from((vec![all_peers.first_peer_id], vec![])))?; Ok(()) } pub fn start(&mut self) -> anyhow::Result<()> { // If this is the only peer in the cluster, tick Raft node a few times to instantly // self-elect itself as Raft leader if self.node.store().peer_count() == 1 { while !self.node.has_ready() { self.node.tick(); } } // If this is the origin peer of the cluster, try to add origin peer to consensus if let Err(err) = self.try_add_origin() { log::error!("Failed to add origin peer to consensus: {err}"); } let tick_period = Duration::from_millis(self.config.tick_period_ms); let mut previous_tick = Instant::now(); let mut idle_cycles = 0_usize; loop { // Wait (for up to `tick_period`) for incoming client requests and Raft messages let raft_messages = self.advance_node(tick_period)?; // Calculate how many ticks passed since the last one let elapsed_ticks = previous_tick.elapsed().div_duration_f32(tick_period) as u32; // Update previous tick timestamp previous_tick += tick_period * elapsed_ticks; // Calculate how many ticks we should *report* to Raft node. // // If last iteration of the loop took too long to complete, and we report all elapsed // ticks to Raft node, it might trigger unnecessary leader election. // // To prevent this, we check if we received new Raft messages (i.e., we are still // connected to Raft leader), and cap how many ticks we report to Raft node. // // By default, election is triggered if no Raft messages were received for 20 ticks, // so we report at most 15 ticks. // // See https://docs.rs/raft/latest/raft/struct.Config.html#structfield.election_tick. let report_ticks = if raft_messages > 0 { // Default `election_tick` is 20, so expected value here is 15 let max_elapsed_ticks = cmp::max(1, self.raft_config.election_tick.saturating_sub(5)); cmp::min(elapsed_ticks, max_elapsed_ticks as u32) } else { elapsed_ticks }; // Report elapsed ticks to Raft node for _ in 0..report_ticks { self.node.tick(); } // Append new entries to the WAL, apply committed entries, etc... let (stop_consensus, is_idle) = self.on_ready()?; if stop_consensus { return Ok(()); } // If we only sent outgoing Raft messages, but did not change any state during `on_ready`, // we consider Raft node to be "idle" if is_idle { // If current node is the only peer in the cluster, or if we received new Raft messages // (i.e., we are still connected to Raft leader/peers), and Raft node is idle, // count "idle cycle" if raft_messages > 0 || (self.is_single_peer() && self.is_leader()) { idle_cycles += 1; } } else { // If Raft state was updated, reset idle cycle counter idle_cycles = 0; } // If Raft node was idle for 3 cycles, try to sync local state to consensus if idle_cycles >= 3 { self.try_sync_local_state()?; } } } fn advance_node(&mut self, tick_period: Duration) -> anyhow::Result<usize> { if self .try_promote_learner() .context("failed to promote learner")? { return Ok(0); } // This method propagates incoming client requests and Raft messages to Raft node // It's more efficient to process multiple events, so we propagate up to 128 events at a time const RAFT_BATCH_SIZE: usize = 128; // We have to tick Raft node periodically, so we can wait for new events for up to `tick_period` let hard_timeout_at = Instant::now() + tick_period; // We also want to react to new events as quickly as possible, so we only wait for `tick_period / 10` // for any consecutive events after the first one let consecutive_message_timeout = tick_period / 10; // Timeout to wait for the *next* event let mut timeout_at = hard_timeout_at; // Track how many *events* we received... let mut events = 0; // ...and how many of these events were *Raft messages* let mut raft_messages = 0; loop { let Ok(message) = self.recv_update(timeout_at) else { break; }; // When we discover conf-change request, we have to break early and process it ASAP, // because Raft node allows to process single conf-change request at a time. // // E.g., without this condition, if two nodes try to join cluster at the same time and // both conf-change requests are processed in the same batch, the second request would // be ignored and the node would fail to join. let is_conf_change = matches!( message, Message::FromClient( ConsensusOperations::AddPeer { .. } | ConsensusOperations::RemovePeer(_) ), ); let is_raft_message = matches!(message, Message::FromPeer(_)); if let Err(err) = self.advance_node_impl(message) { log::warn!("{err}"); continue; } timeout_at = cmp::min( hard_timeout_at, Instant::now() + consecutive_message_timeout, ); events += 1; raft_messages += usize::from(is_raft_message); if events >= RAFT_BATCH_SIZE || is_conf_change { break; } } Ok(raft_messages) } fn recv_update(&mut self, timeout_at: Instant) -> Result<Message, TryRecvUpdateError> { self.runtime.block_on(async { tokio::select! { biased; message = self.receiver.recv() => message.ok_or(TryRecvUpdateError::Closed), _ = tokio::time::sleep_until(timeout_at.into()) => Err(TryRecvUpdateError::Timeout), } }) } fn advance_node_impl(&mut self, message: Message) -> anyhow::Result<()> { match message { Message::FromClient(ConsensusOperations::AddPeer { peer_id, uri }) => { let existing_uris = self .broker .consensus_state .peer_address_by_id() .into_iter() .map(|(peer_id, url)| (url, peer_id)) .collect::<HashMap<_, _>>(); // Don't allow a peer URI to join if already in consensus // - new URIs can always join // - existing URIs can re-join with the same peer ID // See: <https://github.com/qdrant/qdrant/pull/7375> if let Some(registered_peer_id) = existing_uris.get(&uri.parse::<Uri>().context("peer URI is not a valid URI")?) && registered_peer_id != &peer_id { log::warn!( "Rejected peer {peer_id} to join consensus, URI is already registered by peer {registered_peer_id} ({uri})", ); return Err(anyhow!( "peer URI {uri} already used by peer {registered_peer_id}, remove it first or use a different URI", )); } let mut change = ConfChangeV2::default(); change.set_changes(vec![raft_proto::new_conf_change_single( peer_id, ConfChangeType::AddLearnerNode, )]); log::debug!("Proposing network configuration change: {change:?}"); self.node .propose_conf_change(uri.into_bytes(), change) .context("failed to propose conf change")?; } Message::FromClient(ConsensusOperations::RemovePeer(peer_id)) => { let mut change = ConfChangeV2::default(); change.set_changes(vec![raft_proto::new_conf_change_single( peer_id, ConfChangeType::RemoveNode, )]); log::debug!("Proposing network configuration change: {change:?}"); self.node .propose_conf_change(vec![], change) .context("failed to propose conf change")?; } Message::FromClient(ConsensusOperations::RequestSnapshot) => { self.node .request_snapshot() .context("failed to request snapshot")?; } Message::FromClient(ConsensusOperations::ReportSnapshot { peer_id, status }) => { self.node.report_snapshot(peer_id, status.into()); } Message::FromClient(operation) => { let data = serde_cbor::to_vec(&operation).context("failed to serialize operation")?; log::trace!("Proposing entry from client with length: {}", data.len()); self.node .propose(vec![], data) .context("failed to propose entry")?; } Message::FromPeer(message) => { let is_heartbeat = matches!( message.get_msg_type(), MessageType::MsgHeartbeat | MessageType::MsgHeartbeatResponse, ); if !is_heartbeat { log::trace!( "Received a message from peer with progress: {:?}. Message: {:?}", self.node.raft.prs().get(message.from), message, ); } self.node.step(*message).context("failed to step message")?; } } Ok(()) } fn is_single_peer(&self) -> bool { self.node.store().peer_count() == 1 } fn is_leader(&self) -> bool { self.node.status().ss.raft_state == raft::StateRole::Leader } fn try_sync_local_state(&self) -> anyhow::Result<()> { if !self.node.has_ready() { // No updates to process let store = self.node.store(); let pending_operations = store.persistent.read().unapplied_entities_count(); if pending_operations == 0 && store.is_leader_established.check_ready() { // If leader is established and there is nothing else to do on this iteration, // then we can check if there are any un-synchronized local state left. store.sync_local_state()?; } } Ok(()) } /// Tries to propose "origin peer" (the very first peer, that starts new cluster) to consensus fn try_add_origin(&mut self) -> Result<bool, TryAddOriginError> { // We can determine origin peer from consensus state: // - it should be the only peer in the cluster // - and its commit index should be at 0 or 1 // // When we add a new node to existing cluster, we have to bootstrap it from existing cluster // node, and during bootstrap we explicitly add all current peers to consensus state. So, // *all* peers added to the cluster after the origin will always have at least two peers. // // When origin peer starts new cluster, it self-elects itself as a leader and commits empty // operation with index 1. It is impossible to commit anything to consensus before this // operation is committed. And to add another (second/third/etc) peer to the cluster, we // have to commit a conf-change operation. Which means that only origin peer can ever be at // commit index 0 or 1. // Check that we are the only peer in the cluster if self.node.store().peer_count() > 1 { return Ok(false); } let status = self.node.status(); // Check that we are at index 0 or 1 if status.hs.commit > 1 { return Ok(false); } // If we reached this point, we are the origin peer, but it's impossible to propose anything // to consensus, before leader is elected (`propose_conf_change` will return an error), // so we have to wait for a few ticks for self-election if status.ss.raft_state != StateRole::Leader { return Err(TryAddOriginError::NotLeader); } // Propose origin peer to consensus let mut change = ConfChangeV2::default(); change.set_changes(vec![raft_proto::new_conf_change_single( status.id, ConfChangeType::AddNode, )]); let peer_uri = self .node .store() .persistent .read() .peer_address_by_id .read() .get(&status.id) .ok_or_else(|| TryAddOriginError::UriNotFound)? .to_string(); self.node.propose_conf_change(peer_uri.into(), change)?; Ok(true) } /// Returns `true` if learner promotion was proposed, `false` otherwise. /// Learner node does not vote on elections, cause it might not have a big picture yet. /// So consensus should guarantee that learners are promoted one-by-one. /// Promotions are done by leader and only after it has no pending entries, /// that guarantees that learner will start voting only after it applies all the changes in the log fn try_promote_learner(&mut self) -> anyhow::Result<bool> { // Promote only if leader if self.node.status().ss.raft_state != StateRole::Leader { return Ok(false); } // Promote only when there are no uncommitted changes. let store = self.node.store(); let commit = store.hard_state().commit; let last_log_entry = store.last_index()?; if commit != last_log_entry { return Ok(false); } let Some(learner) = self.find_learner_to_promote() else { return Ok(false); }; log::debug!("Proposing promotion for learner {learner} to voter"); let mut change = ConfChangeV2::default(); change.set_changes(vec![raft_proto::new_conf_change_single( learner, ConfChangeType::AddNode, )]); self.node.propose_conf_change(vec![], change)?; Ok(true) } fn find_learner_to_promote(&self) -> Option<u64> { let commit = self.node.store().hard_state().commit; let learners: HashSet<_> = self .node .store() .conf_state() .learners .into_iter() .collect(); let status = self.node.status(); status .progress? .iter()
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
true
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/schema_generator.rs
src/schema_generator.rs
#![allow(dead_code)] use api::rest::models::{CollectionsResponse, ShardKeysResponse, Usage, VersionInfo}; use api::rest::schema::PointInsertOperations; use api::rest::{ FacetRequest, FacetResponse, QueryGroupsRequest, QueryRequest, QueryRequestBatch, QueryResponse, Record, ScoredPoint, SearchMatrixOffsetsResponse, SearchMatrixPairsResponse, SearchMatrixRequest, UpdateVectors, }; use collection::operations::cluster_ops::ClusterOperations; use collection::operations::consistency_params::ReadConsistency; use collection::operations::payload_ops::{DeletePayload, SetPayload}; use collection::operations::point_ops::{PointsSelector, WriteOrdering}; use collection::operations::snapshot_ops::{ ShardSnapshotRecover, SnapshotDescription, SnapshotRecover, }; use collection::operations::types::{ AliasDescription, CollectionClusterInfo, CollectionExistence, CollectionInfo, CollectionsAliasesResponse, CountRequest, CountResult, DiscoverRequest, DiscoverRequestBatch, GroupsResult, OptimizationsResponse, PointGroup, PointRequest, RecommendGroupsRequest, RecommendRequest, RecommendRequestBatch, ScrollRequest, ScrollResult, SearchGroupsRequest, SearchRequest, SearchRequestBatch, UpdateResult, }; use collection::operations::vector_ops::DeleteVectors; use schemars::JsonSchema; use schemars::r#gen::SchemaSettings; use serde::Serialize; use storage::content_manager::collection_meta_ops::{ ChangeAliasesOperation, CreateCollection, UpdateCollection, }; use storage::types::ClusterStatus; use crate::common::telemetry::TelemetryData; use crate::common::update::{CreateFieldIndex, UpdateOperations}; mod actix; mod common; mod settings; mod tracing; #[derive(Serialize, JsonSchema)] struct AllDefinitions { a1: CollectionsResponse, a2: CollectionInfo, // a3: CollectionMetaOperations, a4: PointRequest, a5: Record, a6: SearchRequest, a7: ScoredPoint, a8: UpdateResult, // a9: CollectionUpdateOperations, aa: RecommendRequest, ab: ScrollRequest, ac: ScrollResult, ad: CreateCollection, ae: UpdateCollection, af: ChangeAliasesOperation, ag: CreateFieldIndex, ah: PointsSelector, ai: PointInsertOperations, aj: SetPayload, ak: DeletePayload, al: ClusterStatus, am: SnapshotDescription, an: CountRequest, ao: CountResult, ap: CollectionClusterInfo, aq: TelemetryData, ar: ClusterOperations, at: SearchRequestBatch, au: RecommendRequestBatch, aw: SnapshotRecover, ax: CollectionsAliasesResponse, ay: AliasDescription, az: WriteOrdering, b1: ReadConsistency, b2: UpdateVectors, b3: DeleteVectors, b4: PointGroup, b5: SearchGroupsRequest, b6: RecommendGroupsRequest, b7: GroupsResult, b8: UpdateOperations, b9: ShardSnapshotRecover, ba: DiscoverRequest, bb: DiscoverRequestBatch, bc: VersionInfo, bd: CollectionExistence, be: QueryRequest, bf: QueryRequestBatch, bg: QueryResponse, bh: QueryGroupsRequest, bi: SearchMatrixRequest, bj: SearchMatrixOffsetsResponse, bk: SearchMatrixPairsResponse, bl: FacetRequest, bm: FacetResponse, bn: Usage, bo: ShardKeysResponse, bp: OptimizationsResponse, } fn save_schema<T: JsonSchema>() { let settings = SchemaSettings::draft07(); let generator = settings.into_generator(); let schema = generator.into_root_schema_for::<T>(); let schema_str = serde_json::to_string_pretty(&schema).unwrap(); println!("{schema_str}") } fn main() { save_schema::<AllDefinitions>(); }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tracing/config.rs
src/tracing/config.rs
use std::collections::HashSet; use serde::{Deserialize, Serialize}; use tracing_subscriber::fmt; use super::*; #[derive(Clone, Debug, Default, Eq, PartialEq, Deserialize, Serialize)] #[serde(default)] pub struct LoggerConfig { #[serde(flatten)] pub default: default::Config, #[serde(default)] pub on_disk: on_disk::Config, } impl LoggerConfig { pub fn with_top_level_directive(&self, log_level: Option<String>) -> Self { let mut logger_config = self.clone(); if logger_config.default.log_level.is_some() && log_level.is_some() { eprintln!( "Both top-level `log_level` and `logger.log_level` config directives are used. \ `logger.log_level` takes priority, so top-level `log_level` will be ignored." ); } if logger_config.default.log_level.is_none() { logger_config.default.log_level = log_level; } logger_config } pub fn merge(&mut self, other: Self) { self.default.merge(other.default); self.on_disk.merge(other.on_disk); } } #[derive(Default, Clone, Debug, Eq, PartialEq, Deserialize, Serialize)] #[serde(rename_all = "snake_case")] pub enum LogFormat { #[default] Text, Json, } #[derive(Copy, Clone, Debug, PartialEq, Eq, Deserialize, Serialize, Hash)] #[serde(rename_all = "snake_case")] pub enum SpanEvent { New, Enter, Exit, Close, } impl SpanEvent { pub fn unwrap_or_default_config(events: &Option<HashSet<Self>>) -> fmt::format::FmtSpan { Self::into_fmt_span(events.as_ref().unwrap_or(&HashSet::new()).iter().copied()) } pub fn into_fmt_span(events: impl IntoIterator<Item = Self>) -> fmt::format::FmtSpan { events .into_iter() .fold(fmt::format::FmtSpan::NONE, |events, event| { events | event.into() }) } } impl From<SpanEvent> for fmt::format::FmtSpan { fn from(event: SpanEvent) -> Self { match event { SpanEvent::New => fmt::format::FmtSpan::NEW, SpanEvent::Enter => fmt::format::FmtSpan::ENTER, SpanEvent::Exit => fmt::format::FmtSpan::EXIT, SpanEvent::Close => fmt::format::FmtSpan::CLOSE, } } } #[derive(Copy, Clone, Debug, Eq, PartialEq, Deserialize, Serialize, Default)] #[serde(rename_all = "snake_case")] pub enum Color { #[default] Auto, #[serde(untagged)] Explicit(bool), } impl Color { pub fn to_bool(self) -> bool { match self { Self::Auto => colored::control::SHOULD_COLORIZE.should_colorize(), Self::Explicit(bool) => bool, } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tracing/test.rs
src/tracing/test.rs
use std::collections::HashSet; use serde_json::json; use super::*; #[test] fn deserialize_logger_config() { let json = json!({ "log_level": "debug", "span_events": ["new", "close"], "color": true, "on_disk": { "enabled": true, "log_file": "/logs/qdrant", "log_level": "tracing", "span_events": ["new", "close"], "buffer_size_bytes": 1024, } }); let config = deserialize_config(json); let expected = LoggerConfig { default: default::Config { log_level: Some("debug".into()), span_events: Some(HashSet::from([ config::SpanEvent::New, config::SpanEvent::Close, ])), format: None, color: Some(config::Color::Explicit(true)), }, on_disk: on_disk::Config { enabled: Some(true), log_file: Some("/logs/qdrant".into()), log_level: Some("tracing".into()), span_events: Some(HashSet::from([ config::SpanEvent::New, config::SpanEvent::Close, ])), format: None, buffer_size_bytes: Some(1024), }, }; assert_eq!(config, expected); } #[test] fn deserialize_json_logger_config() { let json = json!({ "log_level": "debug", "span_events": ["new", "close"], "format": "json", "color": true, "on_disk": { "enabled": true, "log_file": "/logs/qdrant", "log_level": "tracing", "span_events": ["new", "close"], "format": "text", "buffer_size_bytes": 1024, } }); let config = deserialize_config(json); let expected = LoggerConfig { default: default::Config { log_level: Some("debug".into()), span_events: Some(HashSet::from([ config::SpanEvent::New, config::SpanEvent::Close, ])), format: Some(config::LogFormat::Json), color: Some(config::Color::Explicit(true)), }, on_disk: on_disk::Config { enabled: Some(true), log_file: Some("/logs/qdrant".into()), log_level: Some("tracing".into()), span_events: Some(HashSet::from([ config::SpanEvent::New, config::SpanEvent::Close, ])), format: Some(config::LogFormat::Text), buffer_size_bytes: Some(1024), }, }; assert_eq!(config, expected); } #[test] fn deserialize_empty_config() { let config = deserialize_config(json!({})); assert_eq!(config, LoggerConfig::default()); } #[test] fn deserialize_config_with_empty_on_disk() { let config = deserialize_config(json!({ "on_disk": {} })); assert_eq!(config, LoggerConfig::default()); } #[test] fn deseriailze_config_with_explicit_nulls() { let json = json!({ "log_level": null, "span_events": null, "format": null, "color": null, "on_disk": { "enabled": null, "log_file": null, "log_level": null, "span_events": null, "format": null, "buffer_size_bytes": null, } }); let config = deserialize_config(json); assert_eq!(config, LoggerConfig::default()); } fn deserialize_config(json: serde_json::Value) -> LoggerConfig { serde_json::from_value(json).unwrap() }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tracing/on_disk.rs
src/tracing/on_disk.rs
use std::collections::HashSet; use std::io; use std::sync::Mutex; use anyhow::Context as _; use common::ext::OptionExt; use fs_err as fs; use serde::{Deserialize, Serialize}; use tracing_subscriber::{Layer, fmt, registry}; use super::*; #[derive(Clone, Debug, Default, Eq, PartialEq, Deserialize, Serialize)] #[serde(default)] pub struct Config { pub enabled: Option<bool>, pub log_file: Option<String>, pub log_level: Option<String>, pub format: Option<config::LogFormat>, pub span_events: Option<HashSet<config::SpanEvent>>, pub buffer_size_bytes: Option<usize>, } impl Config { pub fn merge(&mut self, other: Self) { let Self { enabled, log_file, log_level, span_events, format, buffer_size_bytes, } = other; self.enabled.replace_if_some(enabled); self.log_file.replace_if_some(log_file); self.log_level.replace_if_some(log_level); self.span_events.replace_if_some(span_events); self.format.replace_if_some(format); self.buffer_size_bytes.replace_if_some(buffer_size_bytes); } } pub fn new_logger<S>(config: &mut Config) -> Logger<S> where S: tracing::Subscriber + for<'span> registry::LookupSpan<'span>, { let layer = match new_layer(config) { Ok(layer) => layer, Err(err) => { eprintln!( "failed to enable logging into {} log-file: {err}", config.log_file.as_deref().unwrap_or(""), ); config.enabled = Some(false); None } }; let filter = new_filter(config); layer.with_filter(filter) } pub fn new_layer<S>(config: &Config) -> anyhow::Result<Option<Box<dyn Layer<S> + Send + Sync>>> where S: tracing::Subscriber + for<'span> registry::LookupSpan<'span>, { if !config.enabled.unwrap_or_default() { return Ok(None); } let Some(log_file) = &config.log_file else { return Err(anyhow::format_err!( "log file is not specified (it can only be specified in the config file)" )); }; let writer = fs::OpenOptions::new() .create(true) .append(true) .open(log_file) .with_context(|| format!("failed to open log file {log_file}"))?; let layer = fmt::Layer::default() .with_writer(Mutex::new(io::BufWriter::with_capacity( config.buffer_size_bytes.unwrap_or(8192), writer, ))) .with_span_events(config::SpanEvent::unwrap_or_default_config( &config.span_events, )) .with_ansi(false); let layer = match config.format { None | Some(config::LogFormat::Text) => Box::new(layer) as _, Some(config::LogFormat::Json) => Box::new(layer.json()) as _, }; Ok(Some(layer)) } pub fn new_filter(config: &Config) -> filter::EnvFilter { filter(config.log_level.as_deref().unwrap_or("")) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tracing/default.rs
src/tracing/default.rs
use std::collections::HashSet; use common::ext::OptionExt; use serde::{Deserialize, Serialize}; use tracing_subscriber::{Layer, fmt, registry}; use super::*; #[derive(Clone, Debug, Default, Eq, PartialEq, Deserialize, Serialize)] #[serde(default)] pub struct Config { pub log_level: Option<String>, pub span_events: Option<HashSet<config::SpanEvent>>, pub format: Option<config::LogFormat>, pub color: Option<config::Color>, } impl Config { pub fn merge(&mut self, other: Self) { let Self { log_level, span_events, format, color, } = other; self.log_level.replace_if_some(log_level); self.span_events.replace_if_some(span_events); self.format.replace_if_some(format); self.color.replace_if_some(color); } } pub fn new_logger<S>(config: &Config) -> Logger<S> where S: tracing::Subscriber + for<'span> registry::LookupSpan<'span>, { let layer = new_layer(config); let filter = new_filter(config); Some(layer).with_filter(filter) } pub fn new_layer<S>(config: &Config) -> Box<dyn Layer<S> + Send + Sync> where S: tracing::Subscriber + for<'span> registry::LookupSpan<'span>, { let layer = fmt::Layer::default() .with_span_events(config::SpanEvent::unwrap_or_default_config( &config.span_events, )) .with_ansi(config.color.unwrap_or_default().to_bool()); match config.format { None | Some(config::LogFormat::Text) => Box::new(layer), Some(config::LogFormat::Json) => Box::new(layer.json()), } } pub fn new_filter(config: &Config) -> filter::EnvFilter { filter(config.log_level.as_deref().unwrap_or("")) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tracing/mod.rs
src/tracing/mod.rs
pub mod config; pub mod default; pub mod handle; pub mod on_disk; #[cfg(test)] mod test; use std::fmt::Write as _; use std::str::FromStr as _; use tracing_subscriber::prelude::*; use tracing_subscriber::{filter, reload}; pub use self::config::LoggerConfig; pub use self::handle::LoggerHandle; const DEFAULT_LOG_LEVEL: log::LevelFilter = log::LevelFilter::Info; const DEFAULT_FILTERS: &[(&str, log::LevelFilter)] = &[ ("hyper", log::LevelFilter::Info), ("h2", log::LevelFilter::Error), ("tower", log::LevelFilter::Warn), ("rustls", log::LevelFilter::Info), ("wal", log::LevelFilter::Warn), ("raft", log::LevelFilter::Warn), ]; pub fn setup(mut config: config::LoggerConfig) -> anyhow::Result<LoggerHandle> { // Note that on-disk logger *have* to be initialized *before* default logger! // // If default logger is initialized before on-disk logger, then ANSI escape-sequences (that are // used to apply color and formatting in the terminal, but looks like corrupted text in the text // editor) might appear in the on-disk log-file. // // This happens because when multiple `fmt::Layer`s are initialized in the same subscriber, // the top-level `fmt::Layer` would cache pre-formatted fragments of the log-line // for the next `fmt::Layer`s to reuse. // // And default logger outputs colored log-lines, which on-disk logger reuse even if colors are // disabled for the on-disk logger. :/ let on_disk_logger = on_disk::new_logger(&mut config.on_disk); let (on_disk_logger, on_disk_logger_handle) = reload::Layer::new(on_disk_logger); let reg = tracing_subscriber::registry().with(on_disk_logger); let default_logger = default::new_logger(&config.default); let (default_logger, default_logger_handle) = reload::Layer::new(default_logger); let reg = reg.with(default_logger); let logger_handle = LoggerHandle::new(config, default_logger_handle, on_disk_logger_handle); // Use `console` or `console-subscriber` feature to enable `console-subscriber` // // Note, that `console-subscriber` requires manually enabling // `--cfg tokio_unstable` rust flags during compilation! // // Otherwise `console_subscriber::spawn` call panics! // // See https://docs.rs/tokio/latest/tokio/#unstable-features #[cfg(all(feature = "console-subscriber", tokio_unstable))] let reg = reg.with(console_subscriber::spawn()); #[cfg(all(feature = "console-subscriber", not(tokio_unstable)))] eprintln!( "`console-subscriber` requires manually enabling \ `--cfg tokio_unstable` rust flags during compilation!" ); // Use `tracy` or `tracing-tracy` feature to enable `tracing-tracy` #[cfg(feature = "tracing-tracy")] let reg = reg.with( tracing_tracy::TracyLayer::new(tracing_tracy::DefaultConfig::default()).with_filter( tracing_subscriber::filter::filter_fn(|metadata| metadata.is_span()), ), ); tracing::subscriber::set_global_default(reg)?; tracing_log::LogTracer::init()?; Ok(logger_handle) } fn filter(user_filters: &str) -> filter::EnvFilter { let mut filter = String::new(); let user_log_level = user_filters .rsplit(',') .find_map(|dir| log::LevelFilter::from_str(dir).ok()); if user_log_level.is_none() { write!(&mut filter, "{DEFAULT_LOG_LEVEL}").unwrap(); // Writing into `String` never fails } for &(target, log_level) in DEFAULT_FILTERS { if user_log_level.unwrap_or(DEFAULT_LOG_LEVEL) > log_level { let comma = if filter.is_empty() { "" } else { "," }; write!(&mut filter, "{comma}{target}={log_level}").unwrap(); // Writing into `String` never fails } } let comma = if filter.is_empty() { "" } else { "," }; write!(&mut filter, "{comma}{user_filters}").unwrap(); // Writing into `String` never fails filter::EnvFilter::builder() .with_regex(false) .parse_lossy(filter) } #[rustfmt::skip] // `rustfmt` formats this into unreadable single line type Logger<S> = filter::Filtered< Option<Box<dyn tracing_subscriber::Layer<S> + Send + Sync>>, filter::EnvFilter, S, >;
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tracing/handle.rs
src/tracing/handle.rs
use std::sync::Arc; use tokio::sync::RwLock; use tracing_subscriber::{Registry, layer, reload}; use super::*; #[derive(Clone)] pub struct LoggerHandle { config: Arc<RwLock<config::LoggerConfig>>, default: DefaultLoggerReloadHandle, on_disk: OnDiskLoggerReloadHandle, } #[rustfmt::skip] // `rustfmt` formats this into unreadable single line type DefaultLoggerReloadHandle<S = DefaultLoggerSubscriber> = reload::Handle< Logger<S>, S, >; #[rustfmt::skip] // `rustfmt` formats this into unreadable single line type DefaultLoggerSubscriber<S = Registry> = layer::Layered< reload::Layer<Logger<S>, S>, S, >; #[rustfmt::skip] // `rustfmt` formats this into unreadable single line type OnDiskLoggerReloadHandle<S = Registry> = reload::Handle< Logger<S>, S, >; impl LoggerHandle { pub fn new( config: config::LoggerConfig, default: DefaultLoggerReloadHandle, on_disk: OnDiskLoggerReloadHandle, ) -> Self { Self { config: Arc::new(RwLock::new(config)), default, on_disk, } } pub async fn get_config(&self) -> config::LoggerConfig { self.config.read().await.clone() } pub async fn update_config(&self, new_config: config::LoggerConfig) -> anyhow::Result<()> { let mut config = self.config.write().await; // `tracing-subscriber` does not support `reload`ing `Filtered` layers, so we *have to* use // `modify`. However, `modify` would *deadlock* if provided closure logs anything or produce // any `tracing` event. // // So, we structure `update_config` to only do an absolute minimum of changes and only use // the most trivial operations during `modify`, to guarantee we won't deadlock. // // See: // - https://docs.rs/tracing-subscriber/latest/tracing_subscriber/reload/struct.Handle.html#method.reload // - https://github.com/tokio-rs/tracing/issues/1629 // - https://github.com/tokio-rs/tracing/pull/2657 let mut merged_config = config.clone(); merged_config.merge(new_config); if merged_config.on_disk != config.on_disk { let new_layer = on_disk::new_layer(&merged_config.on_disk)?; let new_filter = on_disk::new_filter(&merged_config.on_disk); self.on_disk.modify(move |logger| { *logger.inner_mut() = new_layer; *logger.filter_mut() = new_filter; })?; config.on_disk = merged_config.on_disk; } if merged_config.default != config.default { let new_layer = default::new_layer(&merged_config.default); let new_filter = default::new_filter(&merged_config.default); self.default.modify(|logger| { *logger.inner_mut() = Some(new_layer); *logger.filter_mut() = new_filter; })?; config.default = merged_config.default; } Ok(()) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/tonic_telemetry.rs
src/tonic/tonic_telemetry.rs
use std::sync::Arc; use std::task::{Context, Poll}; use futures_util::future::BoxFuture; use tower::Service; use tower_layer::Layer; use crate::common::telemetry_ops::requests_telemetry::{ TonicTelemetryCollector, TonicWorkerTelemetryCollector, }; #[derive(Clone)] pub struct TonicTelemetryService<T> { service: T, telemetry_data: Arc<parking_lot::Mutex<TonicWorkerTelemetryCollector>>, } #[derive(Clone)] pub struct TonicTelemetryLayer { telemetry_collector: Arc<parking_lot::Mutex<TonicTelemetryCollector>>, } impl<S> Service<tonic::codegen::http::Request<tonic::transport::Body>> for TonicTelemetryService<S> where S: Service<tonic::codegen::http::Request<tonic::transport::Body>>, S::Future: Send + 'static, { type Response = S::Response; type Error = S::Error; type Future = BoxFuture<'static, Result<S::Response, S::Error>>; fn poll_ready(&mut self, cx: &mut Context<'_>) -> Poll<Result<(), Self::Error>> { self.service.poll_ready(cx) } fn call( &mut self, request: tonic::codegen::http::Request<tonic::transport::Body>, ) -> Self::Future { let method_name = request.uri().path().to_string(); let future = self.service.call(request); let telemetry_data = self.telemetry_data.clone(); Box::pin(async move { let instant = std::time::Instant::now(); let response = future.await?; telemetry_data.lock().add_response(method_name, instant); Ok(response) }) } } impl TonicTelemetryLayer { pub fn new( telemetry_collector: Arc<parking_lot::Mutex<TonicTelemetryCollector>>, ) -> TonicTelemetryLayer { Self { telemetry_collector, } } } impl<S> Layer<S> for TonicTelemetryLayer { type Service = TonicTelemetryService<S>; fn layer(&self, service: S) -> Self::Service { TonicTelemetryService { service, telemetry_data: self .telemetry_collector .lock() .create_grpc_telemetry_collector(), } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/auth.rs
src/tonic/auth.rs
use std::sync::Arc; use std::task::{Context, Poll}; use futures::future::BoxFuture; use storage::rbac::Access; use tonic::Status; use tonic::body::BoxBody; use tower::{Layer, Service}; use crate::common::auth::{AuthError, AuthKeys}; type Request = tonic::codegen::http::Request<tonic::transport::Body>; type Response = tonic::codegen::http::Response<BoxBody>; #[derive(Clone)] pub struct AuthMiddleware<S> { auth_keys: Arc<AuthKeys>, service: S, } async fn check(auth_keys: Arc<AuthKeys>, mut req: Request) -> Result<Request, Status> { // Allow health check endpoints to bypass authentication let path = req.uri().path(); if path == "/qdrant.Qdrant/HealthCheck" || path == "/grpc.health.v1.Health/Check" { // Set default full access for health check endpoints let access = Access::full("Health check endpoints have full access without authentication"); let inference_token = crate::common::inference::token::InferenceToken(None); req.extensions_mut().insert::<Access>(access); req.extensions_mut().insert(inference_token); return Ok(req); } let (access, inference_token) = auth_keys .validate_request(|key| req.headers().get(key).and_then(|val| val.to_str().ok())) .await .map_err(|e| match e { AuthError::Unauthorized(e) => Status::unauthenticated(e), AuthError::Forbidden(e) => Status::permission_denied(e), AuthError::StorageError(e) => Status::from(e), })?; let previous = req.extensions_mut().insert::<Access>(access); debug_assert!( previous.is_none(), "Previous access object should not exist in the request" ); let previous_token = req.extensions_mut().insert(inference_token); debug_assert!( previous_token.is_none(), "Previous inference token should not exist in the request" ); Ok(req) } impl<S> Service<Request> for AuthMiddleware<S> where S: Service<Request, Response = Response> + Clone + Send + 'static, S::Future: Send + 'static, { type Response = S::Response; type Error = S::Error; type Future = BoxFuture<'static, Result<Self::Response, S::Error>>; fn poll_ready(&mut self, cx: &mut Context<'_>) -> Poll<Result<(), Self::Error>> { self.service.poll_ready(cx) } fn call(&mut self, request: Request) -> Self::Future { let auth_keys = self.auth_keys.clone(); let mut service = self.service.clone(); Box::pin(async move { match check(auth_keys, request).await { Ok(req) => service.call(req).await, Err(e) => Ok(e.to_http()), } }) } } #[derive(Clone)] pub struct AuthLayer { auth_keys: Arc<AuthKeys>, } impl AuthLayer { pub fn new(auth_keys: AuthKeys) -> Self { Self { auth_keys: Arc::new(auth_keys), } } } impl<S> Layer<S> for AuthLayer { type Service = AuthMiddleware<S>; fn layer(&self, service: S) -> Self::Service { Self::Service { auth_keys: self.auth_keys.clone(), service, } } } pub fn extract_access<R>(req: &mut tonic::Request<R>) -> Access { req.extensions_mut().remove::<Access>().unwrap_or_else(|| { Access::full("All requests have full by default access when API key is not configured") }) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/mod.rs
src/tonic/mod.rs
mod api; mod auth; mod logging; mod tonic_telemetry; use std::io; use std::net::{IpAddr, SocketAddr}; use std::sync::Arc; use std::time::Duration; use ::api::grpc::QDRANT_DESCRIPTOR_SET; use ::api::grpc::grpc_health_v1::health_check_response::ServingStatus; use ::api::grpc::grpc_health_v1::health_server::{Health, HealthServer}; use ::api::grpc::grpc_health_v1::{ HealthCheckRequest as ProtocolHealthCheckRequest, HealthCheckResponse as ProtocolHealthCheckResponse, }; use ::api::grpc::qdrant::collections_internal_server::CollectionsInternalServer; use ::api::grpc::qdrant::collections_server::CollectionsServer; use ::api::grpc::qdrant::points_internal_server::PointsInternalServer; use ::api::grpc::qdrant::points_server::PointsServer; use ::api::grpc::qdrant::qdrant_internal_server::{QdrantInternal, QdrantInternalServer}; use ::api::grpc::qdrant::qdrant_server::{Qdrant, QdrantServer}; use ::api::grpc::qdrant::shard_snapshots_server::ShardSnapshotsServer; use ::api::grpc::qdrant::snapshots_server::SnapshotsServer; use ::api::grpc::qdrant::{ GetConsensusCommitRequest, GetConsensusCommitResponse, GetPeerTelemetryRequest, GetPeerTelemetryResponse, HealthCheckReply, HealthCheckRequest, WaitOnConsensusCommitRequest, WaitOnConsensusCommitResponse, }; use ::api::rest::models::VersionInfo; use collection::operations::verification::new_unchecked_verification_pass; use storage::content_manager::consensus_manager::ConsensusStateRef; use storage::content_manager::toc::TableOfContent; use storage::dispatcher::Dispatcher; use storage::rbac::Access; use tokio::runtime::Handle; use tokio::signal; use tonic::codec::CompressionEncoding; use tonic::transport::{Server, ServerTlsConfig}; use tonic::{Request, Response, Status}; use crate::common::auth::AuthKeys; use crate::common::helpers; use crate::common::http_client::HttpClient; use crate::common::telemetry_ops::requests_telemetry::TonicTelemetryCollector; use crate::settings::Settings; use crate::tonic::api::collections_api::CollectionsService; use crate::tonic::api::collections_internal_api::CollectionsInternalService; use crate::tonic::api::points_api::PointsService; use crate::tonic::api::points_internal_api::PointsInternalService; use crate::tonic::api::snapshots_api::{ShardSnapshotsService, SnapshotsService}; #[derive(Default)] pub struct QdrantService {} #[tonic::async_trait] impl Qdrant for QdrantService { async fn health_check( &self, _request: Request<HealthCheckRequest>, ) -> Result<Response<HealthCheckReply>, Status> { Ok(Response::new(VersionInfo::default().into())) } } // Additional health check service that follows gRPC health check protocol as described in #2614 #[derive(Default)] pub struct HealthService {} #[tonic::async_trait] impl Health for HealthService { async fn check( &self, _request: Request<ProtocolHealthCheckRequest>, ) -> Result<Response<ProtocolHealthCheckResponse>, Status> { let response = ProtocolHealthCheckResponse { status: ServingStatus::Serving as i32, }; Ok(Response::new(response)) } } pub struct QdrantInternalService { /// Qdrant settings settings: Settings, /// Consensus state consensus_state: ConsensusStateRef, } impl QdrantInternalService { fn new(settings: Settings, consensus_state: ConsensusStateRef) -> Self { Self { settings, consensus_state, } } } #[tonic::async_trait] impl QdrantInternal for QdrantInternalService { async fn get_consensus_commit( &self, _: tonic::Request<GetConsensusCommitRequest>, ) -> Result<Response<GetConsensusCommitResponse>, Status> { let persistent = self.consensus_state.persistent.read(); let commit = persistent.state.hard_state.commit as _; let term = persistent.state.hard_state.term as _; Ok(Response::new(GetConsensusCommitResponse { commit, term })) } async fn wait_on_consensus_commit( &self, request: Request<WaitOnConsensusCommitRequest>, ) -> Result<Response<WaitOnConsensusCommitResponse>, Status> { let request = request.into_inner(); let commit = request.commit as u64; let term = request.term as u64; let timeout = Duration::from_secs(request.timeout as u64); let consensus_tick = Duration::from_millis(self.settings.cluster.consensus.tick_period_ms); let ok = self .consensus_state .wait_for_consensus_commit(commit, term, consensus_tick, timeout) .await .is_ok(); Ok(Response::new(WaitOnConsensusCommitResponse { ok })) } async fn get_peer_telemetry( &self, _request: Request<GetPeerTelemetryRequest>, ) -> Result<Response<GetPeerTelemetryResponse>, Status> { Err(Status::unimplemented("Not implemented")) } } #[cfg(not(unix))] async fn wait_stop_signal(for_what: &str) { signal::ctrl_c().await.unwrap(); log::debug!("Stopping {for_what} on SIGINT"); } #[cfg(unix)] async fn wait_stop_signal(for_what: &str) { let mut term = signal::unix::signal(signal::unix::SignalKind::terminate()).unwrap(); let mut inrt = signal::unix::signal(signal::unix::SignalKind::interrupt()).unwrap(); tokio::select! { _ = term.recv() => log::debug!("Stopping {for_what} on SIGTERM"), _ = inrt.recv() => log::debug!("Stopping {for_what} on SIGINT"), } } pub fn init( dispatcher: Arc<Dispatcher>, telemetry_collector: Arc<parking_lot::Mutex<TonicTelemetryCollector>>, settings: Settings, grpc_port: u16, runtime: Handle, ) -> io::Result<()> { runtime.block_on(async { let socket = SocketAddr::from((settings.service.host.parse::<IpAddr>().unwrap(), grpc_port)); let qdrant_service = QdrantService::default(); let health_service = HealthService::default(); let collections_service = CollectionsService::new(dispatcher.clone()); let points_service = PointsService::new(dispatcher.clone(), settings.service.clone()); let snapshot_service = SnapshotsService::new(dispatcher.clone()); // Only advertise the public services. By default, all services in QDRANT_DESCRIPTOR_SET // will be advertised, so explicitly list the services to be included. let reflection_service = tonic_reflection::server::Builder::configure() .register_encoded_file_descriptor_set(QDRANT_DESCRIPTOR_SET) .with_service_name("qdrant.Collections") .with_service_name("qdrant.Points") .with_service_name("qdrant.Snapshots") .with_service_name("qdrant.Qdrant") .with_service_name("grpc.health.v1.Health") .build() .unwrap(); log::info!("Qdrant gRPC listening on {grpc_port}"); let mut server = Server::builder(); if settings.service.enable_tls { log::info!("TLS enabled for gRPC API (TTL not supported)"); let tls_server_config = helpers::load_tls_external_server_config(settings.tls()?)?; server = server .tls_config(tls_server_config) .map_err(helpers::tonic_error_to_io_error)?; } else { log::info!("TLS disabled for gRPC API"); } // The stack of middleware that our service will be wrapped in let middleware_layer = tower::ServiceBuilder::new() .layer(logging::LoggingMiddlewareLayer::new()) .layer(tonic_telemetry::TonicTelemetryLayer::new( telemetry_collector, )) .option_layer({ AuthKeys::try_create( &settings.service, dispatcher .toc( &Access::full("For tonic auth middleware"), &new_unchecked_verification_pass(), ) .clone(), ) .map(auth::AuthLayer::new) }) .into_inner(); server .layer(middleware_layer) .add_service(reflection_service) .add_service( QdrantServer::new(qdrant_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .add_service( CollectionsServer::new(collections_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .add_service( PointsServer::new(points_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .add_service( SnapshotsServer::new(snapshot_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .add_service( HealthServer::new(health_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .serve_with_shutdown(socket, async { wait_stop_signal("gRPC service").await; }) .await .map_err(helpers::tonic_error_to_io_error) })?; Ok(()) } #[allow(clippy::too_many_arguments)] pub fn init_internal( toc: Arc<TableOfContent>, consensus_state: ConsensusStateRef, telemetry_collector: Arc<parking_lot::Mutex<TonicTelemetryCollector>>, settings: Settings, host: String, internal_grpc_port: u16, tls_config: Option<ServerTlsConfig>, to_consensus: tokio::sync::mpsc::Sender<crate::consensus::Message>, runtime: Handle, ) -> std::io::Result<()> { use ::api::grpc::qdrant::raft_server::RaftServer; use crate::tonic::api::raft_api::RaftService; let http_client = HttpClient::from_settings(&settings)?; runtime .block_on(async { let socket = SocketAddr::from((host.parse::<IpAddr>().unwrap(), internal_grpc_port)); let qdrant_service = QdrantService::default(); let points_internal_service = PointsInternalService::new(toc.clone(), settings.service.clone()); let qdrant_internal_service = QdrantInternalService::new(settings, consensus_state.clone()); let collections_internal_service = CollectionsInternalService::new(toc.clone()); let shard_snapshots_service = ShardSnapshotsService::new(toc.clone(), http_client); let raft_service = RaftService::new(to_consensus, consensus_state, tls_config.is_some()); log::debug!("Qdrant internal gRPC listening on {internal_grpc_port}"); let mut server = Server::builder() // Internally use a high limit for pending accept streams. // We can have a huge number of reset/dropped HTTP2 streams in our internal // communication when there are a lot of clients dropping connections. This // internally causes an GOAWAY/ENHANCE_YOUR_CALM error breaking cluster consensus. // We prefer to keep more pending reset streams even though this may be expensive, // versus an internal error that is very hard to handle. // More info: <https://github.com/qdrant/qdrant/issues/1907> .http2_max_pending_accept_reset_streams(Some(1024)); if let Some(config) = tls_config { log::info!("TLS enabled for internal gRPC API (TTL not supported)"); server = server.tls_config(config)?; } else { log::info!("TLS disabled for internal gRPC API"); }; // The stack of middleware that our service will be wrapped in let middleware_layer = tower::ServiceBuilder::new() .layer(logging::LoggingMiddlewareLayer::new()) .layer(tonic_telemetry::TonicTelemetryLayer::new( telemetry_collector, )) .into_inner(); server .layer(middleware_layer) .add_service( QdrantServer::new(qdrant_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .add_service( QdrantInternalServer::new(qdrant_internal_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .add_service( CollectionsInternalServer::new(collections_internal_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .add_service( PointsInternalServer::new(points_internal_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .add_service( ShardSnapshotsServer::new(shard_snapshots_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .add_service( RaftServer::new(raft_service) .send_compressed(CompressionEncoding::Gzip) .accept_compressed(CompressionEncoding::Gzip) .max_decoding_message_size(usize::MAX), ) .serve_with_shutdown(socket, async { wait_stop_signal("internal gRPC").await; }) .await }) .unwrap(); Ok(()) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/logging.rs
src/tonic/logging.rs
use std::task::{Context, Poll}; use futures_util::future::BoxFuture; use tonic::Code; use tonic::body::BoxBody; use tonic::codegen::http::Response; use tower::Service; use tower_layer::Layer; #[derive(Clone)] pub struct LoggingMiddleware<T> { inner: T, } #[derive(Clone)] pub struct LoggingMiddlewareLayer; impl LoggingMiddlewareLayer { pub fn new() -> Self { Self {} } } impl<S> Service<tonic::codegen::http::Request<tonic::transport::Body>> for LoggingMiddleware<S> where S: Service<tonic::codegen::http::Request<tonic::transport::Body>, Response = Response<BoxBody>> + Clone, S::Future: Send + 'static, { type Response = S::Response; type Error = S::Error; type Future = BoxFuture<'static, Result<S::Response, S::Error>>; fn poll_ready(&mut self, cx: &mut Context<'_>) -> Poll<Result<(), Self::Error>> { self.inner.poll_ready(cx) } fn call( &mut self, request: tonic::codegen::http::Request<tonic::transport::Body>, ) -> Self::Future { let clone = self.inner.clone(); let mut inner = std::mem::replace(&mut self.inner, clone); let method_name = request.uri().path().to_string(); let instant = std::time::Instant::now(); let future = inner.call(request); Box::pin(async move { let response = future.await; let elapsed_sec = instant.elapsed().as_secs_f32(); match response { Err(error) => { log::error!("gRPC request error {method_name}"); Err(error) } Ok(response_tonic) => { let grpc_status = tonic::Status::from_header_map(response_tonic.headers()); if let Some(grpc_status) = grpc_status { match grpc_status.code() { Code::Ok => { log::trace!("gRPC {method_name} Ok {elapsed_sec:.6}"); } Code::Cancelled => { // cluster mode generates a large amount of `stream error received: stream no longer needed` log::trace!("gRPC cancelled {method_name} {elapsed_sec:.6}"); } Code::DeadlineExceeded | Code::Aborted | Code::OutOfRange | Code::ResourceExhausted | Code::NotFound | Code::InvalidArgument | Code::AlreadyExists | Code::FailedPrecondition | Code::PermissionDenied | Code::Unauthenticated => { log::info!( "gRPC {} failed with {} {:?} {:.6}", method_name, grpc_status.code(), grpc_status.message(), elapsed_sec, ); } Code::Internal | Code::Unimplemented | Code::Unavailable | Code::DataLoss | Code::Unknown => log::error!( "gRPC {} unexpectedly failed with {} {:?} {:.6}", method_name, grpc_status.code(), grpc_status.message(), elapsed_sec, ), }; } else { // Fallback to response's `status_code` if no `grpc-status` header found. match response_tonic.status().as_u16() { 100..=199 => { log::trace!( "gRPC information {} {} {:.6}", method_name, response_tonic.status(), elapsed_sec, ); } 200..=299 => { log::trace!( "gRPC success {} {} {:.6}", method_name, response_tonic.status(), elapsed_sec, ); } 300..=399 => { log::debug!( "gRPC redirection {} {} {:.6}", method_name, response_tonic.status(), elapsed_sec, ); } 400..=499 => { log::info!( "gRPC client error {} {} {:.6}", method_name, response_tonic.status(), elapsed_sec, ); } 500..=599 => { log::error!( "gRPC server error {} {} {:.6}", method_name, response_tonic.status(), elapsed_sec, ); } _ => { log::warn!( "gRPC {} unknown status code {} {:.6}", method_name, response_tonic.status(), elapsed_sec, ); } }; } Ok(response_tonic) } } }) } } impl<S> Layer<S> for LoggingMiddlewareLayer { type Service = LoggingMiddleware<S>; fn layer(&self, service: S) -> Self::Service { LoggingMiddleware { inner: service } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/query_common.rs
src/tonic/api/query_common.rs
use std::time::{Duration, Instant}; use api::conversions::json::json_path_from_proto; use api::grpc::qdrant::{ BatchResult, CoreSearchPoints, CountPoints, CountResponse, DiscoverBatchResponse, DiscoverPoints, DiscoverResponse, FacetCounts, FacetResponse, GetPoints, GetResponse, GroupsResult, QueryBatchResponse, QueryGroupsResponse, QueryPointGroups, QueryPoints, QueryResponse, ReadConsistency as ReadConsistencyGrpc, RecommendBatchResponse, RecommendGroupsResponse, RecommendPointGroups, RecommendPoints, RecommendResponse, ScrollPoints, ScrollResponse, SearchBatchResponse, SearchGroupsResponse, SearchMatrixPoints, SearchPointGroups, SearchPoints, SearchResponse, }; use api::grpc::{InferenceUsage, Usage}; use api::rest::OrderByInterface; use collection::collection::distance_matrix::{ CollectionSearchMatrixRequest, CollectionSearchMatrixResponse, }; use collection::operations::consistency_params::ReadConsistency; use collection::operations::conversions::try_discover_request_from_grpc; use collection::operations::shard_selector_internal::ShardSelectorInternal; use collection::operations::types::{ CoreSearchRequest, PointRequestInternal, ScrollRequestInternal, default_exact_count, }; use collection::shards::shard::ShardId; use common::counter::hardware_accumulator::HwMeasurementAcc; use segment::data_types::facets::FacetParams; use segment::data_types::order_by::OrderBy; use segment::data_types::vectors::{DEFAULT_VECTOR_NAME, NamedQuery, VectorInternal}; use shard::query::query_enum::QueryEnum; use shard::search::CoreSearchRequestBatch; use storage::content_manager::toc::TableOfContent; use storage::content_manager::toc::request_hw_counter::RequestHwCounter; use storage::rbac::Access; use tonic::{Response, Status}; use crate::common::inference::params::InferenceParams; use crate::common::inference::query_requests_grpc::{ convert_query_point_groups_from_grpc, convert_query_points_from_grpc, }; use crate::common::query::*; use crate::common::strict_mode::*; pub(crate) fn convert_shard_selector_for_read( shard_id_selector: Option<ShardId>, shard_key_selector: Option<api::grpc::qdrant::ShardKeySelector>, ) -> Result<ShardSelectorInternal, Status> { let res = match (shard_id_selector, shard_key_selector) { (Some(shard_id), None) => ShardSelectorInternal::ShardId(shard_id), (None, Some(shard_key_selector)) => ShardSelectorInternal::try_from(shard_key_selector)?, (None, None) => ShardSelectorInternal::All, (Some(shard_id), Some(_)) => { debug_assert!( false, "Shard selection and shard key selector are mutually exclusive" ); ShardSelectorInternal::ShardId(shard_id) } }; Ok(res) } pub async fn search( toc_provider: impl CheckedTocProvider, search_points: SearchPoints, shard_selection: Option<ShardId>, access: Access, hw_measurement_acc: RequestHwCounter, ) -> Result<Response<SearchResponse>, Status> { let SearchPoints { collection_name, vector, filter, limit, offset, with_payload, params, score_threshold, vector_name, with_vectors, read_consistency, timeout, shard_key_selector, sparse_indices, } = search_points; let vector_internal = VectorInternal::from_vector_and_indices(vector, sparse_indices.map(|v| v.data)); let vector_struct = api::grpc::conversions::into_named_vector_struct(vector_name, vector_internal)?; let shard_selector = convert_shard_selector_for_read(shard_selection, shard_key_selector)?; let search_request = CoreSearchRequest { query: QueryEnum::Nearest(NamedQuery::from(vector_struct)), filter: filter.map(|f| f.try_into()).transpose()?, params: params.map(|p| p.into()), limit: limit as usize, offset: offset.unwrap_or_default() as usize, with_payload: with_payload.map(|wp| wp.try_into()).transpose()?, with_vector: Some( with_vectors .map(|selector| selector.into()) .unwrap_or_default(), ), score_threshold, }; let toc = toc_provider .check_strict_mode( &search_request, &collection_name, timeout.map(|i| i as usize), &access, ) .await?; let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let timing = Instant::now(); let scored_points = do_core_search_points( toc, &collection_name, search_request, read_consistency, shard_selector, access, timeout.map(Duration::from_secs), hw_measurement_acc.get_counter(), ) .await?; let response = SearchResponse { result: scored_points .into_iter() .map(|point| point.into()) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(hw_measurement_acc.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn core_search_batch( toc_provider: impl CheckedTocProvider, collection_name: &str, requests: Vec<(CoreSearchRequest, ShardSelectorInternal)>, read_consistency: Option<ReadConsistencyGrpc>, access: Access, timeout: Option<Duration>, request_hw_counter: RequestHwCounter, ) -> Result<Response<SearchBatchResponse>, Status> { let toc = toc_provider .check_strict_mode_batch( &requests, |i| &i.0, collection_name, timeout.map(|i| i.as_secs() as usize), &access, ) .await?; let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let timing = Instant::now(); let scored_points = do_search_batch_points( toc, collection_name, requests, read_consistency, access, timeout, request_hw_counter.get_counter(), ) .await?; let response = SearchBatchResponse { result: scored_points .into_iter() .map(|points| BatchResult { result: points.into_iter().map(|p| p.into()).collect(), }) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } #[allow(clippy::too_many_arguments)] pub async fn core_search_list( toc: &TableOfContent, collection_name: String, search_points: Vec<CoreSearchPoints>, read_consistency: Option<ReadConsistencyGrpc>, shard_selection: Option<ShardId>, access: Access, timeout: Option<Duration>, request_hw_counter: RequestHwCounter, ) -> Result<Response<SearchBatchResponse>, Status> { let searches: Result<Vec<_>, Status> = search_points .into_iter() .map(CoreSearchRequest::try_from) .collect(); let request = CoreSearchRequestBatch { searches: searches?, }; let timing = Instant::now(); // As this function is handling an internal request, // we can assume that shard_key is already resolved let shard_selection = match shard_selection { None => { debug_assert!(false, "Shard selection is expected for internal request"); ShardSelectorInternal::All } Some(shard_id) => ShardSelectorInternal::ShardId(shard_id), }; let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let scored_points = toc .core_search_batch( &collection_name, request, read_consistency, shard_selection, access, timeout, request_hw_counter.get_counter(), ) .await?; let response = SearchBatchResponse { result: scored_points .into_iter() .map(|points| BatchResult { result: points.into_iter().map(|p| p.into()).collect(), }) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn search_groups( toc_provider: impl CheckedTocProvider, search_point_groups: SearchPointGroups, shard_selection: Option<ShardId>, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<SearchGroupsResponse>, Status> { let search_groups_request = search_point_groups.clone().try_into()?; let SearchPointGroups { collection_name, read_consistency, timeout, shard_key_selector, .. } = search_point_groups; let toc = toc_provider .check_strict_mode( &search_groups_request, &collection_name, timeout.map(|i| i as usize), &access, ) .await?; let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let shard_selector = convert_shard_selector_for_read(shard_selection, shard_key_selector)?; let timing = Instant::now(); let groups_result = crate::common::query::do_search_point_groups( toc, &collection_name, search_groups_request, read_consistency, shard_selector, access, timeout.map(Duration::from_secs), request_hw_counter.get_counter(), ) .await?; let groups_result = GroupsResult::try_from(groups_result) .map_err(|e| Status::internal(format!("Failed to convert groups result: {e}")))?; let response = SearchGroupsResponse { result: Some(groups_result), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn recommend( toc_provider: impl CheckedTocProvider, recommend_points: RecommendPoints, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<RecommendResponse>, Status> { // extract a few fields from the request and convert to internal request let collection_name = recommend_points.collection_name.clone(); let read_consistency = recommend_points.read_consistency.clone(); let shard_key_selector = recommend_points.shard_key_selector.clone(); let timeout = recommend_points.timeout; let request = collection::operations::types::RecommendRequestInternal::try_from(recommend_points)?; let toc = toc_provider .check_strict_mode( &request, &collection_name, timeout.map(|i| i as usize), &access, ) .await?; let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let shard_selector = convert_shard_selector_for_read(None, shard_key_selector)?; let timeout = timeout.map(Duration::from_secs); let timing = Instant::now(); let recommended_points = toc .recommend( &collection_name, request, read_consistency, shard_selector, access, timeout, request_hw_counter.get_counter(), ) .await?; let response = RecommendResponse { result: recommended_points .into_iter() .map(|point| point.into()) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn recommend_batch( toc_provider: impl CheckedTocProvider, collection_name: &str, recommend_points: Vec<RecommendPoints>, read_consistency: Option<ReadConsistencyGrpc>, access: Access, timeout: Option<Duration>, request_hw_counter: RequestHwCounter, ) -> Result<Response<RecommendBatchResponse>, Status> { let mut requests = Vec::with_capacity(recommend_points.len()); for mut request in recommend_points { let shard_selector = convert_shard_selector_for_read(None, request.shard_key_selector.take())?; let internal_request: collection::operations::types::RecommendRequestInternal = request.try_into()?; requests.push((internal_request, shard_selector)); } let toc = toc_provider .check_strict_mode_batch( &requests, |i| &i.0, collection_name, timeout.map(|i| i.as_secs() as usize), &access, ) .await?; let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let timing = Instant::now(); let scored_points = toc .recommend_batch( collection_name, requests, read_consistency, access, timeout, request_hw_counter.get_counter(), ) .await?; let response = RecommendBatchResponse { result: scored_points .into_iter() .map(|points| BatchResult { result: points.into_iter().map(|p| p.into()).collect(), }) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn recommend_groups( toc_provider: impl CheckedTocProvider, recommend_point_groups: RecommendPointGroups, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<RecommendGroupsResponse>, Status> { let recommend_groups_request = recommend_point_groups.clone().try_into()?; let RecommendPointGroups { collection_name, read_consistency, timeout, shard_key_selector, .. } = recommend_point_groups; let toc = toc_provider .check_strict_mode( &recommend_groups_request, &collection_name, timeout.map(|i| i as usize), &access, ) .await?; let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let shard_selector = convert_shard_selector_for_read(None, shard_key_selector)?; let timing = Instant::now(); let groups_result = crate::common::query::do_recommend_point_groups( toc, &collection_name, recommend_groups_request, read_consistency, shard_selector, access, timeout.map(Duration::from_secs), request_hw_counter.get_counter(), ) .await?; let groups_result = GroupsResult::try_from(groups_result) .map_err(|e| Status::internal(format!("Failed to convert groups result: {e}")))?; let response = RecommendGroupsResponse { result: Some(groups_result), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn discover( toc_provider: impl CheckedTocProvider, discover_points: DiscoverPoints, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<DiscoverResponse>, Status> { let (request, collection_name, read_consistency, timeout, shard_key_selector) = try_discover_request_from_grpc(discover_points)?; let toc = toc_provider .check_strict_mode( &request, &collection_name, timeout.map(|i| i.as_secs() as usize), &access, ) .await?; let timing = Instant::now(); let shard_selector = convert_shard_selector_for_read(None, shard_key_selector)?; let discovered_points = toc .discover( &collection_name, request, read_consistency, shard_selector, access, timeout, request_hw_counter.get_counter(), ) .await?; let response = DiscoverResponse { result: discovered_points .into_iter() .map(|point| point.into()) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn discover_batch( toc_provider: impl CheckedTocProvider, collection_name: &str, discover_points: Vec<DiscoverPoints>, read_consistency: Option<ReadConsistencyGrpc>, access: Access, timeout: Option<Duration>, request_hw_counter: RequestHwCounter, ) -> Result<Response<DiscoverBatchResponse>, Status> { let mut requests = Vec::with_capacity(discover_points.len()); for discovery_request in discover_points { let (internal_request, _collection_name, _consistency, _timeout, shard_key_selector) = try_discover_request_from_grpc(discovery_request)?; let shard_selector = convert_shard_selector_for_read(None, shard_key_selector)?; requests.push((internal_request, shard_selector)); } let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let toc = toc_provider .check_strict_mode_batch( &requests, |i| &i.0, collection_name, timeout.map(|i| i.as_secs() as usize), &access, ) .await?; let timing = Instant::now(); let scored_points = toc .discover_batch( collection_name, requests, read_consistency, access, timeout, request_hw_counter.get_counter(), ) .await?; let response = DiscoverBatchResponse { result: scored_points .into_iter() .map(|points| BatchResult { result: points.into_iter().map(|p| p.into()).collect(), }) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn scroll( toc_provider: impl CheckedTocProvider, scroll_points: ScrollPoints, shard_selection: Option<ShardId>, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<ScrollResponse>, Status> { let ScrollPoints { collection_name, filter, offset, limit, with_payload, with_vectors, read_consistency, shard_key_selector, order_by, timeout, } = scroll_points; let scroll_request = ScrollRequestInternal { offset: offset.map(|o| o.try_into()).transpose()?, limit: limit.map(|l| l as usize), filter: filter.map(|f| f.try_into()).transpose()?, with_payload: with_payload.map(|wp| wp.try_into()).transpose()?, with_vector: with_vectors .map(|selector| selector.into()) .unwrap_or_default(), order_by: order_by .map(OrderBy::try_from) .transpose()? .map(OrderByInterface::Struct), }; let toc = toc_provider .check_strict_mode( &scroll_request, &collection_name, timeout.map(|i| i as usize), &access, ) .await?; let timeout = timeout.map(Duration::from_secs); let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let shard_selector = convert_shard_selector_for_read(shard_selection, shard_key_selector)?; let timing = Instant::now(); let scrolled_points = do_scroll_points( toc, &collection_name, scroll_request, read_consistency, timeout, shard_selector, access, request_hw_counter.get_counter(), ) .await?; let points: Result<_, _> = scrolled_points .points .into_iter() .map(api::grpc::qdrant::RetrievedPoint::try_from) .collect(); let points = points.map_err(|e| Status::internal(format!("Failed to convert points: {e}")))?; let response = ScrollResponse { next_page_offset: scrolled_points.next_page_offset.map(|n| n.into()), result: points, time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn count( toc_provider: impl CheckedTocProvider, count_points: CountPoints, shard_selection: Option<ShardId>, access: &Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<CountResponse>, Status> { let CountPoints { collection_name, filter, exact, read_consistency, shard_key_selector, timeout, } = count_points; let count_request = collection::operations::types::CountRequestInternal { filter: filter.map(|f| f.try_into()).transpose()?, exact: exact.unwrap_or_else(default_exact_count), }; let toc = toc_provider .check_strict_mode( &count_request, &collection_name, timeout.map(|i| i as usize), access, ) .await?; let timeout = timeout.map(Duration::from_secs); let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let shard_selector = convert_shard_selector_for_read(shard_selection, shard_key_selector)?; let timing = Instant::now(); let count_result = do_count_points( toc, &collection_name, count_request, read_consistency, timeout, shard_selector, access.clone(), request_hw_counter.get_counter(), ) .await?; let response = CountResponse { result: Some(count_result.into()), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn get( toc_provider: impl CheckedTocProvider, get_points: GetPoints, shard_selection: Option<ShardId>, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<GetResponse>, Status> { let GetPoints { collection_name, ids, with_payload, with_vectors, read_consistency, shard_key_selector, timeout, } = get_points; let point_request = PointRequestInternal { ids: ids .into_iter() .map(|p| p.try_into()) .collect::<Result<_, _>>()?, with_payload: with_payload.map(|wp| wp.try_into()).transpose()?, with_vector: with_vectors .map(|selector| selector.into()) .unwrap_or_default(), }; let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let shard_selector = convert_shard_selector_for_read(shard_selection, shard_key_selector)?; let timing = Instant::now(); let toc = toc_provider .check_strict_mode( &point_request, &collection_name, timeout.map(|i| i as usize), &access, ) .await?; let timeout = timeout.map(Duration::from_secs); let records = do_get_points( toc, &collection_name, point_request, read_consistency, timeout, shard_selector, access, request_hw_counter.get_counter(), ) .await?; let response = GetResponse { result: records.into_iter().map(|point| point.into()).collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn query( toc_provider: impl CheckedTocProvider, query_points: QueryPoints, shard_selection: Option<ShardId>, access: Access, request_hw_counter: RequestHwCounter, inference_params: InferenceParams, ) -> Result<Response<QueryResponse>, Status> { let shard_key_selector = query_points.shard_key_selector.clone(); let shard_selector = convert_shard_selector_for_read(shard_selection, shard_key_selector)?; let read_consistency = query_points .read_consistency .clone() .map(TryFrom::try_from) .transpose()?; let collection_name = query_points.collection_name.clone(); let timeout = query_points.timeout; let (request, inference_usage) = convert_query_points_from_grpc(query_points, inference_params).await?; let toc = toc_provider .check_strict_mode( &request, &collection_name, timeout.map(|i| i as usize), &access, ) .await?; let timeout = timeout.map(Duration::from_secs); let timing = Instant::now(); let scored_points = do_query_points( toc, &collection_name, request, read_consistency, shard_selector, access, timeout, request_hw_counter.get_counter(), ) .await?; let response = QueryResponse { result: scored_points .into_iter() .map(|point| point.into()) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::new(request_hw_counter.to_grpc_api(), Some(inference_usage)).into_non_empty(), }; Ok(Response::new(response)) } #[allow(clippy::too_many_arguments)] pub async fn query_batch( toc_provider: impl CheckedTocProvider, collection_name: &str, points: Vec<QueryPoints>, read_consistency: Option<ReadConsistencyGrpc>, access: Access, timeout: Option<Duration>, request_hw_counter: RequestHwCounter, inference_params: InferenceParams, ) -> Result<Response<QueryBatchResponse>, Status> { let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let mut requests = Vec::with_capacity(points.len()); let mut total_inference_usage = InferenceUsage::default(); for query_points in points { let shard_key_selector = query_points.shard_key_selector.clone(); let shard_selector = convert_shard_selector_for_read(None, shard_key_selector)?; let (request, usage) = convert_query_points_from_grpc(query_points, inference_params.clone()).await?; total_inference_usage.merge(usage); requests.push((request, shard_selector)); } let toc = toc_provider .check_strict_mode_batch( &requests, |i| &i.0, collection_name, timeout.map(|i| i.as_secs() as usize), &access, ) .await?; let timing = Instant::now(); let scored_points = do_query_batch_points( toc, collection_name, requests, read_consistency, access, timeout, request_hw_counter.get_counter(), ) .await?; let response = QueryBatchResponse { result: scored_points .into_iter() .map(|points| BatchResult { result: points.into_iter().map(|p| p.into()).collect(), }) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::new( request_hw_counter.to_grpc_api(), total_inference_usage.into_non_empty(), ) .into_non_empty(), }; Ok(Response::new(response)) } pub async fn query_groups( toc_provider: impl CheckedTocProvider, query_points: QueryPointGroups, shard_selection: Option<ShardId>, access: Access, request_hw_counter: RequestHwCounter, inference_params: InferenceParams, ) -> Result<Response<QueryGroupsResponse>, Status> { let shard_key_selector = query_points.shard_key_selector.clone(); let shard_selector = convert_shard_selector_for_read(shard_selection, shard_key_selector)?; let read_consistency = query_points .read_consistency .clone() .map(TryFrom::try_from) .transpose()?; let timeout = query_points.timeout; let collection_name = query_points.collection_name.clone(); let (request, inference_usage) = convert_query_point_groups_from_grpc(query_points, inference_params).await?; let toc = toc_provider .check_strict_mode( &request, &collection_name, timeout.map(|i| i as usize), &access, ) .await?; let timeout = timeout.map(Duration::from_secs); let timing = Instant::now(); let groups_result = do_query_point_groups( toc, &collection_name, request, read_consistency, shard_selector, access, timeout, request_hw_counter.get_counter(), ) .await?; let grpc_group_result = GroupsResult::try_from(groups_result) .map_err(|err| Status::internal(format!("failed to convert result: {err}")))?; let response = QueryGroupsResponse { result: Some(grpc_group_result), time: timing.elapsed().as_secs_f64(), usage: Usage::new(request_hw_counter.to_grpc_api(), Some(inference_usage)).into_non_empty(), }; Ok(Response::new(response)) } pub async fn facet( toc_provider: impl CheckedTocProvider, facet_counts: FacetCounts, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<FacetResponse>, Status> { let FacetCounts { collection_name, key, filter, exact, limit, read_consistency, shard_key_selector, timeout, } = facet_counts; let facet_request = FacetParams { key: json_path_from_proto(&key)?, filter: filter.map(TryInto::try_into).transpose()?, limit: limit .map(usize::try_from) .transpose() .map_err(|_| Status::invalid_argument("could not parse limit param into usize"))? .unwrap_or(FacetParams::DEFAULT_LIMIT), exact: exact.unwrap_or(FacetParams::DEFAULT_EXACT), }; let toc = toc_provider .check_strict_mode( &facet_request, &collection_name, timeout.map(|i| i as usize), &access, ) .await?; let timeout = timeout.map(Duration::from_secs); let read_consistency = ReadConsistency::try_from_optional(read_consistency)?; let shard_selector = convert_shard_selector_for_read(None, shard_key_selector)?; let timing = Instant::now(); let facet_response = toc .facet( &collection_name, facet_request, shard_selector, read_consistency, access, timeout, request_hw_counter.get_counter(), ) .await?; let segment::data_types::facets::FacetResponse { hits } = facet_response; let response = FacetResponse { hits: hits.into_iter().map(From::from).collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(request_hw_counter.to_grpc_api()).into_non_empty(), }; Ok(Response::new(response)) } pub async fn search_points_matrix( toc_provider: impl CheckedTocProvider, search_matrix_points: SearchMatrixPoints, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<CollectionSearchMatrixResponse, Status> { let SearchMatrixPoints { collection_name, filter, sample, limit, using, read_consistency, shard_key_selector, timeout, } = search_matrix_points; let search_matrix_request = CollectionSearchMatrixRequest { filter: filter.map(TryInto::try_into).transpose()?, sample_size: sample .map(usize::try_from) .transpose() .map_err(|_| Status::invalid_argument("could not parse 'sample' param into usize"))? .unwrap_or(CollectionSearchMatrixRequest::DEFAULT_SAMPLE), limit_per_sample: limit .map(usize::try_from) .transpose() .map_err(|_| Status::invalid_argument("could not parse 'limit' param into usize"))? .unwrap_or(CollectionSearchMatrixRequest::DEFAULT_LIMIT_PER_SAMPLE), using: using.unwrap_or_else(|| DEFAULT_VECTOR_NAME.to_owned()), }; let toc = toc_provider .check_strict_mode( &search_matrix_request, &collection_name,
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
true
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/collections_internal_api.rs
src/tonic/api/collections_internal_api.rs
use std::sync::Arc; use std::time::{Duration, Instant}; use api::grpc::qdrant::collections_internal_server::CollectionsInternal; use api::grpc::qdrant::{ CollectionOperationResponse, GetCollectionInfoRequestInternal, GetCollectionInfoResponse, GetShardRecoveryPointRequest, GetShardRecoveryPointResponse, InitiateShardTransferRequest, UpdateShardCutoffPointRequest, WaitForShardStateRequest, }; use storage::content_manager::toc::TableOfContent; use storage::rbac::{Access, AccessRequirements, CollectionPass}; use tonic::{Request, Response, Status}; use super::validate_and_log; use crate::tonic::api::collections_common::get; const FULL_ACCESS: Access = Access::full("Internal API"); fn full_access_pass(collection_name: &str) -> Result<CollectionPass<'_>, Status> { FULL_ACCESS .check_collection_access(collection_name, AccessRequirements::new()) .map_err(Status::from) } pub struct CollectionsInternalService { toc: Arc<TableOfContent>, } impl CollectionsInternalService { pub fn new(toc: Arc<TableOfContent>) -> Self { Self { toc } } } #[tonic::async_trait] impl CollectionsInternal for CollectionsInternalService { async fn get( &self, request: Request<GetCollectionInfoRequestInternal>, ) -> Result<Response<GetCollectionInfoResponse>, Status> { validate_and_log(request.get_ref()); let GetCollectionInfoRequestInternal { get_collection_info_request, shard_id, } = request.into_inner(); let get_collection_info_request = get_collection_info_request .ok_or_else(|| Status::invalid_argument("GetCollectionInfoRequest is missing"))?; get( self.toc.as_ref(), get_collection_info_request, FULL_ACCESS.clone(), Some(shard_id), ) .await } async fn initiate( &self, request: Request<InitiateShardTransferRequest>, ) -> Result<Response<CollectionOperationResponse>, Status> { // TODO: Ensure cancel safety! validate_and_log(request.get_ref()); let timing = Instant::now(); let InitiateShardTransferRequest { collection_name, shard_id, } = request.into_inner(); // TODO: Ensure cancel safety! self.toc .initiate_receiving_shard(collection_name, shard_id) .await?; let response = CollectionOperationResponse { result: true, time: timing.elapsed().as_secs_f64(), }; Ok(Response::new(response)) } async fn wait_for_shard_state( &self, request: Request<WaitForShardStateRequest>, ) -> Result<Response<CollectionOperationResponse>, Status> { let request = request.into_inner(); validate_and_log(&request); let timing = Instant::now(); let WaitForShardStateRequest { collection_name, shard_id, state, timeout, } = request; let state = state.try_into()?; let timeout = Duration::from_secs(timeout); let collection_read = self .toc .get_collection(&full_access_pass(&collection_name)?) .await .map_err(|err| { Status::not_found(format!( "Collection {collection_name} could not be found: {err}" )) })?; // Wait for replica state collection_read .wait_local_shard_replica_state(shard_id, state, timeout) .await .map_err(|err| { Status::aborted(format!( "Failed to wait for shard {shard_id} to get into {state:?} state: {err}" )) })?; let response = CollectionOperationResponse { result: true, time: timing.elapsed().as_secs_f64(), }; Ok(Response::new(response)) } async fn get_shard_recovery_point( &self, request: Request<GetShardRecoveryPointRequest>, ) -> Result<Response<GetShardRecoveryPointResponse>, Status> { validate_and_log(request.get_ref()); let timing = Instant::now(); let GetShardRecoveryPointRequest { collection_name, shard_id, } = request.into_inner(); let collection_read = self .toc .get_collection(&full_access_pass(&collection_name)?) .await .map_err(|err| { Status::not_found(format!( "Collection {collection_name} could not be found: {err}" )) })?; // Get shard recovery point let recovery_point = collection_read .shard_recovery_point(shard_id) .await .map_err(|err| { Status::internal(format!( "Failed to get recovery point for shard {shard_id}: {err}" )) })?; let response = GetShardRecoveryPointResponse { recovery_point: Some(recovery_point.into()), time: timing.elapsed().as_secs_f64(), }; Ok(Response::new(response)) } async fn update_shard_cutoff_point( &self, request: Request<UpdateShardCutoffPointRequest>, ) -> Result<Response<CollectionOperationResponse>, Status> { validate_and_log(request.get_ref()); let timing = Instant::now(); let UpdateShardCutoffPointRequest { collection_name, shard_id, cutoff, } = request.into_inner(); let cutoff = cutoff.ok_or_else(|| Status::invalid_argument("Missing cutoff point"))?; let collection_read = self .toc .get_collection(&full_access_pass(&collection_name)?) .await .map_err(|err| { Status::not_found(format!( "Collection {collection_name} could not be found: {err}" )) })?; // Set the shard cutoff point collection_read .update_shard_cutoff_point(shard_id, &cutoff.try_into()?) .await .map_err(|err| { Status::internal(format!( "Failed to set shard cutoff point for shard {shard_id}: {err}" )) })?; let response = CollectionOperationResponse { result: true, time: timing.elapsed().as_secs_f64(), }; Ok(Response::new(response)) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/mod.rs
src/tonic/api/mod.rs
pub mod collections_api; pub mod collections_internal_api; pub mod points_api; pub mod points_internal_api; pub mod raft_api; pub mod snapshots_api; mod collections_common; mod query_common; mod update_common; use collection::operations::validation; use tonic::Status; use validator::Validate; /// Validate the given request and fail on error. /// /// Returns validation error on failure. fn validate(request: &impl Validate) -> Result<(), Status> { request.validate().map_err(|ref err| { Status::invalid_argument(validation::label_errors("Validation error in body", err)) }) } /// Validate the given request. Returns validation error on failure. fn validate_and_log(request: &impl Validate) { if let Err(ref err) = request.validate() { validation::warn_validation_errors("Internal gRPC", err); } } #[cfg(test)] mod tests { use validator::Validate; use super::*; #[derive(Validate, Debug)] struct SomeThing { #[validate(range(min = 1))] pub idx: usize, } #[derive(Validate, Debug)] struct OtherThing { #[validate(nested)] pub things: Vec<SomeThing>, } #[test] fn test_validation() { use tonic::Code; let bad_config = OtherThing { things: vec![SomeThing { idx: 0 }], }; let validation = validate(&bad_config).expect_err("validation of bad request payload should fail"); assert_eq!(validation.code(), Code::InvalidArgument); assert_eq!( validation.message(), "Validation error in body: [things[0].idx: value 0 invalid, must be 1 or larger]" ) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/collections_api.rs
src/tonic/api/collections_api.rs
use std::sync::Arc; use std::time::{Duration, Instant}; use api::grpc::qdrant::collections_server::Collections; use api::grpc::qdrant::{ ChangeAliases, CollectionClusterInfoRequest, CollectionClusterInfoResponse, CollectionExistsRequest, CollectionExistsResponse, CollectionOperationResponse, CreateCollection, CreateShardKeyRequest, CreateShardKeyResponse, DeleteCollection, DeleteShardKeyRequest, DeleteShardKeyResponse, GetCollectionInfoRequest, GetCollectionInfoResponse, ListAliasesRequest, ListAliasesResponse, ListCollectionAliasesRequest, ListCollectionsRequest, ListCollectionsResponse, ListShardKeysRequest, ListShardKeysResponse, UpdateCollection, UpdateCollectionClusterSetupRequest, UpdateCollectionClusterSetupResponse, }; use collection::operations::cluster_ops::{ ClusterOperations, CreateShardingKeyOperation, DropShardingKeyOperation, }; use collection::operations::types::CollectionsAliasesResponse; use collection::operations::verification::new_unchecked_verification_pass; use storage::dispatcher::Dispatcher; use tonic::{Request, Response, Status}; use super::validate; use crate::common::collections::*; use crate::tonic::api::collections_common::get; use crate::tonic::auth::extract_access; pub struct CollectionsService { dispatcher: Arc<Dispatcher>, } impl CollectionsService { pub fn new(dispatcher: Arc<Dispatcher>) -> Self { Self { dispatcher } } async fn perform_operation<O>( &self, mut request: Request<O>, ) -> Result<Response<CollectionOperationResponse>, Status> where O: WithTimeout + TryInto< storage::content_manager::collection_meta_ops::CollectionMetaOperations, Error = Status, >, { let timing = Instant::now(); let access = extract_access(&mut request); let operation = request.into_inner(); let wait_timeout = operation.wait_timeout(); let result = self .dispatcher .submit_collection_meta_op(operation.try_into()?, access, wait_timeout) .await?; let response = CollectionOperationResponse::from((timing, result)); Ok(Response::new(response)) } } #[tonic::async_trait] impl Collections for CollectionsService { async fn get( &self, mut request: Request<GetCollectionInfoRequest>, ) -> Result<Response<GetCollectionInfoResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); // Nothing to verify here. let pass = new_unchecked_verification_pass(); get( self.dispatcher.toc(&access, &pass), request.into_inner(), access, None, ) .await } async fn list( &self, mut request: Request<ListCollectionsRequest>, ) -> Result<Response<ListCollectionsResponse>, Status> { validate(request.get_ref())?; let timing = Instant::now(); let access = extract_access(&mut request); // Nothing to verify here. let pass = new_unchecked_verification_pass(); let result = do_list_collections(self.dispatcher.toc(&access, &pass), access).await?; let response = ListCollectionsResponse::from((timing, result)); Ok(Response::new(response)) } async fn create( &self, request: Request<CreateCollection>, ) -> Result<Response<CollectionOperationResponse>, Status> { validate(request.get_ref())?; self.perform_operation(request).await } async fn update( &self, request: Request<UpdateCollection>, ) -> Result<Response<CollectionOperationResponse>, Status> { validate(request.get_ref())?; self.perform_operation(request).await } async fn delete( &self, request: Request<DeleteCollection>, ) -> Result<Response<CollectionOperationResponse>, Status> { validate(request.get_ref())?; self.perform_operation(request).await } async fn update_aliases( &self, request: Request<ChangeAliases>, ) -> Result<Response<CollectionOperationResponse>, Status> { validate(request.get_ref())?; self.perform_operation(request).await } async fn list_collection_aliases( &self, mut request: Request<ListCollectionAliasesRequest>, ) -> Result<Response<ListAliasesResponse>, Status> { validate(request.get_ref())?; let timing = Instant::now(); let access = extract_access(&mut request); // Nothing to verify here. let pass = new_unchecked_verification_pass(); let ListCollectionAliasesRequest { collection_name } = request.into_inner(); let CollectionsAliasesResponse { aliases } = do_list_collection_aliases( self.dispatcher.toc(&access, &pass), access, &collection_name, ) .await?; let response = ListAliasesResponse { aliases: aliases.into_iter().map(|alias| alias.into()).collect(), time: timing.elapsed().as_secs_f64(), }; Ok(Response::new(response)) } async fn list_aliases( &self, mut request: Request<ListAliasesRequest>, ) -> Result<Response<ListAliasesResponse>, Status> { validate(request.get_ref())?; let timing = Instant::now(); let access = extract_access(&mut request); // Nothing to verify here. let pass = new_unchecked_verification_pass(); let CollectionsAliasesResponse { aliases } = do_list_aliases(self.dispatcher.toc(&access, &pass), access).await?; let response = ListAliasesResponse { aliases: aliases.into_iter().map(|alias| alias.into()).collect(), time: timing.elapsed().as_secs_f64(), }; Ok(Response::new(response)) } async fn collection_exists( &self, mut request: Request<CollectionExistsRequest>, ) -> Result<Response<CollectionExistsResponse>, Status> { let timing = Instant::now(); validate(request.get_ref())?; let access = extract_access(&mut request); // Nothing to verify here. let pass = new_unchecked_verification_pass(); let CollectionExistsRequest { collection_name } = request.into_inner(); let result = do_collection_exists( self.dispatcher.toc(&access, &pass), access, &collection_name, ) .await?; let response = CollectionExistsResponse { result: Some(result), time: timing.elapsed().as_secs_f64(), }; Ok(Response::new(response)) } async fn collection_cluster_info( &self, mut request: Request<CollectionClusterInfoRequest>, ) -> Result<Response<CollectionClusterInfoResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); // Nothing to verify here. let pass = new_unchecked_verification_pass(); let response = do_get_collection_cluster( self.dispatcher.toc(&access, &pass), access, request.into_inner().collection_name.as_str(), ) .await? .into(); Ok(Response::new(response)) } async fn update_collection_cluster_setup( &self, mut request: Request<UpdateCollectionClusterSetupRequest>, ) -> Result<Response<UpdateCollectionClusterSetupResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let UpdateCollectionClusterSetupRequest { collection_name, operation, timeout, .. } = request.into_inner(); let result = do_update_collection_cluster( self.dispatcher.as_ref(), collection_name, operation .ok_or_else(|| Status::new(tonic::Code::InvalidArgument, "empty operation"))? .try_into()?, access, timeout.map(std::time::Duration::from_secs), ) .await?; Ok(Response::new(UpdateCollectionClusterSetupResponse { result, })) } async fn list_shard_keys( &self, mut request: Request<ListShardKeysRequest>, ) -> Result<Response<ListShardKeysResponse>, Status> { validate(request.get_ref())?; let timing = Instant::now(); let access = extract_access(&mut request); // Nothing to verify here. let pass = new_unchecked_verification_pass(); let result = do_get_collection_shard_keys( self.dispatcher.toc(&access, &pass), access, request.into_inner().collection_name.as_str(), ) .await?; let response = ListShardKeysResponse::from((timing, result)); Ok(Response::new(response)) } async fn create_shard_key( &self, mut request: Request<CreateShardKeyRequest>, ) -> Result<Response<CreateShardKeyResponse>, Status> { let access = extract_access(&mut request); let CreateShardKeyRequest { collection_name, request, timeout, } = request.into_inner(); let Some(request) = request else { return Err(Status::new(tonic::Code::InvalidArgument, "empty request")); }; let timeout = timeout.map(std::time::Duration::from_secs); let operation = ClusterOperations::CreateShardingKey(CreateShardingKeyOperation { create_sharding_key: request.try_into()?, }); let result = do_update_collection_cluster( self.dispatcher.as_ref(), collection_name, operation, access, timeout, ) .await?; Ok(Response::new(CreateShardKeyResponse { result })) } async fn delete_shard_key( &self, mut request: Request<DeleteShardKeyRequest>, ) -> Result<Response<DeleteShardKeyResponse>, Status> { let access = extract_access(&mut request); let DeleteShardKeyRequest { collection_name, request, timeout, } = request.into_inner(); let Some(request) = request else { return Err(Status::new(tonic::Code::InvalidArgument, "empty request")); }; let timeout = timeout.map(std::time::Duration::from_secs); let operation = ClusterOperations::DropShardingKey(DropShardingKeyOperation { drop_sharding_key: request.try_into()?, }); let result = do_update_collection_cluster( self.dispatcher.as_ref(), collection_name, operation, access, timeout, ) .await?; Ok(Response::new(DeleteShardKeyResponse { result })) } } trait WithTimeout { fn wait_timeout(&self) -> Option<Duration>; } macro_rules! impl_with_timeout { ($operation:ty) => { impl WithTimeout for $operation { fn wait_timeout(&self) -> Option<Duration> { self.timeout.map(Duration::from_secs) } } }; } impl_with_timeout!(CreateCollection); impl_with_timeout!(UpdateCollection); impl_with_timeout!(DeleteCollection); impl_with_timeout!(ChangeAliases); impl_with_timeout!(UpdateCollectionClusterSetupRequest);
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/raft_api.rs
src/tonic/api/raft_api.rs
use api::grpc::qdrant::raft_server::Raft; use api::grpc::qdrant::{ AddPeerToKnownMessage, AllPeers, Peer, PeerId, RaftMessage as RaftMessageBytes, Uri as UriStr, }; use itertools::Itertools; use raft::eraftpb::Message as RaftMessage; use storage::content_manager::consensus_manager::ConsensusStateRef; use storage::content_manager::consensus_ops::ConsensusOperations; use tokio::sync::mpsc::Sender; use tonic::transport::Uri; use tonic::{Request, Response, Status, async_trait}; use super::validate; use crate::consensus; pub struct RaftService { message_sender: Sender<consensus::Message>, consensus_state: ConsensusStateRef, use_tls: bool, } impl RaftService { pub fn new( sender: Sender<consensus::Message>, consensus_state: ConsensusStateRef, use_tls: bool, ) -> Self { Self { message_sender: sender, consensus_state, use_tls, } } } #[async_trait] impl Raft for RaftService { async fn send(&self, mut request: Request<RaftMessageBytes>) -> Result<Response<()>, Status> { let message = <RaftMessage as prost_for_raft::Message>::decode(&request.get_mut().message[..]) .map_err(|err| { Status::invalid_argument(format!("Failed to parse raft message: {err}")) })?; self.message_sender .send(consensus::Message::FromPeer(Box::new(message))) .await .map_err(|_| Status::internal("Can't send Raft message over channel"))?; Ok(Response::new(())) } async fn who_is( &self, request: tonic::Request<PeerId>, ) -> Result<tonic::Response<UriStr>, tonic::Status> { let addresses = self.consensus_state.peer_address_by_id(); let uri = addresses .get(&request.get_ref().id) .ok_or_else(|| Status::internal("Peer not found"))?; Ok(Response::new(UriStr { uri: uri.to_string(), })) } async fn add_peer_to_known( &self, request: tonic::Request<AddPeerToKnownMessage>, ) -> Result<tonic::Response<AllPeers>, tonic::Status> { validate(request.get_ref())?; let peer = request.get_ref(); let uri_string = if let Some(uri) = &peer.uri { uri.clone() } else { let ip = request .remote_addr() .ok_or_else(|| { Status::failed_precondition("Remote address unavailable due to the used IO") })? .ip(); let port = peer .port .ok_or_else(|| Status::invalid_argument("URI or port should be supplied"))?; if self.use_tls { format!("https://{ip}:{port}") } else { format!("http://{ip}:{port}") } }; let uri: Uri = uri_string .parse() .map_err(|err| Status::internal(format!("Failed to parse uri: {err}")))?; let peer = request.into_inner(); // the consensus operation can take up to DEFAULT_META_OP_WAIT self.consensus_state .propose_consensus_op_with_await( ConsensusOperations::AddPeer { peer_id: peer.id, uri: uri.to_string(), }, None, ) .await .map_err(|err| Status::internal(format!("Failed to add peer: {err}")))?; let mut addresses = self.consensus_state.peer_address_by_id(); // Make sure that the new peer is now present in the known addresses if !addresses.values().contains(&uri) { return Err(Status::internal(format!( "Failed to add peer after consensus: {uri}" ))); } let first_peer_id = self.consensus_state.first_voter(); // If `first_peer_id` is not present in the list of peers, it means it was removed from // cluster at some point. // // Before Qdrant version 1.11.6 origin peer was not committed to consensus, so if it was // removed from cluster, any node added to the cluster after this would not recognize it as // being part of the cluster in the past and will end up with a broken consensus state. // // To prevent this, we add `first_peer_id` (with a fake URI) to the list of peers. // // `add_peer_to_known` is used to add new peers to the cluster, and so `first_peer_id` (and // its fake URI) would be removed from new peer's state shortly, while it will be synchronizing // and applying past Raft log. addresses.entry(first_peer_id).or_default(); Ok(Response::new(AllPeers { all_peers: addresses .into_iter() .map(|(id, uri)| Peer { id, uri: uri.to_string(), }) .collect(), first_peer_id, })) } // Left for compatibility - does nothing async fn add_peer_as_participant( &self, _request: tonic::Request<PeerId>, ) -> Result<tonic::Response<()>, tonic::Status> { Ok(Response::new(())) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/points_api.rs
src/tonic/api/points_api.rs
use std::sync::Arc; use std::time::{Duration, Instant}; use api::grpc::Usage; use api::grpc::qdrant::points_server::Points; use api::grpc::qdrant::{ ClearPayloadPoints, CountPoints, CountResponse, CreateFieldIndexCollection, DeleteFieldIndexCollection, DeletePayloadPoints, DeletePointVectors, DeletePoints, DiscoverBatchPoints, DiscoverBatchResponse, DiscoverPoints, DiscoverResponse, FacetCounts, FacetResponse, GetPoints, GetResponse, PointsOperationResponse, QueryBatchPoints, QueryBatchResponse, QueryGroupsResponse, QueryPointGroups, QueryPoints, QueryResponse, RecommendBatchPoints, RecommendBatchResponse, RecommendGroupsResponse, RecommendPointGroups, RecommendPoints, RecommendResponse, ScrollPoints, ScrollResponse, SearchBatchPoints, SearchBatchResponse, SearchGroupsResponse, SearchMatrixOffsets, SearchMatrixOffsetsResponse, SearchMatrixPairs, SearchMatrixPairsResponse, SearchMatrixPoints, SearchPointGroups, SearchPoints, SearchResponse, SetPayloadPoints, UpdateBatchPoints, UpdateBatchResponse, UpdatePointVectors, UpsertPoints, }; use collection::operations::types::CoreSearchRequest; use common::counter::hardware_accumulator::HwMeasurementAcc; use storage::content_manager::toc::request_hw_counter::RequestHwCounter; use storage::dispatcher::Dispatcher; use tonic::{Request, Response, Status}; use super::query_common::*; use super::update_common::*; use super::validate; use crate::common::inference::params::InferenceParams; use crate::common::inference::token::extract_token; use crate::common::strict_mode::*; use crate::common::update::InternalUpdateParams; use crate::settings::ServiceConfig; use crate::tonic::auth::extract_access; pub struct PointsService { dispatcher: Arc<Dispatcher>, service_config: ServiceConfig, } impl PointsService { pub fn new(dispatcher: Arc<Dispatcher>, service_config: ServiceConfig) -> Self { Self { dispatcher, service_config, } } fn get_request_collection_hw_usage_counter( &self, collection_name: String, wait: Option<bool>, ) -> RequestHwCounter { let counter = HwMeasurementAcc::new_with_metrics_drain( self.dispatcher.get_collection_hw_metrics(collection_name), ); let waiting = wait != Some(false); RequestHwCounter::new(counter, self.service_config.hardware_reporting() && waiting) } } #[tonic::async_trait] impl Points for PointsService { async fn upsert( &self, mut request: Request<UpsertPoints>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let inference_token = extract_token(&request); let timeout = request.get_ref().timeout.map(Duration::from_secs); let inference_params = InferenceParams::new(inference_token, timeout); let collection_name = request.get_ref().collection_name.clone(); let wait = Some(request.get_ref().wait.unwrap_or(false)); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, wait); upsert( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), InternalUpdateParams::default(), access, inference_params, hw_metrics, ) .await .map(|resp| resp.map(PointsOperationResponse::from)) } async fn delete( &self, mut request: Request<DeletePoints>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let wait = Some(request.get_ref().wait.unwrap_or(false)); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, wait); delete( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), InternalUpdateParams::default(), access, hw_metrics, ) .await .map(|resp| resp.map(PointsOperationResponse::from)) } async fn get(&self, mut request: Request<GetPoints>) -> Result<Response<GetResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let inner_request = request.into_inner(); let hw_metrics = self .get_request_collection_hw_usage_counter(inner_request.collection_name.clone(), None); get( StrictModeCheckedTocProvider::new(&self.dispatcher), inner_request, None, access, hw_metrics, ) .await } async fn update_vectors( &self, mut request: Request<UpdatePointVectors>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; // Nothing to verify here. let access = extract_access(&mut request); let inference_token = extract_token(&request); let timeout = request.get_ref().timeout.map(Duration::from_secs); let inference_params = InferenceParams::new(inference_token, timeout); let collection_name = request.get_ref().collection_name.clone(); let wait = Some(request.get_ref().wait.unwrap_or(false)); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, wait); update_vectors( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), InternalUpdateParams::default(), access, inference_params, hw_metrics, ) .await .map(|resp| resp.map(PointsOperationResponse::from)) } async fn delete_vectors( &self, mut request: Request<DeletePointVectors>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let hw_metrics = self.get_request_collection_hw_usage_counter( request.get_ref().collection_name.clone(), None, ); delete_vectors( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), InternalUpdateParams::default(), access, hw_metrics, ) .await .map(|resp| resp.map(Into::into)) } async fn set_payload( &self, mut request: Request<SetPayloadPoints>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let wait = Some(request.get_ref().wait.unwrap_or(false)); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, wait); set_payload( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), InternalUpdateParams::default(), access, hw_metrics, ) .await .map(|resp| resp.map(Into::into)) } async fn overwrite_payload( &self, mut request: Request<SetPayloadPoints>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let wait = Some(request.get_ref().wait.unwrap_or(false)); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, wait); overwrite_payload( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), InternalUpdateParams::default(), access, hw_metrics, ) .await .map(|resp| resp.map(Into::into)) } async fn delete_payload( &self, mut request: Request<DeletePayloadPoints>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let wait = Some(request.get_ref().wait.unwrap_or(false)); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, wait); delete_payload( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), InternalUpdateParams::default(), access, hw_metrics, ) .await .map(|resp| resp.map(Into::into)) } async fn clear_payload( &self, mut request: Request<ClearPayloadPoints>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let wait = Some(request.get_ref().wait.unwrap_or(false)); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, wait); clear_payload( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), InternalUpdateParams::default(), access, hw_metrics, ) .await .map(|resp| resp.map(Into::into)) } async fn update_batch( &self, mut request: Request<UpdateBatchPoints>, ) -> Result<Response<UpdateBatchResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let inference_token = extract_token(&request); let timeout = request.get_ref().timeout.map(Duration::from_secs); let inference_params = InferenceParams::new(inference_token, timeout); let collection_name = request.get_ref().collection_name.clone(); let wait = Some(request.get_ref().wait.unwrap_or(false)); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, wait); update_batch( &self.dispatcher, request.into_inner(), InternalUpdateParams::default(), access, inference_params, hw_metrics, ) .await } async fn create_field_index( &self, mut request: Request<CreateFieldIndexCollection>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let wait = Some(request.get_ref().wait.unwrap_or(false)); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, wait); create_field_index( self.dispatcher.clone(), request.into_inner(), InternalUpdateParams::default(), access, hw_metrics, ) .await .map(|resp| resp.map(Into::into)) } async fn delete_field_index( &self, mut request: Request<DeleteFieldIndexCollection>, ) -> Result<Response<PointsOperationResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); delete_field_index( self.dispatcher.clone(), request.into_inner(), InternalUpdateParams::default(), access, ) .await .map(|resp| resp.map(Into::into)) } async fn search( &self, mut request: Request<SearchPoints>, ) -> Result<Response<SearchResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let res = search( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), None, access, hw_metrics, ) .await?; Ok(res) } async fn search_batch( &self, mut request: Request<SearchBatchPoints>, ) -> Result<Response<SearchBatchResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let SearchBatchPoints { collection_name, search_points, read_consistency, timeout, } = request.into_inner(); let timeout = timeout.map(Duration::from_secs); let mut requests = Vec::new(); for mut search_point in search_points { let shard_key = search_point.shard_key_selector.take(); let shard_selector = convert_shard_selector_for_read(None, shard_key)?; let core_search_request = CoreSearchRequest::try_from(search_point)?; requests.push((core_search_request, shard_selector)); } let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name.clone(), None); let res = core_search_batch( StrictModeCheckedTocProvider::new(&self.dispatcher), &collection_name, requests, read_consistency, access, timeout, hw_metrics, ) .await?; Ok(res) } async fn search_groups( &self, mut request: Request<SearchPointGroups>, ) -> Result<Response<SearchGroupsResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let res = search_groups( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), None, access, hw_metrics, ) .await?; Ok(res) } async fn scroll( &self, mut request: Request<ScrollPoints>, ) -> Result<Response<ScrollResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let inner_request = request.into_inner(); let hw_metrics = self .get_request_collection_hw_usage_counter(inner_request.collection_name.clone(), None); scroll( StrictModeCheckedTocProvider::new(&self.dispatcher), inner_request, None, access, hw_metrics, ) .await } async fn recommend( &self, mut request: Request<RecommendPoints>, ) -> Result<Response<RecommendResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let res = recommend( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), access, hw_metrics, ) .await?; Ok(res) } async fn recommend_batch( &self, mut request: Request<RecommendBatchPoints>, ) -> Result<Response<RecommendBatchResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let RecommendBatchPoints { collection_name, recommend_points, read_consistency, timeout, } = request.into_inner(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name.clone(), None); let res = recommend_batch( StrictModeCheckedTocProvider::new(&self.dispatcher), &collection_name, recommend_points, read_consistency, access, timeout.map(Duration::from_secs), hw_metrics, ) .await?; Ok(res) } async fn recommend_groups( &self, mut request: Request<RecommendPointGroups>, ) -> Result<Response<RecommendGroupsResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let res = recommend_groups( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), access, hw_metrics, ) .await?; Ok(res) } async fn discover( &self, mut request: Request<DiscoverPoints>, ) -> Result<Response<DiscoverResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let res = discover( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), access, hw_metrics, ) .await?; Ok(res) } async fn discover_batch( &self, mut request: Request<DiscoverBatchPoints>, ) -> Result<Response<DiscoverBatchResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let DiscoverBatchPoints { collection_name, discover_points, read_consistency, timeout, } = request.into_inner(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name.clone(), None); let res = discover_batch( StrictModeCheckedTocProvider::new(&self.dispatcher), &collection_name, discover_points, read_consistency, access, timeout.map(Duration::from_secs), hw_metrics, ) .await?; Ok(res) } async fn count( &self, mut request: Request<CountPoints>, ) -> Result<Response<CountResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let res = count( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), None, &access, hw_metrics, ) .await?; Ok(res) } async fn query( &self, mut request: Request<QueryPoints>, ) -> Result<Response<QueryResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let inference_token = extract_token(&request); let timeout = request.get_ref().timeout.map(Duration::from_secs); let inference_params = InferenceParams::new(inference_token, timeout); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let res = query( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), None, access, hw_metrics, inference_params, ) .await?; Ok(res) } async fn query_batch( &self, mut request: Request<QueryBatchPoints>, ) -> Result<Response<QueryBatchResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let inference_token = extract_token(&request); let timeout = request.get_ref().timeout.map(Duration::from_secs); let inference_params = InferenceParams::new(inference_token, timeout); let request = request.into_inner(); let QueryBatchPoints { collection_name, query_points, read_consistency, timeout, } = request; let timeout = timeout.map(Duration::from_secs); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name.clone(), None); let res = query_batch( StrictModeCheckedTocProvider::new(&self.dispatcher), &collection_name, query_points, read_consistency, access, timeout, hw_metrics, inference_params, ) .await?; Ok(res) } async fn query_groups( &self, mut request: Request<QueryPointGroups>, ) -> Result<Response<QueryGroupsResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let inference_token = extract_token(&request); let timeout = request.get_ref().timeout.map(Duration::from_secs); let inference_params = InferenceParams::new(inference_token, timeout); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let res = query_groups( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), None, access, hw_metrics, inference_params, ) .await?; Ok(res) } async fn facet( &self, mut request: Request<FacetCounts>, ) -> Result<Response<FacetResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let hw_metrics = self.get_request_collection_hw_usage_counter( request.get_ref().collection_name.clone(), None, ); facet( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), access, hw_metrics, ) .await } async fn search_matrix_pairs( &self, mut request: Request<SearchMatrixPoints>, ) -> Result<Response<SearchMatrixPairsResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let timing = Instant::now(); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let search_matrix_response = search_points_matrix( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), access, hw_metrics.get_counter(), ) .await?; let pairs_response = SearchMatrixPairsResponse { result: Some(SearchMatrixPairs::from(search_matrix_response)), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(hw_metrics.to_grpc_api()).into_non_empty(), }; Ok(Response::new(pairs_response)) } async fn search_matrix_offsets( &self, mut request: Request<SearchMatrixPoints>, ) -> Result<Response<SearchMatrixOffsetsResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let timing = Instant::now(); let collection_name = request.get_ref().collection_name.clone(); let hw_metrics = self.get_request_collection_hw_usage_counter(collection_name, None); let search_matrix_response = search_points_matrix( StrictModeCheckedTocProvider::new(&self.dispatcher), request.into_inner(), access, hw_metrics.get_counter(), ) .await?; let offsets_response = SearchMatrixOffsetsResponse { result: Some(SearchMatrixOffsets::from(search_matrix_response)), time: timing.elapsed().as_secs_f64(), usage: Usage::from_hardware_usage(hw_metrics.to_grpc_api()).into_non_empty(), }; Ok(Response::new(offsets_response)) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/collections_common.rs
src/tonic/api/collections_common.rs
use std::time::Instant; use api::grpc::qdrant::{GetCollectionInfoRequest, GetCollectionInfoResponse}; use collection::shards::shard::ShardId; use storage::content_manager::toc::TableOfContent; use storage::rbac::Access; use tonic::{Response, Status}; use crate::common::collections::do_get_collection; pub async fn get( toc: &TableOfContent, get_collection_info_request: GetCollectionInfoRequest, access: Access, shard_selection: Option<ShardId>, ) -> Result<Response<GetCollectionInfoResponse>, Status> { let timing = Instant::now(); let collection_name = get_collection_info_request.collection_name; let result = do_get_collection(toc, access, &collection_name, shard_selection).await?; let response = GetCollectionInfoResponse { result: Some(result.into()), time: timing.elapsed().as_secs_f64(), }; Ok(Response::new(response)) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/points_internal_api.rs
src/tonic/api/points_internal_api.rs
use std::any; use std::str::FromStr; use std::sync::Arc; use std::time::{Duration, Instant}; use api::grpc::HardwareUsage; use api::grpc::qdrant::points_internal_server::PointsInternal; use api::grpc::qdrant::{ ClearPayloadPointsInternal, CoreSearchBatchPointsInternal, CountPointsInternal, CountResponse, CreateFieldIndexCollectionInternal, DeleteFieldIndexCollectionInternal, DeletePayloadPointsInternal, DeletePointsInternal, DeleteVectorsInternal, FacetCountsInternal, FacetResponseInternal, GetPointsInternal, GetResponse, IntermediateResult, PointsOperationResponseInternal, QueryBatchPointsInternal, QueryBatchResponseInternal, QueryResultInternal, QueryShardPoints, RecommendPointsInternal, RecommendResponse, ScrollPointsInternal, ScrollResponse, SearchBatchResponse, SetPayloadPointsInternal, SyncPointsInternal, UpdateBatchInternal, UpdateVectorsInternal, UpsertPointsInternal, }; use api::grpc::update_operation::Update; use collection::operations::shard_selector_internal::ShardSelectorInternal; use collection::operations::universal_query::shard_query::ShardQueryRequest; use collection::shards::shard::ShardId; use common::counter::hardware_accumulator::HwMeasurementAcc; use itertools::Itertools; use segment::data_types::facets::{FacetParams, FacetResponse}; use segment::json_path::JsonPath; use segment::types::Filter; use storage::content_manager::toc::TableOfContent; use storage::content_manager::toc::request_hw_counter::RequestHwCounter; use storage::rbac::Access; use tonic::{Request, Response, Status}; use super::query_common::*; use super::update_common::*; use super::validate_and_log; use crate::common::inference::params::InferenceParams; use crate::common::inference::token::extract_token; use crate::common::strict_mode::*; use crate::common::update::InternalUpdateParams; use crate::settings::ServiceConfig; const FULL_ACCESS: Access = Access::full("Internal API"); /// This API is intended for P2P communication within a distributed deployment. pub struct PointsInternalService { toc: Arc<TableOfContent>, service_config: ServiceConfig, } impl PointsInternalService { pub fn new(toc: Arc<TableOfContent>, service_config: ServiceConfig) -> Self { Self { toc, service_config, } } async fn sync_internal( &self, sync_points_internal: SyncPointsInternal, inference_params: InferenceParams, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let SyncPointsInternal { sync_points, shard_id, clock_tag, } = sync_points_internal; let sync_points = extract_internal_request(sync_points)?; // Exclude the `inference_usage` for internal call let (response, _inference_usage) = sync( self.toc.clone(), sync_points, InternalUpdateParams::from_grpc(shard_id, clock_tag), FULL_ACCESS.clone(), inference_params, ) .await? .into_inner(); Ok(Response::new(response)) } async fn upsert_internal( &self, upsert_points_internal: UpsertPointsInternal, inference_params: InferenceParams, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let UpsertPointsInternal { upsert_points, shard_id, clock_tag, } = upsert_points_internal; let upsert_points = extract_internal_request(upsert_points)?; let hw_metrics = self.get_request_collection_hw_usage_counter_for_internal( upsert_points.collection_name.clone(), ); upsert( StrictModeCheckedInternalTocProvider::new(&self.toc), upsert_points, InternalUpdateParams::from_grpc(shard_id, clock_tag), FULL_ACCESS.clone(), inference_params.clone(), hw_metrics, ) .await } async fn delete_internal( &self, delete_points_internal: DeletePointsInternal, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let DeletePointsInternal { delete_points, shard_id, clock_tag, } = delete_points_internal; let delete_points = extract_internal_request(delete_points)?; let hw_metrics = self.get_request_collection_hw_usage_counter_for_internal( delete_points.collection_name.clone(), ); delete( UncheckedTocProvider::new_unchecked(&self.toc), delete_points, InternalUpdateParams::from_grpc(shard_id, clock_tag), FULL_ACCESS.clone(), hw_metrics, ) .await } async fn update_vectors_internal( &self, update_vectors_internal: UpdateVectorsInternal, inference_params: InferenceParams, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let UpdateVectorsInternal { update_vectors, shard_id, clock_tag, } = update_vectors_internal; let update_point_vectors = extract_internal_request(update_vectors)?; let hw_metrics = self.get_request_collection_hw_usage_counter_for_internal( update_point_vectors.collection_name.clone(), ); crate::tonic::api::update_common::update_vectors( StrictModeCheckedInternalTocProvider::new(&self.toc), update_point_vectors, InternalUpdateParams::from_grpc(shard_id, clock_tag), FULL_ACCESS.clone(), inference_params.clone(), hw_metrics, ) .await } async fn delete_vectors_internal( &self, delete_vectors_internal: DeleteVectorsInternal, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let DeleteVectorsInternal { delete_vectors, shard_id, clock_tag, } = delete_vectors_internal; let delete_point_vectors = extract_internal_request(delete_vectors)?; let hw_metrics = self.get_request_collection_hw_usage_counter_for_internal( delete_point_vectors.collection_name.clone(), ); crate::tonic::api::update_common::delete_vectors( UncheckedTocProvider::new_unchecked(&self.toc), delete_point_vectors, InternalUpdateParams::from_grpc(shard_id, clock_tag), FULL_ACCESS.clone(), hw_metrics, ) .await } async fn set_payload_internal( &self, set_payload_internal: SetPayloadPointsInternal, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let SetPayloadPointsInternal { set_payload_points, shard_id, clock_tag, } = set_payload_internal; let set_payload_points = extract_internal_request(set_payload_points)?; let hw_metrics = self.get_request_collection_hw_usage_counter_for_internal( set_payload_points.collection_name.clone(), ); set_payload( StrictModeCheckedInternalTocProvider::new(&self.toc), set_payload_points, InternalUpdateParams::from_grpc(shard_id, clock_tag), FULL_ACCESS.clone(), hw_metrics, ) .await } async fn overwrite_payload_internal( &self, overwrite_payload_internal: SetPayloadPointsInternal, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let SetPayloadPointsInternal { set_payload_points, shard_id, clock_tag, } = overwrite_payload_internal; let set_payload_points = extract_internal_request(set_payload_points)?; let hw_metrics = self.get_request_collection_hw_usage_counter_for_internal( set_payload_points.collection_name.clone(), ); overwrite_payload( StrictModeCheckedInternalTocProvider::new(&self.toc), set_payload_points, InternalUpdateParams::from_grpc(shard_id, clock_tag), FULL_ACCESS.clone(), hw_metrics, ) .await } async fn delete_payload_internal( &self, delete_payload_internal: DeletePayloadPointsInternal, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let DeletePayloadPointsInternal { delete_payload_points, shard_id, clock_tag, } = delete_payload_internal; let delete_payload_points = extract_internal_request(delete_payload_points)?; let hw_metrics = self.get_request_collection_hw_usage_counter_for_internal( delete_payload_points.collection_name.clone(), ); delete_payload( UncheckedTocProvider::new_unchecked(&self.toc), delete_payload_points, InternalUpdateParams::from_grpc(shard_id, clock_tag), FULL_ACCESS.clone(), hw_metrics, ) .await } async fn clear_payload_internal( &self, clear_payload_internal: ClearPayloadPointsInternal, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let ClearPayloadPointsInternal { clear_payload_points, shard_id, clock_tag, } = clear_payload_internal; let clear_payload_points = extract_internal_request(clear_payload_points)?; let hw_metrics = self.get_request_collection_hw_usage_counter_for_internal( clear_payload_points.collection_name.clone(), ); clear_payload( UncheckedTocProvider::new_unchecked(&self.toc), clear_payload_points, InternalUpdateParams::from_grpc(shard_id, clock_tag), FULL_ACCESS.clone(), hw_metrics, ) .await } async fn create_field_index_internal( &self, create_field_index_collection: CreateFieldIndexCollectionInternal, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let CreateFieldIndexCollectionInternal { create_field_index_collection, shard_id, clock_tag, } = create_field_index_collection; create_field_index_internal( self.toc.clone(), extract_internal_request(create_field_index_collection)?, InternalUpdateParams::from_grpc(shard_id, clock_tag), ) .await } async fn delete_field_index_internal( &self, delete_field_index_collection: DeleteFieldIndexCollectionInternal, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let DeleteFieldIndexCollectionInternal { delete_field_index_collection, shard_id, clock_tag, } = delete_field_index_collection; delete_field_index_internal( self.toc.clone(), extract_internal_request(delete_field_index_collection)?, InternalUpdateParams::from_grpc(shard_id, clock_tag), ) .await } } pub async fn query_batch_internal( toc: &TableOfContent, collection_name: String, query_points: Vec<QueryShardPoints>, shard_selection: Option<ShardId>, timeout: Option<Duration>, request_hw_data: RequestHwCounter, ) -> Result<Response<QueryBatchResponseInternal>, Status> { let batch_requests: Vec<_> = query_points .into_iter() .map(ShardQueryRequest::try_from) .try_collect()?; let timing = Instant::now(); // As this function is handling an internal request, // we can assume that shard_key is already resolved let shard_selection = match shard_selection { None => { debug_assert!(false, "Shard selection is expected for internal request"); ShardSelectorInternal::All } Some(shard_id) => ShardSelectorInternal::ShardId(shard_id), }; let batch_response = toc .query_batch_internal( &collection_name, batch_requests, shard_selection, timeout, request_hw_data.get_counter(), ) .await?; let response = QueryBatchResponseInternal { results: batch_response .into_iter() .map(|response| QueryResultInternal { intermediate_results: response .into_iter() .map(|intermediate| IntermediateResult { result: intermediate.into_iter().map(From::from).collect_vec(), }) .collect_vec(), }) .collect(), time: timing.elapsed().as_secs_f64(), hardware_usage: request_hw_data.to_grpc_api(), inference_usage: None, // No inference in internal API }; Ok(Response::new(response)) } async fn facet_counts_internal( toc: &TableOfContent, request: FacetCountsInternal, request_hw_data: RequestHwCounter, ) -> Result<Response<FacetResponseInternal>, Status> { let timing = Instant::now(); let FacetCountsInternal { collection_name, key, filter, limit, exact, shard_id, timeout, } = request; let shard_selection = ShardSelectorInternal::ShardId(shard_id); let request = FacetParams { key: JsonPath::from_str(&key) .map_err(|_| Status::invalid_argument("Failed to parse facet key"))?, limit: limit as usize, filter: filter.map(Filter::try_from).transpose()?, exact, }; let response = toc .facet_internal( &collection_name, request, shard_selection, timeout.map(Duration::from_secs), request_hw_data.get_counter(), ) .await?; let FacetResponse { hits } = response; let response = FacetResponseInternal { hits: hits.into_iter().map(From::from).collect_vec(), time: timing.elapsed().as_secs_f64(), usage: request_hw_data.to_grpc_api(), }; Ok(Response::new(response)) } impl PointsInternalService { /// Generates a new `RequestHwCounter` for the request. /// This counter is indented to be used for internal requests. fn get_request_collection_hw_usage_counter_for_internal( &self, collection_name: String, ) -> RequestHwCounter { let counter = HwMeasurementAcc::new_with_metrics_drain( self.toc.get_collection_hw_metrics(collection_name), ); RequestHwCounter::new(counter, self.service_config.hardware_reporting()) } } #[tonic::async_trait] impl PointsInternal for PointsInternalService { async fn upsert( &self, request: Request<UpsertPointsInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); let inference_token = extract_token(&request); let inference_params = InferenceParams::new(inference_token.clone(), None); self.upsert_internal(request.into_inner(), inference_params) .await } async fn delete( &self, request: Request<DeletePointsInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); self.delete_internal(request.into_inner()).await } async fn update_vectors( &self, request: Request<UpdateVectorsInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); let inference_token = extract_token(&request); let inference_params = InferenceParams::new(inference_token.clone(), None); self.update_vectors_internal(request.into_inner(), inference_params) .await } async fn delete_vectors( &self, request: Request<DeleteVectorsInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); self.delete_vectors_internal(request.into_inner()).await } async fn set_payload( &self, request: Request<SetPayloadPointsInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); self.set_payload_internal(request.into_inner()).await } async fn overwrite_payload( &self, request: Request<SetPayloadPointsInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); self.overwrite_payload_internal(request.into_inner()).await } async fn delete_payload( &self, request: Request<DeletePayloadPointsInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); self.delete_payload_internal(request.into_inner()).await } async fn clear_payload( &self, request: Request<ClearPayloadPointsInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); self.clear_payload_internal(request.into_inner()).await } async fn create_field_index( &self, request: Request<CreateFieldIndexCollectionInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); self.create_field_index_internal(request.into_inner()).await } async fn delete_field_index( &self, request: Request<DeleteFieldIndexCollectionInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); self.delete_field_index_internal(request.into_inner()).await } async fn update_batch( &self, request: Request<UpdateBatchInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); let inference_token = extract_token(&request); // Update operation doesn't specify explicit timeout yet let inference_params = InferenceParams::new(inference_token.clone(), None); let request_inner = request.into_inner(); let mut total_usage = HardwareUsage::default(); let mut last_result = None; // This API: // - Sequentially applies all operations // - If one operation fails, it will report the error immediately // - If no operations are present, it will return an empty response // - If all operations are successful, it will return the last operation result for update in request_inner.operations { let result = match update.update { None => { return Err(Status::invalid_argument("Update is missing")); } Some(update) => match update { Update::Sync(sync) => { self.sync_internal(sync, inference_params.clone()).await? } Update::Upsert(upsert) => { self.upsert_internal(upsert, inference_params.clone()) .await? } Update::Delete(delete) => self.delete_internal(delete).await?, Update::UpdateVectors(update_vectors) => { self.update_vectors_internal(update_vectors, inference_params.clone()) .await? } Update::DeleteVectors(delete_vectors) => { self.delete_vectors_internal(delete_vectors).await? } Update::SetPayload(set_payload) => { self.set_payload_internal(set_payload).await? } Update::OverwritePayload(overwrite_payload) => { self.overwrite_payload_internal(overwrite_payload).await? } Update::DeletePayload(delete_payload) => { self.delete_payload_internal(delete_payload).await? } Update::ClearPayload(clear_payload) => { self.clear_payload_internal(clear_payload).await? } Update::CreateFieldIndex(create_field_index) => { self.create_field_index_internal(create_field_index).await? } Update::DeleteFieldIndex(delete_field_index) => { self.delete_field_index_internal(delete_field_index).await? } }, }; let mut response = result.into_inner(); if let Some(usage) = response.hardware_usage.take() { total_usage.add(usage); } last_result = Some(response) } if let Some(mut last_result) = last_result.take() { last_result.hardware_usage = Some(total_usage); Ok(Response::new(last_result)) } else { // This response is possible if there are no operations in the request Ok(Response::new(PointsOperationResponseInternal { result: None, time: 0.0, hardware_usage: None, inference_usage: None, // No inference in internal API })) } } async fn core_search_batch( &self, request: Request<CoreSearchBatchPointsInternal>, ) -> Result<Response<SearchBatchResponse>, Status> { validate_and_log(request.get_ref()); let CoreSearchBatchPointsInternal { collection_name, search_points, shard_id, timeout, } = request.into_inner(); let timeout = timeout.map(Duration::from_secs); // Individual `read_consistency` values are ignored by `core_search_batch`... // // search_points // .iter_mut() // .for_each(|search_points| search_points.read_consistency = None); let hw_data = self.get_request_collection_hw_usage_counter_for_internal(collection_name.clone()); let res = core_search_list( self.toc.as_ref(), collection_name, search_points, None, // *Has* to be `None`! shard_id, FULL_ACCESS.clone(), timeout, hw_data, ) .await?; Ok(res) } async fn recommend( &self, request: Request<RecommendPointsInternal>, ) -> Result<Response<RecommendResponse>, Status> { validate_and_log(request.get_ref()); let RecommendPointsInternal { recommend_points, .. // shard_id - is not used in internal API, // because it is transformed into regular search requests on the first node } = request.into_inner(); let mut recommend_points = recommend_points .ok_or_else(|| Status::invalid_argument("RecommendPoints is missing"))?; recommend_points.read_consistency = None; // *Have* to be `None`! let collection_name = recommend_points.collection_name.clone(); let hw_data = self.get_request_collection_hw_usage_counter_for_internal(collection_name); let res = recommend( UncheckedTocProvider::new_unchecked(&self.toc), recommend_points, FULL_ACCESS.clone(), hw_data, ) .await?; Ok(res) } async fn scroll( &self, request: Request<ScrollPointsInternal>, ) -> Result<Response<ScrollResponse>, Status> { validate_and_log(request.get_ref()); let ScrollPointsInternal { scroll_points, shard_id, } = request.into_inner(); let mut scroll_points = scroll_points.ok_or_else(|| Status::invalid_argument("ScrollPoints is missing"))?; scroll_points.read_consistency = None; // *Have* to be `None`! let hw_data = self.get_request_collection_hw_usage_counter_for_internal( scroll_points.collection_name.clone(), ); scroll( UncheckedTocProvider::new_unchecked(&self.toc), scroll_points, shard_id, FULL_ACCESS.clone(), hw_data, ) .await } async fn get( &self, request: Request<GetPointsInternal>, ) -> Result<Response<GetResponse>, Status> { validate_and_log(request.get_ref()); let GetPointsInternal { get_points, shard_id, } = request.into_inner(); let mut get_points = get_points.ok_or_else(|| Status::invalid_argument("GetPoints is missing"))?; get_points.read_consistency = None; // *Have* to be `None`! let hw_data = self.get_request_collection_hw_usage_counter_for_internal( get_points.collection_name.clone(), ); get( UncheckedTocProvider::new_unchecked(&self.toc), get_points, shard_id, FULL_ACCESS.clone(), hw_data, ) .await } async fn count( &self, request: Request<CountPointsInternal>, ) -> Result<Response<CountResponse>, Status> { validate_and_log(request.get_ref()); let CountPointsInternal { count_points, shard_id, } = request.into_inner(); let count_points = count_points.ok_or_else(|| Status::invalid_argument("CountPoints is missing"))?; let hw_data = self.get_request_collection_hw_usage_counter_for_internal( count_points.collection_name.clone(), ); let res = count( UncheckedTocProvider::new_unchecked(&self.toc), count_points, shard_id, &FULL_ACCESS, hw_data, ) .await?; Ok(res) } async fn sync( &self, request: Request<SyncPointsInternal>, ) -> Result<Response<PointsOperationResponseInternal>, Status> { validate_and_log(request.get_ref()); let inference_token = extract_token(&request); // Internal operation, we don't expect timeout here let inference_params = InferenceParams::new(inference_token, None); self.sync_internal(request.into_inner(), inference_params) .await } async fn query_batch( &self, request: Request<QueryBatchPointsInternal>, ) -> Result<Response<QueryBatchResponseInternal>, Status> { validate_and_log(request.get_ref()); let QueryBatchPointsInternal { collection_name, shard_id, query_points, timeout, } = request.into_inner(); let timeout = timeout.map(Duration::from_secs); let hw_data = self.get_request_collection_hw_usage_counter_for_internal(collection_name.clone()); query_batch_internal( self.toc.as_ref(), collection_name, query_points, shard_id, timeout, hw_data, ) .await } async fn facet( &self, request: Request<FacetCountsInternal>, ) -> Result<Response<FacetResponseInternal>, Status> { validate_and_log(request.get_ref()); let request_inner = request.into_inner(); let hw_data = self.get_request_collection_hw_usage_counter_for_internal( request_inner.collection_name.clone(), ); facet_counts_internal(self.toc.as_ref(), request_inner, hw_data).await } } fn extract_internal_request<T>(request: Option<T>) -> Result<T, tonic::Status> { request.ok_or_else(|| { tonic::Status::invalid_argument(format!("{} is missing", any::type_name::<T>())) }) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/update_common.rs
src/tonic/api/update_common.rs
use std::sync::Arc; use std::time::Instant; use api::conversions::json::{json_path_from_proto, proto_to_payloads}; use api::grpc; use api::grpc::qdrant::payload_index_params::IndexParams; use api::grpc::qdrant::points_update_operation::{ClearPayload, Operation, PointStructList}; use api::grpc::qdrant::{ ClearPayloadPoints, CreateFieldIndexCollection, DeleteFieldIndexCollection, DeletePayloadPoints, DeletePointVectors, DeletePoints, FieldType, PayloadIndexParams, PointsOperationResponseInternal, PointsSelector, SetPayloadPoints, SyncPoints, UpdateBatchPoints, UpdateBatchResponse, UpdatePointVectors, UpsertPoints, points_update_operation, }; use api::grpc::{HardwareUsage, InferenceUsage, Usage}; use api::rest::schema::{PointInsertOperations, PointsList}; use api::rest::{PointStruct, PointVectors, ShardKeySelector, UpdateVectors, VectorStruct}; use collection::operations::CollectionUpdateOperations; use collection::operations::conversions::try_points_selector_from_grpc; use collection::operations::payload_ops::DeletePayload; use collection::operations::point_ops::{self, PointOperations, PointSyncOperation}; use collection::operations::vector_ops::DeleteVectors; use common::counter::hardware_accumulator::HwMeasurementAcc; use itertools::Itertools; use segment::types::{ ExtendedPointId, Filter, PayloadFieldSchema, PayloadSchemaParams, PayloadSchemaType, }; use storage::content_manager::toc::TableOfContent; use storage::content_manager::toc::request_hw_counter::RequestHwCounter; use storage::dispatcher::Dispatcher; use storage::rbac::Access; use tonic::{Response, Status}; use crate::common::inference::params::InferenceParams; use crate::common::inference::service::InferenceType; use crate::common::inference::update_requests::convert_point_struct; use crate::common::strict_mode::*; use crate::common::update::*; pub async fn upsert( toc_provider: impl CheckedTocProvider, upsert_points: UpsertPoints, internal_params: InternalUpdateParams, access: Access, inference_params: InferenceParams, request_hw_counter: RequestHwCounter, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let UpsertPoints { collection_name, wait, points, ordering, shard_key_selector, update_filter, timeout, } = upsert_points; let points: Result<_, _> = points.into_iter().map(PointStruct::try_from).collect(); let operation = PointInsertOperations::PointsList(PointsList { points: points?, shard_key: shard_key_selector .map(ShardKeySelector::try_from) .transpose()?, update_filter: update_filter .map(segment::types::Filter::try_from) .transpose()?, }); let timing = Instant::now(); let (result, inference_usage) = do_upsert_points( toc_provider, collection_name, operation, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, inference_params, request_hw_counter.get_counter(), ) .await?; let response = points_operation_response_internal_with_inference_usage( timing, result, request_hw_counter.to_grpc_api(), inference_usage.map(grpc::InferenceUsage::from), ); Ok(Response::new(response)) } pub async fn delete( toc_provider: impl CheckedTocProvider, delete_points: DeletePoints, internal_params: InternalUpdateParams, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let DeletePoints { collection_name, wait, points, ordering, shard_key_selector, timeout, } = delete_points; let points_selector = match points { None => return Err(Status::invalid_argument("PointSelector is missing")), Some(p) => try_points_selector_from_grpc(p, shard_key_selector)?, }; let timing = Instant::now(); let result = do_delete_points( toc_provider, collection_name, points_selector, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, request_hw_counter.get_counter(), ) .await?; let response = points_operation_response_internal(timing, result, request_hw_counter.to_grpc_api()); Ok(Response::new(response)) } pub async fn update_vectors( toc_provider: impl CheckedTocProvider, update_point_vectors: UpdatePointVectors, internal_params: InternalUpdateParams, access: Access, inference_params: InferenceParams, request_hw_counter: RequestHwCounter, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let UpdatePointVectors { collection_name, wait, points, ordering, shard_key_selector, update_filter, timeout, } = update_point_vectors; // Build list of operation points let mut op_points = Vec::with_capacity(points.len()); for point in points { let id = match point.id { Some(id) => id.try_into()?, None => return Err(Status::invalid_argument("id is expected")), }; let vector = match point.vectors { Some(vectors) => VectorStruct::try_from(vectors)?, None => return Err(Status::invalid_argument("vectors is expected")), }; op_points.push(PointVectors { id, vector }); } let operation = UpdateVectors { points: op_points, shard_key: shard_key_selector .map(ShardKeySelector::try_from) .transpose()?, update_filter: update_filter .map(segment::types::Filter::try_from) .transpose()?, }; let timing = Instant::now(); let (result, usage) = do_update_vectors( toc_provider, collection_name, operation, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, inference_params, request_hw_counter.get_counter(), ) .await?; let response = points_operation_response_internal_with_inference_usage( timing, result, request_hw_counter.to_grpc_api(), usage.map(grpc::InferenceUsage::from), ); Ok(Response::new(response)) } pub async fn delete_vectors( toc_provider: impl CheckedTocProvider, delete_point_vectors: DeletePointVectors, internal_params: InternalUpdateParams, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let DeletePointVectors { collection_name, wait, points_selector, vectors, ordering, shard_key_selector, timeout, } = delete_point_vectors; let (points, filter) = extract_points_selector(points_selector)?; let vector_names = match vectors { Some(vectors) => vectors.names, None => return Err(Status::invalid_argument("vectors is expected")), }; let operation = DeleteVectors { points, filter, vector: vector_names.into_iter().collect(), shard_key: shard_key_selector .map(ShardKeySelector::try_from) .transpose()?, }; let timing = Instant::now(); let result = do_delete_vectors( toc_provider, collection_name, operation, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, request_hw_counter.get_counter(), ) .await?; let response = points_operation_response_internal(timing, result, request_hw_counter.to_grpc_api()); Ok(Response::new(response)) } pub async fn set_payload( toc_provider: impl CheckedTocProvider, set_payload_points: SetPayloadPoints, internal_params: InternalUpdateParams, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let SetPayloadPoints { collection_name, wait, payload, points_selector, ordering, shard_key_selector, key, timeout, } = set_payload_points; let key = key.map(|k| json_path_from_proto(&k)).transpose()?; let (points, filter) = extract_points_selector(points_selector)?; let operation = collection::operations::payload_ops::SetPayload { payload: proto_to_payloads(payload)?, points, filter, shard_key: shard_key_selector .map(ShardKeySelector::try_from) .transpose()?, key, }; let timing = Instant::now(); let result = do_set_payload( toc_provider, collection_name, operation, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, request_hw_counter.get_counter(), ) .await?; let response = points_operation_response_internal(timing, result, request_hw_counter.to_grpc_api()); Ok(Response::new(response)) } pub async fn overwrite_payload( toc_provider: impl CheckedTocProvider, set_payload_points: SetPayloadPoints, internal_params: InternalUpdateParams, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let SetPayloadPoints { collection_name, wait, payload, points_selector, ordering, shard_key_selector, timeout, .. } = set_payload_points; let (points, filter) = extract_points_selector(points_selector)?; let operation = collection::operations::payload_ops::SetPayload { payload: proto_to_payloads(payload)?, points, filter, shard_key: shard_key_selector .map(ShardKeySelector::try_from) .transpose()?, // overwrite operation don't support indicate path of property key: None, }; let timing = Instant::now(); let result = do_overwrite_payload( toc_provider, collection_name, operation, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, request_hw_counter.get_counter(), ) .await?; let response = points_operation_response_internal(timing, result, request_hw_counter.to_grpc_api()); Ok(Response::new(response)) } pub async fn delete_payload( toc_provider: impl CheckedTocProvider, delete_payload_points: DeletePayloadPoints, internal_params: InternalUpdateParams, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let DeletePayloadPoints { collection_name, wait, keys, points_selector, ordering, shard_key_selector, timeout, } = delete_payload_points; let keys = keys.iter().map(|k| json_path_from_proto(k)).try_collect()?; let (points, filter) = extract_points_selector(points_selector)?; let operation = DeletePayload { keys, points, filter, shard_key: shard_key_selector .map(ShardKeySelector::try_from) .transpose()?, }; let timing = Instant::now(); let result = do_delete_payload( toc_provider, collection_name, operation, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, request_hw_counter.get_counter(), ) .await?; let response = points_operation_response_internal(timing, result, request_hw_counter.to_grpc_api()); Ok(Response::new(response)) } pub async fn clear_payload( toc_provider: impl CheckedTocProvider, clear_payload_points: ClearPayloadPoints, internal_params: InternalUpdateParams, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let ClearPayloadPoints { collection_name, wait, points, ordering, shard_key_selector, timeout, } = clear_payload_points; let points_selector = match points { None => return Err(Status::invalid_argument("PointSelector is missing")), Some(p) => try_points_selector_from_grpc(p, shard_key_selector)?, }; let timing = Instant::now(); let result = do_clear_payload( toc_provider, collection_name, points_selector, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, request_hw_counter.get_counter(), ) .await?; let response = points_operation_response_internal(timing, result, request_hw_counter.to_grpc_api()); Ok(Response::new(response)) } pub async fn update_batch( dispatcher: &Dispatcher, update_batch_points: UpdateBatchPoints, internal_params: InternalUpdateParams, access: Access, inference_params: InferenceParams, request_hw_counter: RequestHwCounter, ) -> Result<Response<UpdateBatchResponse>, Status> { let UpdateBatchPoints { collection_name, wait, operations, ordering, timeout, } = update_batch_points; let timing = Instant::now(); let mut results = Vec::with_capacity(operations.len()); let mut total_inference_usage = InferenceUsage::default(); for op in operations { let operation = op .operation .ok_or_else(|| Status::invalid_argument("Operation is missing"))?; let collection_name = collection_name.clone(); let ordering = ordering.clone(); let mut result = match operation { points_update_operation::Operation::Upsert(PointStructList { points, shard_key_selector, update_filter, }) => { upsert( StrictModeCheckedTocProvider::new(dispatcher), UpsertPoints { collection_name, wait, points, ordering, shard_key_selector, update_filter, timeout, }, internal_params, access.clone(), inference_params.clone(), request_hw_counter.clone(), ) .await } points_update_operation::Operation::DeleteDeprecated(points) => { delete( StrictModeCheckedTocProvider::new(dispatcher), DeletePoints { collection_name, wait, points: Some(points), ordering, shard_key_selector: None, timeout, }, internal_params, access.clone(), request_hw_counter.clone(), ) .await } points_update_operation::Operation::SetPayload( points_update_operation::SetPayload { payload, points_selector, shard_key_selector, key, }, ) => { set_payload( StrictModeCheckedTocProvider::new(dispatcher), SetPayloadPoints { collection_name, wait, payload, points_selector, ordering, shard_key_selector, key, timeout, }, internal_params, access.clone(), request_hw_counter.clone(), ) .await } points_update_operation::Operation::OverwritePayload( points_update_operation::OverwritePayload { payload, points_selector, shard_key_selector, .. }, ) => { overwrite_payload( StrictModeCheckedTocProvider::new(dispatcher), SetPayloadPoints { collection_name, wait, payload, points_selector, ordering, shard_key_selector, // overwrite operation doesn't support it key: None, timeout, }, internal_params, access.clone(), request_hw_counter.clone(), ) .await } points_update_operation::Operation::DeletePayload( points_update_operation::DeletePayload { keys, points_selector, shard_key_selector, }, ) => { delete_payload( StrictModeCheckedTocProvider::new(dispatcher), DeletePayloadPoints { collection_name, wait, keys, points_selector, ordering, shard_key_selector, timeout, }, internal_params, access.clone(), request_hw_counter.clone(), ) .await } points_update_operation::Operation::ClearPayload(ClearPayload { points, shard_key_selector, }) => { clear_payload( StrictModeCheckedTocProvider::new(dispatcher), ClearPayloadPoints { collection_name, wait, points, ordering, shard_key_selector, timeout, }, internal_params, access.clone(), request_hw_counter.clone(), ) .await } points_update_operation::Operation::UpdateVectors( points_update_operation::UpdateVectors { points, shard_key_selector, update_filter, }, ) => { update_vectors( StrictModeCheckedTocProvider::new(dispatcher), UpdatePointVectors { collection_name, wait, points, ordering, shard_key_selector, update_filter, timeout, }, internal_params, access.clone(), inference_params.clone(), request_hw_counter.clone(), ) .await } points_update_operation::Operation::DeleteVectors( points_update_operation::DeleteVectors { points_selector, vectors, shard_key_selector, }, ) => { delete_vectors( StrictModeCheckedTocProvider::new(dispatcher), DeletePointVectors { collection_name, wait, points_selector, vectors, ordering, shard_key_selector, timeout, }, internal_params, access.clone(), request_hw_counter.clone(), ) .await } Operation::ClearPayloadDeprecated(selector) => { clear_payload( StrictModeCheckedTocProvider::new(dispatcher), ClearPayloadPoints { collection_name, wait, points: Some(selector), ordering, shard_key_selector: None, timeout, }, internal_params, access.clone(), request_hw_counter.clone(), ) .await } Operation::DeletePoints(points_update_operation::DeletePoints { points, shard_key_selector, }) => { delete( StrictModeCheckedTocProvider::new(dispatcher), DeletePoints { collection_name, wait, points, ordering, shard_key_selector, timeout, }, internal_params, access.clone(), request_hw_counter.clone(), ) .await } }?; total_inference_usage.merge_opt(result.get_mut().inference_usage.take()); results.push(result); } Ok(Response::new(UpdateBatchResponse { result: results .into_iter() .map(|response| grpc::UpdateResult::from(response.into_inner().result.unwrap())) .collect(), time: timing.elapsed().as_secs_f64(), usage: Usage::new( request_hw_counter.to_grpc_api(), total_inference_usage.into_non_empty(), ) .into_non_empty(), })) } pub async fn create_field_index( dispatcher: Arc<Dispatcher>, create_field_index_collection: CreateFieldIndexCollection, internal_params: InternalUpdateParams, access: Access, request_hw_counter: RequestHwCounter, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let CreateFieldIndexCollection { collection_name, wait, field_name, field_type, field_index_params, ordering, timeout, } = create_field_index_collection; let field_name = json_path_from_proto(&field_name)?; let field_schema = convert_field_type(field_type, field_index_params)?; let operation = CreateFieldIndex { field_name, field_schema, }; let timing = Instant::now(); let result = do_create_index( dispatcher, collection_name, operation, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, request_hw_counter.get_counter(), ) .await?; let response = points_operation_response_internal( timing, result, None, // Do not measure API usage for this operation, as it might be inaccurate due to consensus involvement ); Ok(Response::new(response)) } pub async fn create_field_index_internal( toc: Arc<TableOfContent>, create_field_index_collection: CreateFieldIndexCollection, internal_params: InternalUpdateParams, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let CreateFieldIndexCollection { collection_name, wait, field_name, field_type, field_index_params, ordering, timeout, } = create_field_index_collection; let field_name = json_path_from_proto(&field_name)?; let field_schema = convert_field_type(field_type, field_index_params)?; let timing = Instant::now(); let result = do_create_index_internal( toc, collection_name, field_name, field_schema, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, HwMeasurementAcc::disposable(), // API unmeasured ) .await?; let response = points_operation_response_internal(timing, result, None); Ok(Response::new(response)) } pub async fn delete_field_index( dispatcher: Arc<Dispatcher>, delete_field_index_collection: DeleteFieldIndexCollection, internal_params: InternalUpdateParams, access: Access, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let DeleteFieldIndexCollection { collection_name, wait, field_name, ordering, timeout, } = delete_field_index_collection; let field_name = json_path_from_proto(&field_name)?; let timing = Instant::now(); let result = do_delete_index( dispatcher, collection_name, field_name, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, access, HwMeasurementAcc::disposable(), // API unmeasured ) .await?; let response = points_operation_response_internal(timing, result, None); Ok(Response::new(response)) } pub async fn delete_field_index_internal( toc: Arc<TableOfContent>, delete_field_index_collection: DeleteFieldIndexCollection, internal_params: InternalUpdateParams, ) -> Result<Response<PointsOperationResponseInternal>, Status> { let DeleteFieldIndexCollection { collection_name, wait, field_name, ordering, timeout, } = delete_field_index_collection; let field_name = json_path_from_proto(&field_name)?; let timing = Instant::now(); let result = do_delete_index_internal( toc, collection_name, field_name, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, HwMeasurementAcc::disposable(), // API unmeasured ) .await?; let response = points_operation_response_internal(timing, result, None); Ok(Response::new(response)) } pub async fn sync( toc: Arc<TableOfContent>, sync_points: SyncPoints, internal_params: InternalUpdateParams, access: Access, inference_params: InferenceParams, ) -> Result<Response<(PointsOperationResponseInternal, InferenceUsage)>, Status> { let SyncPoints { collection_name, wait, points, from_id, to_id, ordering, timeout, } = sync_points; let timing = Instant::now(); let point_structs: Result<_, _> = points.into_iter().map(PointStruct::try_from).collect(); // No actual inference should happen here, as we are just syncing existing points // So, this function is used for consistency only let (points, usage) = convert_point_struct(point_structs?, InferenceType::Update, inference_params).await?; let operation = PointSyncOperation { points, from_id: from_id.map(|x| x.try_into()).transpose()?, to_id: to_id.map(|x| x.try_into()).transpose()?, }; let operation = CollectionUpdateOperations::PointOperation(PointOperations::SyncPoints(operation)); let result = update( &toc, &collection_name, operation, internal_params, UpdateParams::from_grpc(wait, ordering, timeout.map(std::time::Duration::from_secs))?, None, access, HwMeasurementAcc::disposable(), // API unmeasured ) .await?; let response = points_operation_response_internal(timing, result, None); Ok(Response::new((response, usage.unwrap_or_default().into()))) } pub fn points_operation_response_internal_with_inference_usage( timing: Instant, update_result: collection::operations::types::UpdateResult, hardware_usage: Option<HardwareUsage>, inference_usage: Option<InferenceUsage>, ) -> PointsOperationResponseInternal { PointsOperationResponseInternal { result: Some(update_result.into()), time: timing.elapsed().as_secs_f64(), hardware_usage, inference_usage, } } pub fn points_operation_response_internal( timing: Instant, update_result: collection::operations::types::UpdateResult, hardware_usage: Option<HardwareUsage>, ) -> PointsOperationResponseInternal { points_operation_response_internal_with_inference_usage( timing, update_result, hardware_usage, None, // No inference usage for this operation ) } fn extract_points_selector( points_selector: Option<PointsSelector>, ) -> Result<(Option<Vec<ExtendedPointId>>, Option<Filter>), Status> { let (points, filter) = if let Some(points_selector) = points_selector { let points_selector = try_points_selector_from_grpc(points_selector, None)?; match points_selector { point_ops::PointsSelector::PointIdsSelector(points) => (Some(points.points), None), point_ops::PointsSelector::FilterSelector(filter) => (None, Some(filter.filter)), } } else { return Err(Status::invalid_argument("points_selector is expected")); }; Ok((points, filter)) } fn convert_field_type( field_type: Option<i32>, field_index_params: Option<PayloadIndexParams>, ) -> Result<Option<PayloadFieldSchema>, Status> { let field_type_parsed = field_type .map(|x| FieldType::try_from(x).ok()) .ok_or_else(|| Status::invalid_argument("cannot convert field_type"))?; let field_schema = match (field_type_parsed, field_index_params) { ( Some(field_type), Some(PayloadIndexParams { index_params: Some(index_params), }), ) => { let schema_params = match index_params { // Parameterized keyword type IndexParams::KeywordIndexParams(keyword_index_params) => { matches!(field_type, FieldType::Keyword).then(|| { TryFrom::try_from(keyword_index_params).map(PayloadSchemaParams::Keyword) }) } IndexParams::IntegerIndexParams(integer_index_params) => { matches!(field_type, FieldType::Integer).then(|| { TryFrom::try_from(integer_index_params).map(PayloadSchemaParams::Integer) }) } // Parameterized float type IndexParams::FloatIndexParams(float_index_params) => { matches!(field_type, FieldType::Float).then(|| { TryFrom::try_from(float_index_params).map(PayloadSchemaParams::Float) }) } IndexParams::GeoIndexParams(geo_index_params) => { matches!(field_type, FieldType::Geo) .then(|| TryFrom::try_from(geo_index_params).map(PayloadSchemaParams::Geo)) } // Parameterized text type IndexParams::TextIndexParams(text_index_params) => { matches!(field_type, FieldType::Text).then(|| { TryFrom::try_from(text_index_params).map(PayloadSchemaParams::Text) }) } // Parameterized bool type IndexParams::BoolIndexParams(bool_index_params) => { matches!(field_type, FieldType::Bool).then(|| { TryFrom::try_from(bool_index_params).map(PayloadSchemaParams::Bool) }) } // Parameterized Datetime type IndexParams::DatetimeIndexParams(datetime_index_params) => { matches!(field_type, FieldType::Datetime).then(|| { TryFrom::try_from(datetime_index_params).map(PayloadSchemaParams::Datetime) }) } // Parameterized Uuid type
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
true
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/tonic/api/snapshots_api.rs
src/tonic/api/snapshots_api.rs
use std::sync::Arc; use std::time::Instant; use api::grpc::qdrant::shard_snapshots_server::ShardSnapshots; use api::grpc::qdrant::snapshots_server::Snapshots; use api::grpc::qdrant::{ CreateFullSnapshotRequest, CreateShardSnapshotRequest, CreateSnapshotRequest, CreateSnapshotResponse, DeleteFullSnapshotRequest, DeleteShardSnapshotRequest, DeleteSnapshotRequest, DeleteSnapshotResponse, ListFullSnapshotsRequest, ListShardSnapshotsRequest, ListSnapshotsRequest, ListSnapshotsResponse, RecoverShardSnapshotRequest, RecoverSnapshotResponse, }; use collection::operations::verification::new_unchecked_verification_pass; use storage::content_manager::snapshots::{ do_create_full_snapshot, do_delete_collection_snapshot, do_delete_full_snapshot, do_list_full_snapshots, }; use storage::content_manager::toc::TableOfContent; use storage::dispatcher::Dispatcher; use tonic::{Request, Response, Status, async_trait}; use super::{validate, validate_and_log}; use crate::common; use crate::common::collections::{do_create_snapshot, do_list_snapshots}; use crate::common::http_client::HttpClient; use crate::tonic::auth::extract_access; pub struct SnapshotsService { dispatcher: Arc<Dispatcher>, } impl SnapshotsService { pub fn new(dispatcher: Arc<Dispatcher>) -> Self { Self { dispatcher } } } #[async_trait] impl Snapshots for SnapshotsService { async fn create( &self, mut request: Request<CreateSnapshotRequest>, ) -> Result<Response<CreateSnapshotResponse>, Status> { validate(request.get_ref())?; let access = extract_access(&mut request); let collection_name = request.into_inner().collection_name; let timing = Instant::now(); let dispatcher = self.dispatcher.clone(); // Nothing to verify here. let pass = new_unchecked_verification_pass(); let response = do_create_snapshot( Arc::clone(dispatcher.toc(&access, &pass)), access, &collection_name, ) .await?; Ok(Response::new(CreateSnapshotResponse { snapshot_description: Some(response.into()), time: timing.elapsed().as_secs_f64(), })) } async fn list( &self, mut request: Request<ListSnapshotsRequest>, ) -> Result<Response<ListSnapshotsResponse>, Status> { validate(request.get_ref())?; let timing = Instant::now(); let access = extract_access(&mut request); let ListSnapshotsRequest { collection_name } = request.into_inner(); // Nothing to verify here. let pass = new_unchecked_verification_pass(); let snapshots = do_list_snapshots( self.dispatcher.toc(&access, &pass), access, &collection_name, ) .await?; Ok(Response::new(ListSnapshotsResponse { snapshot_descriptions: snapshots.into_iter().map(|s| s.into()).collect(), time: timing.elapsed().as_secs_f64(), })) } async fn delete( &self, mut request: Request<DeleteSnapshotRequest>, ) -> Result<Response<DeleteSnapshotResponse>, Status> { validate(request.get_ref())?; let timing = Instant::now(); let access = extract_access(&mut request); let DeleteSnapshotRequest { collection_name, snapshot_name, } = request.into_inner(); let _response = do_delete_collection_snapshot( &self.dispatcher, access, &collection_name, &snapshot_name, ) .await?; Ok(Response::new(DeleteSnapshotResponse { time: timing.elapsed().as_secs_f64(), })) } async fn create_full( &self, mut request: Request<CreateFullSnapshotRequest>, ) -> Result<Response<CreateSnapshotResponse>, Status> { validate(request.get_ref())?; let timing = Instant::now(); let access = extract_access(&mut request); let response = do_create_full_snapshot(&self.dispatcher, access).await?; Ok(Response::new(CreateSnapshotResponse { snapshot_description: Some(response.into()), time: timing.elapsed().as_secs_f64(), })) } async fn list_full( &self, mut request: Request<ListFullSnapshotsRequest>, ) -> Result<Response<ListSnapshotsResponse>, Status> { validate(request.get_ref())?; let timing = Instant::now(); let access = extract_access(&mut request); // Nothing to verify here. let pass = new_unchecked_verification_pass(); let snapshots = do_list_full_snapshots(self.dispatcher.toc(&access, &pass), access).await?; Ok(Response::new(ListSnapshotsResponse { snapshot_descriptions: snapshots.into_iter().map(|s| s.into()).collect(), time: timing.elapsed().as_secs_f64(), })) } async fn delete_full( &self, mut request: Request<DeleteFullSnapshotRequest>, ) -> Result<Response<DeleteSnapshotResponse>, Status> { validate(request.get_ref())?; let timing = Instant::now(); let access = extract_access(&mut request); let snapshot_name = request.into_inner().snapshot_name; let _response = do_delete_full_snapshot(&self.dispatcher, access, &snapshot_name).await?; Ok(Response::new(DeleteSnapshotResponse { time: timing.elapsed().as_secs_f64(), })) } } pub struct ShardSnapshotsService { toc: Arc<TableOfContent>, http_client: HttpClient, } impl ShardSnapshotsService { pub fn new(toc: Arc<TableOfContent>, http_client: HttpClient) -> Self { Self { toc, http_client } } } #[async_trait] impl ShardSnapshots for ShardSnapshotsService { async fn create( &self, mut request: Request<CreateShardSnapshotRequest>, ) -> Result<Response<CreateSnapshotResponse>, Status> { let access = extract_access(&mut request); let request = request.into_inner(); validate_and_log(&request); let timing = Instant::now(); let snapshot_description = common::snapshots::create_shard_snapshot( self.toc.clone(), access, request.collection_name, request.shard_id, ) .await?; Ok(Response::new(CreateSnapshotResponse { snapshot_description: Some(snapshot_description.into()), time: timing.elapsed().as_secs_f64(), })) } async fn list( &self, mut request: Request<ListShardSnapshotsRequest>, ) -> Result<Response<ListSnapshotsResponse>, Status> { let access = extract_access(&mut request); let request = request.into_inner(); validate_and_log(&request); let timing = Instant::now(); let snapshot_descriptions = common::snapshots::list_shard_snapshots( self.toc.clone(), access, request.collection_name, request.shard_id, ) .await?; Ok(Response::new(ListSnapshotsResponse { snapshot_descriptions: snapshot_descriptions.into_iter().map(Into::into).collect(), time: timing.elapsed().as_secs_f64(), })) } async fn delete( &self, mut request: Request<DeleteShardSnapshotRequest>, ) -> Result<Response<DeleteSnapshotResponse>, Status> { let access = extract_access(&mut request); let request = request.into_inner(); validate_and_log(&request); let timing = Instant::now(); common::snapshots::delete_shard_snapshot( self.toc.clone(), access, request.collection_name, request.shard_id, request.snapshot_name, ) .await?; Ok(Response::new(DeleteSnapshotResponse { time: timing.elapsed().as_secs_f64(), })) } async fn recover( &self, mut request: Request<RecoverShardSnapshotRequest>, ) -> Result<Response<RecoverSnapshotResponse>, Status> { let access = extract_access(&mut request); let request = request.into_inner(); validate_and_log(&request); let timing = Instant::now(); common::snapshots::recover_shard_snapshot( self.toc.clone(), access, request.collection_name, request.shard_id, request.snapshot_location.try_into()?, request.snapshot_priority.try_into()?, request.checksum, self.http_client.clone(), request.api_key, ) .await?; Ok(Response::new(RecoverSnapshotResponse { time: timing.elapsed().as_secs_f64(), })) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/telemetry_reporting.rs
src/common/telemetry_reporting.rs
use std::sync::Arc; use std::time::Duration; use common::defaults::APP_USER_AGENT; use common::types::{DetailsLevel, TelemetryDetail}; use reqwest::Client; use segment::common::anonymize::Anonymize; use storage::content_manager::errors::StorageResult; use storage::rbac::Access; use tokio::sync::Mutex; use super::telemetry::TelemetryCollector; const DETAIL: TelemetryDetail = TelemetryDetail { level: DetailsLevel::Level2, histograms: false, }; const REPORTING_INTERVAL: Duration = Duration::from_secs(60 * 60); // One hour pub struct TelemetryReporter { telemetry_url: String, telemetry: Arc<Mutex<TelemetryCollector>>, } const FULL_ACCESS: Access = Access::full("Telemetry reporter"); impl TelemetryReporter { fn new(telemetry: Arc<Mutex<TelemetryCollector>>) -> Self { let telemetry_url = if cfg!(debug_assertions) { "https://staging-telemetry.qdrant.io".to_string() } else { "https://telemetry.qdrant.io".to_string() }; Self { telemetry_url, telemetry, } } async fn report(&self, client: &Client) -> StorageResult<()> { let data = self .telemetry .lock() .await .prepare_data(&FULL_ACCESS, DETAIL, None) .await? .anonymize(); let data = serde_json::to_string(&data)?; let resp = client .post(&self.telemetry_url) .body(data) .header("Content-Type", "application/json") .send() .await?; if !resp.status().is_success() { log::error!( "Failed to report telemetry: resp status:{:?} resp body:{:?}", resp.status(), resp.text().await? ); } Ok(()) } pub async fn run(telemetry: Arc<Mutex<TelemetryCollector>>) { let reporter = Self::new(telemetry); let client = Client::builder() .user_agent(APP_USER_AGENT.as_str()) .build() .unwrap(); loop { if let Err(err) = reporter.report(&client).await { log::error!("Failed to report telemetry {err}") } tokio::time::sleep(REPORTING_INTERVAL).await; } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/pyroscope_state.rs
src/common/pyroscope_state.rs
#[cfg(target_os = "linux")] pub mod pyro { use pyroscope::pyroscope::PyroscopeAgentRunning; use pyroscope::{PyroscopeAgent, PyroscopeError}; use pyroscope_pprofrs::{PprofConfig, pprof_backend}; use crate::common::debugger::PyroscopeConfig; pub struct PyroscopeState { pub config: PyroscopeConfig, pub agent: Option<PyroscopeAgent<PyroscopeAgentRunning>>, } impl PyroscopeState { fn build_agent( config: &PyroscopeConfig, ) -> Result<PyroscopeAgent<PyroscopeAgentRunning>, PyroscopeError> { let pprof_config = PprofConfig::new().sample_rate(config.sampling_rate.unwrap_or(100)); let backend_impl = pprof_backend(pprof_config); log::info!( "Starting pyroscope agent with identifier {}", &config.identifier ); // TODO: Add more tags like peerId and peerUrl let agent = PyroscopeAgent::builder(config.url.clone(), "qdrant".to_string()) .backend(backend_impl) .tags(vec![("app", "Qdrant"), ("identifier", &config.identifier)]) .build()?; let running_agent = agent.start()?; Ok(running_agent) } pub fn from_config(config: Option<PyroscopeConfig>) -> Option<Self> { match config { Some(pyro_config) => { let agent = PyroscopeState::build_agent(&pyro_config); match agent { Ok(agent) => Some(PyroscopeState { config: pyro_config, agent: Some(agent), }), Err(err) => { log::warn!("Pyroscope agent failed to start {err}"); None } } } None => None, } } pub fn stop_agent(&mut self) -> bool { log::info!("Stopping pyroscope agent"); if let Some(agent) = self.agent.take() { return match agent.stop() { Ok(stopped_agent) => { log::info!("Stopped pyroscope agent. Shutting it down"); stopped_agent.shutdown(); log::info!("Pyroscope agent shut down completed."); true } Err(err) => { log::warn!("Pyroscope agent failed to stop {err}"); false } }; } true } } impl Drop for PyroscopeState { fn drop(&mut self) { self.stop_agent(); } } } #[cfg(not(target_os = "linux"))] pub mod pyro { use crate::common::debugger::PyroscopeConfig; pub struct PyroscopeState {} impl PyroscopeState { pub fn from_config(_config: Option<PyroscopeConfig>) -> Option<Self> { None } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/http_client.rs
src/common/http_client.rs
use std::path::Path; use std::{io, result}; use common::defaults::APP_USER_AGENT; use fs_err as fs; use reqwest::header::{HeaderMap, HeaderValue, InvalidHeaderValue}; use storage::content_manager::errors::StorageError; use super::auth::HTTP_HEADER_API_KEY; use crate::settings::{Settings, TlsConfig}; #[derive(Clone)] pub struct HttpClient { tls_config: Option<TlsConfig>, verify_https_client_certificate: bool, } impl HttpClient { pub fn from_settings(settings: &Settings) -> Result<Self> { let tls_config = if settings.service.enable_tls { let Some(tls_config) = settings.tls.clone() else { return Err(Error::TlsConfigUndefined); }; Some(tls_config) } else { None }; let verify_https_client_certificate = settings.service.verify_https_client_certificate; let http_client = Self { tls_config, verify_https_client_certificate, }; Ok(http_client) } /// Create a new HTTP(S) client /// /// An API key can be optionally provided to be used in this HTTP client. It'll send the API /// key as `Api-key` header in every request. /// /// # Warning /// /// Setting an API key may leak when the client is used to send a request to a malicious /// server. This is potentially dangerous if a user has control over what URL is accessed. /// /// For this reason the API key is not set by default as provided in the configuration. It must /// be explicitly provided when creating the HTTP client. pub fn client(&self, api_key: Option<&str>) -> Result<reqwest::Client> { https_client( api_key, self.tls_config.as_ref(), self.verify_https_client_certificate, ) } } fn https_client( api_key: Option<&str>, tls_config: Option<&TlsConfig>, verify_https_client_certificate: bool, ) -> Result<reqwest::Client> { let mut builder = reqwest::Client::builder().user_agent(APP_USER_AGENT.as_str()); // Configure TLS root certificate and validation if let Some(tls_config) = tls_config { if let Some(ca_cert) = &tls_config.ca_cert { match https_client_ca_cert(ca_cert) { Ok(ca_cert) => builder = builder.add_root_certificate(ca_cert), Err(err) => { // I think it might be Ok to not fail here, if root certificate is not found // There are 2 possible scenarios: // // 1. Server TLS is either not used or it uses some other CA certificate (like global one) // 2. Server TLS is using self-signed certificate, and we should have it. // // In first case, we don't need to load the CA certificate, everything will work in either case. // In second case, we should have the CA certificate, request will fail because of invalid certificate. // // So both scenarios work exactly the same way if we fail early or not. // Warning message is needed for easier debugging in case of second scenario. log::warn!( "Failed to load CA certificate, skipping HTTPS client CA certificate configuration: {err}", ); } } } if verify_https_client_certificate { builder = builder.identity(https_client_identity( tls_config.cert.as_ref(), tls_config.key.as_ref(), )?); } } // Attach API key as sensitive header if let Some(api_key) = api_key { let mut headers = HeaderMap::new(); let mut api_key_value = HeaderValue::from_str(api_key).map_err(Error::MalformedApiKey)?; api_key_value.set_sensitive(true); headers.insert(HTTP_HEADER_API_KEY, api_key_value); builder = builder.default_headers(headers); } let client = builder.build()?; Ok(client) } fn https_client_ca_cert(ca_cert: impl AsRef<Path>) -> Result<reqwest::tls::Certificate> { let ca_cert_pem = fs::read(ca_cert.as_ref()) .map_err(|err| Error::failed_to_read(err, "CA certificate", ca_cert.as_ref()))?; let ca_cert = reqwest::Certificate::from_pem(&ca_cert_pem)?; Ok(ca_cert) } fn https_client_identity(cert: &Path, key: &Path) -> Result<reqwest::tls::Identity> { let mut identity_pem = fs::read(cert).map_err(|err| Error::failed_to_read(err, "certificate", cert))?; let mut key_file = fs::File::open(key).map_err(|err| Error::failed_to_read(err, "key", key))?; // Concatenate certificate and key into a single PEM bytes io::copy(&mut key_file, &mut identity_pem) .map_err(|err| Error::failed_to_read(err, "key", key))?; let identity = reqwest::Identity::from_pem(&identity_pem)?; Ok(identity) } pub type Result<T, E = Error> = result::Result<T, E>; #[derive(Debug, thiserror::Error)] pub enum Error { #[error("TLS config is not defined in the Qdrant config file")] TlsConfigUndefined, #[error("{1}: {0}")] Io(#[source] io::Error, String), #[error("failed to setup HTTPS client: {0}")] Reqwest(#[from] reqwest::Error), #[error("malformed API key")] MalformedApiKey(#[source] InvalidHeaderValue), } impl Error { pub fn io(source: io::Error, context: impl Into<String>) -> Self { Self::Io(source, context.into()) } pub fn failed_to_read(source: io::Error, file: &str, path: &Path) -> Self { Self::io( source, format!("failed to read HTTPS client {file} file {}", path.display()), ) } } impl From<Error> for StorageError { fn from(err: Error) -> Self { StorageError::service_error(format!("failed to initialize HTTP(S) client: {err}")) } } impl From<Error> for io::Error { fn from(err: Error) -> Self { io::Error::other(err) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/error_reporting.rs
src/common/error_reporting.rs
use std::time::Duration; use common::defaults::APP_USER_AGENT; pub struct ErrorReporter; impl ErrorReporter { fn get_url() -> String { if cfg!(debug_assertions) { "https://staging-telemetry.qdrant.io".to_string() } else { "https://telemetry.qdrant.io".to_string() } } pub fn report(error: &str, reporting_id: &str, backtrace: Option<&str>) { let client = reqwest::blocking::Client::builder() .user_agent(APP_USER_AGENT.as_str()) .build() .unwrap(); let report = serde_json::json!({ "id": reporting_id, "error": error, "backtrace": backtrace.unwrap_or(""), }); let data = serde_json::to_string(&report).unwrap(); let _resp = client .post(Self::get_url()) .body(data) .header("Content-Type", "application/json") .timeout(Duration::from_secs(1)) .send(); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/update.rs
src/common/update.rs
use std::sync::Arc; use std::time::Duration; use api::rest::models::InferenceUsage; use api::rest::*; use collection::collection::Collection; use collection::operations::conversions::write_ordering_from_proto; use collection::operations::point_ops::*; use collection::operations::shard_selector_internal::ShardSelectorInternal; use collection::operations::types::{CollectionError, CollectionResult, UpdateResult}; use collection::operations::vector_ops::*; use collection::operations::verification::*; use collection::shards::shard::ShardId; use common::counter::hardware_accumulator::HwMeasurementAcc; use schemars::JsonSchema; use segment::json_path::JsonPath; use segment::types::{Filter, PayloadFieldSchema, PayloadKeyType, StrictModeConfig}; use serde::{Deserialize, Serialize}; use shard::operations::payload_ops::*; use shard::operations::*; use storage::content_manager::collection_meta_ops::*; use storage::content_manager::collection_verification::check_strict_mode; use storage::content_manager::errors::StorageError; use storage::content_manager::toc::TableOfContent; use storage::dispatcher::Dispatcher; use storage::rbac::Access; use validator::Validate; use crate::common::inference::params::InferenceParams; use crate::common::inference::service::InferenceType; use crate::common::inference::update_requests::*; use crate::common::strict_mode::*; #[derive(Copy, Clone, Debug, Deserialize, Serialize, Validate)] pub struct UpdateParams { #[serde(default)] pub wait: bool, #[serde(default)] pub ordering: WriteOrdering, #[serde(default)] pub timeout: Option<Duration>, } impl UpdateParams { pub fn from_grpc( wait: Option<bool>, ordering: Option<api::grpc::qdrant::WriteOrdering>, timeout: Option<Duration>, ) -> tonic::Result<Self> { let params = Self { wait: wait.unwrap_or(false), ordering: write_ordering_from_proto(ordering)?, timeout, }; Ok(params) } pub(crate) fn timeout_as_secs(&self) -> Option<usize> { self.timeout.map(|timeout| timeout.as_secs() as usize) } } #[derive(Copy, Clone, Debug, Default)] pub struct InternalUpdateParams { pub shard_id: Option<ShardId>, pub clock_tag: Option<ClockTag>, } impl InternalUpdateParams { pub fn from_grpc( shard_id: Option<ShardId>, clock_tag: Option<api::grpc::qdrant::ClockTag>, ) -> Self { Self { shard_id, clock_tag: clock_tag.map(ClockTag::from), } } } #[derive(Deserialize, Serialize, JsonSchema, Validate)] pub struct UpdateOperations { #[validate(nested)] pub operations: Vec<UpdateOperation>, } #[derive(Deserialize, Serialize, JsonSchema)] #[serde(rename_all = "snake_case")] #[serde(untagged)] pub enum UpdateOperation { Upsert(UpsertOperation), Delete(DeleteOperation), SetPayload(SetPayloadOperation), OverwritePayload(OverwritePayloadOperation), DeletePayload(DeletePayloadOperation), ClearPayload(ClearPayloadOperation), UpdateVectors(UpdateVectorsOperation), DeleteVectors(DeleteVectorsOperation), } impl Validate for UpdateOperation { fn validate(&self) -> Result<(), validator::ValidationErrors> { match self { UpdateOperation::Upsert(op) => op.validate(), UpdateOperation::Delete(op) => op.validate(), UpdateOperation::SetPayload(op) => op.validate(), UpdateOperation::OverwritePayload(op) => op.validate(), UpdateOperation::DeletePayload(op) => op.validate(), UpdateOperation::ClearPayload(op) => op.validate(), UpdateOperation::UpdateVectors(op) => op.validate(), UpdateOperation::DeleteVectors(op) => op.validate(), } } } impl StrictModeVerification for UpdateOperation { fn query_limit(&self) -> Option<usize> { None } fn indexed_filter_read(&self) -> Option<&segment::types::Filter> { None } fn indexed_filter_write(&self) -> Option<&segment::types::Filter> { None } fn request_exact(&self) -> Option<bool> { None } fn request_search_params(&self) -> Option<&segment::types::SearchParams> { None } async fn check_strict_mode( &self, collection: &Collection, strict_mode_config: &StrictModeConfig, ) -> CollectionResult<()> { match self { UpdateOperation::Upsert(op) => { op.upsert .check_strict_mode(collection, strict_mode_config) .await } UpdateOperation::Delete(op) => { op.delete .check_strict_mode(collection, strict_mode_config) .await } UpdateOperation::SetPayload(op) => { op.set_payload .check_strict_mode(collection, strict_mode_config) .await } UpdateOperation::OverwritePayload(op) => { op.overwrite_payload .check_strict_mode(collection, strict_mode_config) .await } UpdateOperation::DeletePayload(op) => { op.delete_payload .check_strict_mode(collection, strict_mode_config) .await } UpdateOperation::ClearPayload(op) => { op.clear_payload .check_strict_mode(collection, strict_mode_config) .await } UpdateOperation::UpdateVectors(op) => { op.update_vectors .check_strict_mode(collection, strict_mode_config) .await } UpdateOperation::DeleteVectors(op) => { op.delete_vectors .check_strict_mode(collection, strict_mode_config) .await } } } } impl StrictModeVerification for CreateFieldIndex { async fn check_custom( &self, collection: &Collection, strict_mode_config: &StrictModeConfig, ) -> CollectionResult<()> { if let Some(max_payload_index_count) = strict_mode_config.max_payload_index_count { let collection_info = collection.info(&ShardSelectorInternal::All).await?; if collection_info.payload_schema.len() >= max_payload_index_count { return Err(CollectionError::strict_mode( format!( "Collection already has the maximum number of payload indices ({max_payload_index_count})" ), "Please delete an existing index before creating a new one.", )); } } Ok(()) } fn indexed_filter_write(&self) -> Option<&Filter> { None } fn query_limit(&self) -> Option<usize> { None } fn indexed_filter_read(&self) -> Option<&Filter> { None } fn request_exact(&self) -> Option<bool> { None } fn request_search_params(&self) -> Option<&segment::types::SearchParams> { None } } #[derive(Deserialize, Serialize, JsonSchema, Validate)] pub struct UpsertOperation { #[validate(nested)] upsert: PointInsertOperations, } #[derive(Deserialize, Serialize, JsonSchema, Validate)] pub struct DeleteOperation { #[validate(nested)] delete: PointsSelector, } #[derive(Deserialize, Serialize, JsonSchema, Validate)] pub struct SetPayloadOperation { #[validate(nested)] set_payload: SetPayload, } #[derive(Deserialize, Serialize, JsonSchema, Validate)] pub struct OverwritePayloadOperation { #[validate(nested)] overwrite_payload: SetPayload, } #[derive(Deserialize, Serialize, JsonSchema, Validate)] pub struct DeletePayloadOperation { #[validate(nested)] delete_payload: DeletePayload, } #[derive(Deserialize, Serialize, JsonSchema, Validate)] pub struct ClearPayloadOperation { #[validate(nested)] clear_payload: PointsSelector, } #[derive(Deserialize, Serialize, JsonSchema, Validate)] pub struct UpdateVectorsOperation { #[validate(nested)] update_vectors: UpdateVectors, } #[derive(Deserialize, Serialize, JsonSchema, Validate)] pub struct DeleteVectorsOperation { #[validate(nested)] delete_vectors: DeleteVectors, } #[derive(Debug, Deserialize, Serialize, JsonSchema, Validate)] pub struct CreateFieldIndex { pub field_name: PayloadKeyType, #[serde(alias = "field_type")] #[validate(nested)] pub field_schema: Option<PayloadFieldSchema>, } #[expect(clippy::too_many_arguments)] pub async fn do_upsert_points( toc_provider: impl CheckedTocProvider, collection_name: String, operation: PointInsertOperations, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, inference_params: InferenceParams, hw_measurement_acc: HwMeasurementAcc, ) -> Result<(UpdateResult, Option<models::InferenceUsage>), StorageError> { let toc = toc_provider .check_strict_mode( &operation, &collection_name, params.timeout_as_secs(), &access, ) .await?; let (operation, shard_key, usage, update_filter) = match operation { PointInsertOperations::PointsBatch(batch) => { let PointsBatch { batch, shard_key, update_filter, } = batch; let (batch, usage) = convert_batch(batch, inference_params).await?; let operation = PointInsertOperationsInternal::PointsBatch(batch); (operation, shard_key, usage, update_filter) } PointInsertOperations::PointsList(list) => { let PointsList { points, shard_key, update_filter, } = list; let (list, usage) = convert_point_struct(points, InferenceType::Update, inference_params).await?; let operation = PointInsertOperationsInternal::PointsList(list); (operation, shard_key, usage, update_filter) } }; let operation = if let Some(condition) = update_filter { CollectionUpdateOperations::PointOperation(PointOperations::UpsertPointsConditional( ConditionalInsertOperationInternal { points_op: operation, condition, }, )) } else { CollectionUpdateOperations::PointOperation(PointOperations::UpsertPoints(operation)) }; let result = update( toc, &collection_name, operation, internal_params, params, shard_key, access, hw_measurement_acc, ) .await?; Ok((result, usage)) } pub async fn do_delete_points( toc_provider: impl CheckedTocProvider, collection_name: String, points: PointsSelector, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { let toc = toc_provider .check_strict_mode(&points, &collection_name, params.timeout_as_secs(), &access) .await?; let (operation, shard_key) = match points { PointsSelector::PointIdsSelector(PointIdsList { points, shard_key }) => { (PointOperations::DeletePoints { ids: points }, shard_key) } PointsSelector::FilterSelector(FilterSelector { filter, shard_key }) => { (PointOperations::DeletePointsByFilter(filter), shard_key) } }; let operation = CollectionUpdateOperations::PointOperation(operation); update( toc, &collection_name, operation, internal_params, params, shard_key, access, hw_measurement_acc, ) .await } #[expect(clippy::too_many_arguments)] pub async fn do_update_vectors( toc_provider: impl CheckedTocProvider, collection_name: String, operation: UpdateVectors, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, inference_params: InferenceParams, hw_measurement_acc: HwMeasurementAcc, ) -> Result<(UpdateResult, Option<models::InferenceUsage>), StorageError> { let toc = toc_provider .check_strict_mode( &operation, &collection_name, params.timeout_as_secs(), &access, ) .await?; let UpdateVectors { points, shard_key, update_filter, } = operation; let (points, usage) = convert_point_vectors(points, InferenceType::Update, inference_params).await?; let operation = CollectionUpdateOperations::VectorOperation(VectorOperations::UpdateVectors( UpdateVectorsOp { points, update_filter, }, )); let result = update( toc, &collection_name, operation, internal_params, params, shard_key, access, hw_measurement_acc, ) .await?; Ok((result, usage)) } pub async fn do_delete_vectors( toc_provider: impl CheckedTocProvider, collection_name: String, operation: DeleteVectors, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { // TODO: Is this cancel safe!? let toc = toc_provider .check_strict_mode( &operation, &collection_name, params.timeout_as_secs(), &access, ) .await?; let DeleteVectors { vector, filter, points, shard_key, } = operation; let vector_names: Vec<_> = vector.into_iter().collect(); let mut result = None; if let Some(filter) = filter { let vectors_operation = VectorOperations::DeleteVectorsByFilter(filter, vector_names.clone()); let operation = CollectionUpdateOperations::VectorOperation(vectors_operation); result = Some( update( toc, &collection_name, operation, internal_params, params, shard_key.clone(), access.clone(), hw_measurement_acc.clone(), ) .await?, ); } if let Some(points) = points { let vectors_operation = VectorOperations::DeleteVectors(points.into(), vector_names); let operation = CollectionUpdateOperations::VectorOperation(vectors_operation); result = Some( update( toc, &collection_name, operation, internal_params, params, shard_key, access, hw_measurement_acc, ) .await?, ); } result.ok_or_else(|| StorageError::bad_request("No filter or points provided")) } pub async fn do_set_payload( toc_provider: impl CheckedTocProvider, collection_name: String, operation: SetPayload, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { let toc = toc_provider .check_strict_mode( &operation, &collection_name, params.timeout_as_secs(), &access, ) .await?; let SetPayload { points, payload, filter, shard_key, key, } = operation; let operation = CollectionUpdateOperations::PayloadOperation(PayloadOps::SetPayload(SetPayloadOp { payload, points, filter, key, })); update( toc, &collection_name, operation, internal_params, params, shard_key, access, hw_measurement_acc, ) .await } pub async fn do_overwrite_payload( toc_provider: impl CheckedTocProvider, collection_name: String, operation: SetPayload, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { let toc = toc_provider .check_strict_mode( &operation, &collection_name, params.timeout_as_secs(), &access, ) .await?; let SetPayload { points, payload, filter, shard_key, key: _, } = operation; let operation = CollectionUpdateOperations::PayloadOperation(PayloadOps::OverwritePayload(SetPayloadOp { payload, points, filter, // overwrite operation doesn't support payload selector key: None, })); update( toc, &collection_name, operation, internal_params, params, shard_key, access, hw_measurement_acc, ) .await } pub async fn do_delete_payload( toc_provider: impl CheckedTocProvider, collection_name: String, operation: DeletePayload, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { let toc = toc_provider .check_strict_mode( &operation, &collection_name, params.timeout_as_secs(), &access, ) .await?; let DeletePayload { keys, points, filter, shard_key, } = operation; let operation = CollectionUpdateOperations::PayloadOperation(PayloadOps::DeletePayload(DeletePayloadOp { keys, points, filter, })); update( toc, &collection_name, operation, internal_params, params, shard_key, access, hw_measurement_acc, ) .await } pub async fn do_clear_payload( toc_provider: impl CheckedTocProvider, collection_name: String, points: PointsSelector, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { let toc = toc_provider .check_strict_mode(&points, &collection_name, params.timeout_as_secs(), &access) .await?; let (point_operation, shard_key) = match points { PointsSelector::PointIdsSelector(PointIdsList { points, shard_key }) => { (PayloadOps::ClearPayload { points }, shard_key) } PointsSelector::FilterSelector(FilterSelector { filter, shard_key }) => { (PayloadOps::ClearPayloadByFilter(filter), shard_key) } }; let operation = CollectionUpdateOperations::PayloadOperation(point_operation); update( toc, &collection_name, operation, internal_params, params, shard_key, access, hw_measurement_acc, ) .await } #[expect(clippy::too_many_arguments)] pub async fn do_batch_update_points( toc_provider: impl CheckedTocProvider + Clone, collection_name: String, operations: Vec<UpdateOperation>, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, inference_params: InferenceParams, hw_measurement_acc: HwMeasurementAcc, ) -> Result<(Vec<UpdateResult>, Option<InferenceUsage>), StorageError> { // Check strict mode for all batch operations, *before applying* them let mut toc = None; for operation in &operations { toc = toc_provider .check_strict_mode( operation, &collection_name, params.timeout_as_secs(), &access, ) .await? .into(); } let Some(toc) = toc else { // Batch is empty, return empty result vector return Ok((Vec::new(), None)); }; // Pass unchecked ToC provider into `do_*` functions, because we already checked strict mode let toc_provider = UncheckedTocProvider::new_unchecked(toc); let mut results = Vec::with_capacity(operations.len()); let mut inference_usage = InferenceUsage::default(); for operation in operations { let current_update_result = match operation { UpdateOperation::Upsert(operation) => { let (result, usage) = do_upsert_points( toc_provider.clone(), collection_name.clone(), operation.upsert, internal_params, params, access.clone(), inference_params.clone(), hw_measurement_acc.clone(), ) .await?; inference_usage.merge_opt(usage); result } UpdateOperation::Delete(operation) => { do_delete_points( toc_provider.clone(), collection_name.clone(), operation.delete, internal_params, params, access.clone(), hw_measurement_acc.clone(), ) .await? } UpdateOperation::SetPayload(operation) => { do_set_payload( toc_provider.clone(), collection_name.clone(), operation.set_payload, internal_params, params, access.clone(), hw_measurement_acc.clone(), ) .await? } UpdateOperation::OverwritePayload(operation) => { do_overwrite_payload( toc_provider.clone(), collection_name.clone(), operation.overwrite_payload, internal_params, params, access.clone(), hw_measurement_acc.clone(), ) .await? } UpdateOperation::DeletePayload(operation) => { do_delete_payload( toc_provider.clone(), collection_name.clone(), operation.delete_payload, internal_params, params, access.clone(), hw_measurement_acc.clone(), ) .await? } UpdateOperation::ClearPayload(operation) => { do_clear_payload( toc_provider.clone(), collection_name.clone(), operation.clear_payload, internal_params, params, access.clone(), hw_measurement_acc.clone(), ) .await? } UpdateOperation::UpdateVectors(operation) => { let (result, usage) = do_update_vectors( toc_provider.clone(), collection_name.clone(), operation.update_vectors, internal_params, params, access.clone(), inference_params.clone(), hw_measurement_acc.clone(), ) .await?; inference_usage.merge_opt(usage); result } UpdateOperation::DeleteVectors(operation) => { do_delete_vectors( toc_provider.clone(), collection_name.clone(), operation.delete_vectors, internal_params, params, access.clone(), hw_measurement_acc.clone(), ) .await? } }; results.push(current_update_result); } Ok((results, inference_usage.into_non_empty())) } pub async fn do_create_index( dispatcher: Arc<Dispatcher>, collection_name: String, operation: CreateFieldIndex, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { // TODO: Is this cancel safe!? // Use per-request timeout from params if provided let wait_timeout = params.timeout; // Check strict mode before submitting consensus operation let pass = check_strict_mode( &operation, wait_timeout.map(|d| d.as_secs() as usize), &collection_name, &dispatcher, &access, ) .await?; let Some(field_schema) = operation.field_schema else { return Err(StorageError::bad_request( "Can't auto-detect field type, please specify `field_schema` in the request", )); }; let consensus_op = CollectionMetaOperations::CreatePayloadIndex(CreatePayloadIndex { collection_name: collection_name.clone(), field_name: operation.field_name.clone(), field_schema: field_schema.clone(), }); let toc = dispatcher.toc(&access, &pass).clone(); // TODO: Is `submit_collection_meta_op` cancel-safe!? Should be, I think?.. 🤔 dispatcher .submit_collection_meta_op(consensus_op, access, wait_timeout) .await?; // This function is required as long as we want to maintain interface compatibility // for `wait` parameter and return type. // The idea is to migrate from the point-like interface to consensus-like interface in the next few versions do_create_index_internal( toc, collection_name, operation.field_name, Some(field_schema), internal_params, params, hw_measurement_acc, ) .await } pub async fn do_create_index_internal( toc: Arc<TableOfContent>, collection_name: String, field_name: PayloadKeyType, field_schema: Option<PayloadFieldSchema>, internal_params: InternalUpdateParams, params: UpdateParams, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { let operation = CollectionUpdateOperations::FieldIndexOperation( FieldIndexOperations::CreateIndex(CreateIndex { field_name, field_schema, }), ); update( &toc, &collection_name, operation, internal_params, params, None, Access::full("Internal API"), hw_measurement_acc, ) .await } pub async fn do_delete_index( dispatcher: Arc<Dispatcher>, collection_name: String, index_name: JsonPath, internal_params: InternalUpdateParams, params: UpdateParams, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { // TODO: Is this cancel safe!? let consensus_op = CollectionMetaOperations::DropPayloadIndex(DropPayloadIndex { collection_name: collection_name.clone(), field_name: index_name.clone(), }); // Use per-request timeout from params if provided let wait_timeout = params.timeout; // Nothing to verify here. let pass = new_unchecked_verification_pass(); let toc = dispatcher.toc(&access, &pass).clone(); // TODO: Is `submit_collection_meta_op` cancel-safe!? Should be, I think?.. 🤔 dispatcher .submit_collection_meta_op(consensus_op, access, wait_timeout) .await?; do_delete_index_internal( toc, collection_name, index_name, internal_params, params, hw_measurement_acc, ) .await } pub async fn do_delete_index_internal( toc: Arc<TableOfContent>, collection_name: String, index_name: JsonPath, internal_params: InternalUpdateParams, params: UpdateParams, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { let operation = CollectionUpdateOperations::FieldIndexOperation( FieldIndexOperations::DeleteIndex(index_name), ); update( &toc, &collection_name, operation, internal_params, params, None, Access::full("Internal API"), hw_measurement_acc, ) .await } #[expect(clippy::too_many_arguments)] pub async fn update( toc: &TableOfContent, collection_name: &str, operation: CollectionUpdateOperations, internal_params: InternalUpdateParams, params: UpdateParams, shard_key: Option<ShardKeySelector>, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<UpdateResult, StorageError> { let InternalUpdateParams { shard_id, clock_tag, } = internal_params; let UpdateParams { wait, ordering, timeout: _, } = params; let shard_selector = match operation { CollectionUpdateOperations::PointOperation(point_ops::PointOperations::SyncPoints(_)) => { debug_assert_eq!( shard_key, None, "Sync points operations can't specify shard key" ); match shard_id { Some(shard_id) => ShardSelectorInternal::ShardId(shard_id), None => { debug_assert!(false, "Sync operation is supposed to select shard directly"); ShardSelectorInternal::Empty } } } CollectionUpdateOperations::FieldIndexOperation(_) => { debug_assert_eq!( shard_key, None, "Field index operations can't specify shard key" ); match shard_id { Some(shard_id) => ShardSelectorInternal::ShardId(shard_id), None => ShardSelectorInternal::All, } } _ => get_shard_selector_for_update(shard_id, shard_key), }; toc.update( collection_name, OperationWithClockTag::new(operation, clock_tag), wait, ordering, shard_selector, access, hw_measurement_acc, ) .await } /// Converts a pair of parameters into a shard selector /// suitable for update operations. /// /// The key difference from selector for search operations is that /// empty shard selector in case of update means default shard, /// while empty shard selector in case of search means all shards. /// /// Parameters: /// - shard_selection: selection of the exact shard ID, always have priority over shard_key /// - shard_key: selection of the shard key, can be a single key or a list of keys /// /// Returns: /// - ShardSelectorInternal - resolved shard selector fn get_shard_selector_for_update( shard_selection: Option<ShardId>, shard_key: Option<ShardKeySelector>, ) -> ShardSelectorInternal { match (shard_selection, shard_key) { (Some(shard_selection), None) => ShardSelectorInternal::ShardId(shard_selection), (Some(shard_selection), Some(_)) => { debug_assert!( false, "Shard selection and shard key are mutually exclusive" ); ShardSelectorInternal::ShardId(shard_selection) } (None, Some(shard_key)) => ShardSelectorInternal::from(shard_key), (None, None) => ShardSelectorInternal::Empty, } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/telemetry.rs
src/common/telemetry.rs
use std::sync::Arc; use std::time::Duration; use collection::operations::verification::new_unchecked_verification_pass; use common::types::{DetailsLevel, TelemetryDetail}; use parking_lot::Mutex; use schemars::JsonSchema; use segment::common::anonymize::Anonymize; use serde::Serialize; use shard::common::stopping_guard::StoppingGuard; use storage::content_manager::errors::{StorageError, StorageResult}; use storage::dispatcher::Dispatcher; use storage::rbac::Access; use tokio::time::error::Elapsed; use tokio_util::task::AbortOnDropHandle; use uuid::Uuid; use crate::common::telemetry_ops::app_telemetry::{AppBuildTelemetry, AppBuildTelemetryCollector}; use crate::common::telemetry_ops::cluster_telemetry::ClusterTelemetry; use crate::common::telemetry_ops::collections_telemetry::CollectionsTelemetry; use crate::common::telemetry_ops::hardware::HardwareTelemetry; use crate::common::telemetry_ops::memory_telemetry::MemoryTelemetry; use crate::common::telemetry_ops::requests_telemetry::{ ActixTelemetryCollector, RequestsTelemetry, TonicTelemetryCollector, }; use crate::settings::Settings; // Keep in sync with openapi/openapi-service.ytt.yaml const DEFAULT_TELEMETRY_TIMEOUT: Duration = Duration::from_secs(60); pub struct TelemetryCollector { process_id: Uuid, settings: Settings, dispatcher: Arc<Dispatcher>, pub app_telemetry_collector: AppBuildTelemetryCollector, pub actix_telemetry_collector: Arc<Mutex<ActixTelemetryCollector>>, pub tonic_telemetry_collector: Arc<Mutex<TonicTelemetryCollector>>, } // Whole telemetry data #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] pub struct TelemetryData { #[anonymize(false)] id: String, pub(crate) app: AppBuildTelemetry, pub(crate) collections: CollectionsTelemetry, #[serde(skip_serializing_if = "Option::is_none")] pub(crate) cluster: Option<ClusterTelemetry>, #[serde(skip_serializing_if = "Option::is_none")] pub(crate) requests: Option<RequestsTelemetry>, #[serde(skip_serializing_if = "Option::is_none")] pub(crate) memory: Option<MemoryTelemetry>, #[serde(skip_serializing_if = "Option::is_none")] pub(crate) hardware: Option<HardwareTelemetry>, } impl TelemetryCollector { pub fn reporting_id(&self) -> String { self.process_id.to_string() } pub fn generate_id() -> Uuid { Uuid::new_v4() } pub fn new(settings: Settings, dispatcher: Arc<Dispatcher>, id: Uuid) -> Self { Self { process_id: id, settings, dispatcher, app_telemetry_collector: AppBuildTelemetryCollector::new(), actix_telemetry_collector: Arc::new(Mutex::new(ActixTelemetryCollector { workers: Vec::new(), })), tonic_telemetry_collector: Arc::new(Mutex::new(TonicTelemetryCollector { workers: Vec::new(), })), } } pub async fn prepare_data( &self, access: &Access, detail: TelemetryDetail, timeout: Option<Duration>, ) -> StorageResult<TelemetryData> { let timeout = timeout.unwrap_or(DEFAULT_TELEMETRY_TIMEOUT); // Use blocking pool because the collection telemetry acquires several sync. locks. let is_stopped_guard = StoppingGuard::new(); let is_stopped = is_stopped_guard.get_is_stopped(); let collections_telemetry_handle = { let toc = self .dispatcher .toc(access, &new_unchecked_verification_pass()) .clone(); let runtime_handle = toc.general_runtime_handle().clone(); let access_collection = access.clone(); let handle = runtime_handle.spawn_blocking(move || { // Re-enter the async runtime in this blocking thread tokio::runtime::Handle::current().block_on(async move { CollectionsTelemetry::collect( detail, &access_collection, &toc, timeout, &is_stopped, ) .await }) }); AbortOnDropHandle::new(handle) }; let collections_telemetry = tokio::time::timeout(timeout, collections_telemetry_handle) .await .map_err(|_: Elapsed| StorageError::timeout(timeout, "collections telemetry"))???; Ok(TelemetryData { id: self.process_id.to_string(), collections: collections_telemetry, app: AppBuildTelemetry::collect(detail, &self.app_telemetry_collector, &self.settings), cluster: ClusterTelemetry::collect(access, detail, &self.dispatcher, &self.settings), requests: RequestsTelemetry::collect( access, &self.actix_telemetry_collector.lock(), &self.tonic_telemetry_collector.lock(), detail, ), memory: (detail.level > DetailsLevel::Level0) .then(|| MemoryTelemetry::collect(access)) .flatten(), hardware: (detail.level > DetailsLevel::Level0) .then(|| HardwareTelemetry::new(&self.dispatcher, access)), }) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/helpers.rs
src/common/helpers.rs
use std::cmp::max; use std::io; use std::sync::atomic::{AtomicUsize, Ordering}; use fs_err as fs; use tokio::runtime; use tokio::runtime::Runtime; use tonic::transport::{Certificate, ClientTlsConfig, Identity, ServerTlsConfig}; use crate::settings::{Settings, TlsConfig}; pub fn create_search_runtime(max_search_threads: usize) -> io::Result<Runtime> { let num_threads = common::defaults::search_thread_count(max_search_threads); runtime::Builder::new_multi_thread() .worker_threads(num_threads) .max_blocking_threads(num_threads) .enable_all() .thread_name_fn(|| { static ATOMIC_ID: AtomicUsize = AtomicUsize::new(0); let id = ATOMIC_ID.fetch_add(1, Ordering::SeqCst); format!("search-{id}") }) .build() } pub fn create_update_runtime(max_optimization_threads: usize) -> io::Result<Runtime> { let mut update_runtime_builder = runtime::Builder::new_multi_thread(); let num_cpus = common::cpu::get_num_cpus(); update_runtime_builder .enable_time() .enable_io() .worker_threads(num_cpus) .thread_name_fn(move || { static ATOMIC_ID: AtomicUsize = AtomicUsize::new(0); let update_id = ATOMIC_ID.fetch_add(1, Ordering::SeqCst); format!("update-{update_id}") }); if max_optimization_threads > 0 { // panics if val is not larger than 0. update_runtime_builder.max_blocking_threads(max_optimization_threads); } update_runtime_builder.build() } pub fn create_general_purpose_runtime() -> io::Result<Runtime> { runtime::Builder::new_multi_thread() .enable_time() .enable_io() .worker_threads(max(common::cpu::get_num_cpus(), 2)) .thread_name_fn(|| { static ATOMIC_ID: AtomicUsize = AtomicUsize::new(0); let general_id = ATOMIC_ID.fetch_add(1, Ordering::SeqCst); format!("general-{general_id}") }) .build() } /// Load client TLS configuration. pub fn load_tls_client_config(settings: &Settings) -> io::Result<Option<ClientTlsConfig>> { if settings.cluster.p2p.enable_tls { let tls_config = &settings.tls()?; Ok(Some( ClientTlsConfig::new() .identity(load_identity(tls_config)?) .ca_certificate(load_ca_certificate(tls_config)?), )) } else { Ok(None) } } /// Load server TLS configuration for external gRPC pub fn load_tls_external_server_config(tls_config: &TlsConfig) -> io::Result<ServerTlsConfig> { Ok(ServerTlsConfig::new().identity(load_identity(tls_config)?)) } /// Load server TLS configuration for internal gRPC, check client certificate against CA pub fn load_tls_internal_server_config(tls_config: &TlsConfig) -> io::Result<ServerTlsConfig> { Ok(ServerTlsConfig::new() .identity(load_identity(tls_config)?) .client_ca_root(load_ca_certificate(tls_config)?)) } fn load_identity(tls_config: &TlsConfig) -> io::Result<Identity> { let cert = fs::read_to_string(&tls_config.cert)?; let key = fs::read_to_string(&tls_config.key)?; Ok(Identity::from_pem(cert, key)) } fn load_ca_certificate(tls_config: &TlsConfig) -> io::Result<Certificate> { let Some(ca_cert_path) = &tls_config.ca_cert else { return Err(io::Error::new( io::ErrorKind::InvalidInput, "CA certificate is required for TLS configuration", )); }; let pem = fs::read_to_string(ca_cert_path)?; Ok(Certificate::from_pem(pem)) } pub fn tonic_error_to_io_error(err: tonic::transport::Error) -> io::Error { io::Error::other(err) } #[cfg(test)] mod tests { use std::sync::Arc; use std::thread; use std::thread::sleep; use std::time::Duration; use collection::common::is_ready::IsReady; #[test] fn test_is_ready() { let is_ready = Arc::new(IsReady::default()); let is_ready_clone = is_ready.clone(); let join = thread::spawn(move || { is_ready_clone.await_ready(); eprintln!( "is_ready_clone.check_ready() = {:#?}", is_ready_clone.check_ready() ); }); sleep(Duration::from_millis(500)); eprintln!("Making ready"); is_ready.make_ready(); sleep(Duration::from_millis(500)); join.join().unwrap() } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/collections.rs
src/common/collections.rs
use std::collections::HashMap; use std::sync::Arc; use std::time::Duration; use api::grpc::qdrant::CollectionExists; use api::rest::models::{ CollectionDescription, CollectionsResponse, ShardKeyDescription, ShardKeysResponse, }; use collection::config::ShardingMethod; #[cfg(feature = "staging")] use collection::operations::cluster_ops::TestSlowDownOperation; use collection::operations::cluster_ops::{ AbortTransferOperation, ClusterOperations, DropReplicaOperation, MoveShardOperation, ReplicatePoints, ReplicatePointsOperation, ReplicateShardOperation, ReshardingDirection, RestartTransfer, RestartTransferOperation, StartResharding, }; use collection::operations::shard_selector_internal::ShardSelectorInternal; use collection::operations::snapshot_ops::SnapshotDescription; use collection::operations::types::{ AliasDescription, CollectionClusterInfo, CollectionInfo, CollectionsAliasesResponse, }; use collection::operations::verification::new_unchecked_verification_pass; use collection::shards::replica_set; use collection::shards::replica_set::replica_set_state; use collection::shards::resharding::ReshardKey; use collection::shards::shard::{PeerId, ShardId, ShardsPlacement}; use collection::shards::transfer::{ ShardTransfer, ShardTransferKey, ShardTransferMethod, ShardTransferRestart, }; use itertools::Itertools; use rand::prelude::SliceRandom; use rand::seq::IteratorRandom; use storage::content_manager::collection_meta_ops::ShardTransferOperations::{Abort, Start}; #[cfg(feature = "staging")] use storage::content_manager::collection_meta_ops::TestSlowDown; use storage::content_manager::collection_meta_ops::{ CollectionMetaOperations, CreateShardKey, DropShardKey, ReshardingOperation, SetShardReplicaState, ShardTransferOperations, UpdateCollectionOperation, }; use storage::content_manager::errors::StorageError; use storage::content_manager::toc::TableOfContent; use storage::dispatcher::Dispatcher; use storage::rbac::{Access, AccessRequirements}; use uuid::Uuid; pub async fn do_collection_exists( toc: &TableOfContent, access: Access, name: &str, ) -> Result<CollectionExists, StorageError> { let collection_pass = access.check_collection_access(name, AccessRequirements::new())?; // if this returns Ok, it means the collection exists. // if not, we check that the error is NotFound let Err(error) = toc.get_collection(&collection_pass).await else { return Ok(CollectionExists { exists: true }); }; match error { StorageError::NotFound { .. } => Ok(CollectionExists { exists: false }), e => Err(e), } } pub async fn do_get_collection( toc: &TableOfContent, access: Access, name: &str, shard_selection: Option<ShardId>, ) -> Result<CollectionInfo, StorageError> { let collection_pass = access.check_collection_access(name, AccessRequirements::new())?; let collection = toc.get_collection(&collection_pass).await?; let shard_selection = match shard_selection { None => ShardSelectorInternal::All, Some(shard_id) => ShardSelectorInternal::ShardId(shard_id), }; Ok(collection.info(&shard_selection).await?) } pub async fn do_list_collections( toc: &TableOfContent, access: Access, ) -> Result<CollectionsResponse, StorageError> { let collections = toc .all_collections(&access) .await .into_iter() .map(|pass| CollectionDescription { name: pass.name().to_string(), }) .collect_vec(); Ok(CollectionsResponse { collections }) } pub async fn do_get_collection_shard_keys( toc: &TableOfContent, access: Access, name: &str, ) -> Result<ShardKeysResponse, StorageError> { let collection_pass = access.check_collection_access(name, AccessRequirements::new())?; let collection = toc.get_collection(&collection_pass).await?; let state = collection.state().await; let shard_keys = match state.config.params.sharding_method.unwrap_or_default() { ShardingMethod::Auto => None, ShardingMethod::Custom => Some( state .shards_key_mapping .iter_shard_keys() .map(|k| ShardKeyDescription { key: k.clone() }) .collect(), ), }; Ok(ShardKeysResponse { shard_keys }) } /// Construct shards-replicas layout for the shard from the given scope of peers /// Example: /// Shards: 3 /// Replicas: 2 /// Peers: [A, B, C] /// /// Placement: /// [ /// [A, B] /// [B, C] /// [A, C] /// ] fn generate_even_placement( mut pool: Vec<PeerId>, shard_number: usize, replication_factor: usize, ) -> ShardsPlacement { let mut exact_placement = Vec::new(); let mut rng = rand::rng(); pool.shuffle(&mut rng); let mut loop_iter = pool.iter().cycle(); // pool: [1,2,3,4] // shuf_pool: [2,3,4,1] // // loop_iter: [2, 3, 4, 1, 2, 3, 4, 1, 2, 3, 4, 1, 2, 3, 4, 1,...] // shard_placement: [2, 3, 4][1, 2, 3][4, 1, 2][3, 4, 1][2, 3, 4] let max_replication_factor = std::cmp::min(replication_factor, pool.len()); for _shard in 0..shard_number { let mut shard_placement = Vec::new(); for _replica in 0..max_replication_factor { shard_placement.push(*loop_iter.next().unwrap()); } exact_placement.push(shard_placement); } exact_placement } pub async fn do_list_collection_aliases( toc: &TableOfContent, access: Access, collection_name: &str, ) -> Result<CollectionsAliasesResponse, StorageError> { let collection_pass = access.check_collection_access(collection_name, AccessRequirements::new())?; let aliases: Vec<AliasDescription> = toc .collection_aliases(&collection_pass, &access) .await? .into_iter() .map(|alias| AliasDescription { alias_name: alias, collection_name: collection_name.to_string(), }) .collect(); Ok(CollectionsAliasesResponse { aliases }) } pub async fn do_list_aliases( toc: &TableOfContent, access: Access, ) -> Result<CollectionsAliasesResponse, StorageError> { let aliases = toc.list_aliases(&access).await?; Ok(CollectionsAliasesResponse { aliases }) } pub async fn do_list_snapshots( toc: &TableOfContent, access: Access, collection_name: &str, ) -> Result<Vec<SnapshotDescription>, StorageError> { let collection_pass = access.check_collection_access(collection_name, AccessRequirements::new().extras())?; Ok(toc .get_collection(&collection_pass) .await? .list_snapshots() .await?) } pub async fn do_create_snapshot( toc: Arc<TableOfContent>, access: Access, collection_name: &str, ) -> Result<SnapshotDescription, StorageError> { let collection_pass = access .check_collection_access(collection_name, AccessRequirements::new().write().extras())? .into_static(); let result = tokio::spawn(async move { toc.create_snapshot(&collection_pass).await }).await??; Ok(result) } pub async fn do_get_collection_cluster( toc: &TableOfContent, access: Access, name: &str, ) -> Result<CollectionClusterInfo, StorageError> { let collection_pass = access.check_collection_access(name, AccessRequirements::new().extras())?; let collection = toc.get_collection(&collection_pass).await?; Ok(collection.cluster_info(toc.this_peer_id).await?) } pub async fn do_update_collection_cluster( dispatcher: &Dispatcher, collection_name: String, operation: ClusterOperations, access: Access, wait_timeout: Option<Duration>, ) -> Result<bool, StorageError> { let collection_pass = access.check_collection_access( &collection_name, AccessRequirements::new().write().manage().extras(), )?; if dispatcher.consensus_state().is_none() { return Err(StorageError::BadRequest { description: "Distributed mode disabled".to_string(), }); } let consensus_state = dispatcher.consensus_state().unwrap(); let get_all_peer_ids = || { consensus_state .persistent .read() .peer_address_by_id .read() .keys() .cloned() .collect_vec() }; let validate_peer_exists = |peer_id| { let target_peer_exist = consensus_state .persistent .read() .peer_address_by_id .read() .contains_key(&peer_id); if !target_peer_exist { return Err(StorageError::BadRequest { description: format!("Peer {peer_id} does not exist"), }); } Ok(()) }; // All checks should've been done at this point. let pass = new_unchecked_verification_pass(); let collection = dispatcher .toc(&access, &pass) .get_collection(&collection_pass) .await?; match operation { ClusterOperations::MoveShard(MoveShardOperation { move_shard }) => { // validate shard to move if !collection.contains_shard(move_shard.shard_id).await { return Err(StorageError::BadRequest { description: format!( "Shard {} of {} does not exist", move_shard.shard_id, collection_name ), }); }; // validate target and source peer exists validate_peer_exists(move_shard.to_peer_id)?; validate_peer_exists(move_shard.from_peer_id)?; // submit operation to consensus dispatcher .submit_collection_meta_op( CollectionMetaOperations::TransferShard( collection_name, Start(ShardTransfer { shard_id: move_shard.shard_id, to_shard_id: move_shard.to_shard_id, to: move_shard.to_peer_id, from: move_shard.from_peer_id, sync: false, method: move_shard.method, filter: None, }), ), access, wait_timeout, ) .await } ClusterOperations::ReplicateShard(ReplicateShardOperation { replicate_shard }) => { // validate shard to move if !collection.contains_shard(replicate_shard.shard_id).await { return Err(StorageError::BadRequest { description: format!( "Shard {} of {} does not exist", replicate_shard.shard_id, collection_name ), }); }; // validate target peer exists validate_peer_exists(replicate_shard.to_peer_id)?; // validate source peer exists validate_peer_exists(replicate_shard.from_peer_id)?; // submit operation to consensus dispatcher .submit_collection_meta_op( CollectionMetaOperations::TransferShard( collection_name, Start(ShardTransfer { shard_id: replicate_shard.shard_id, to_shard_id: replicate_shard.to_shard_id, to: replicate_shard.to_peer_id, from: replicate_shard.from_peer_id, sync: true, method: replicate_shard.method, filter: None, }), ), access, wait_timeout, ) .await } ClusterOperations::ReplicatePoints(ReplicatePointsOperation { replicate_points }) => { let ReplicatePoints { filter, from_shard_key, to_shard_key, } = replicate_points; let from_shard_ids = collection.get_shard_ids(&from_shard_key).await?; // Temporary, before we support multi-source transfers if from_shard_ids.len() != 1 { return Err(StorageError::BadRequest { description: format!( "Only replicating from shard keys with exactly one shard is supported. Shard key {from_shard_key} has {} shards", from_shard_ids.len() ), }); } // validate shard key exists let from_replicas = collection.get_replicas(&from_shard_key).await?; let to_replicas = collection.get_replicas(&to_shard_key).await?; debug_assert!(!from_replicas.is_empty()); if to_replicas.len() != 1 { return Err(StorageError::BadRequest { description: format!( "Only replicating to shard keys with exactly one replica is supported. Shard key {to_shard_key} has {} replicas", to_replicas.len() ), }); } let (from_shard_id, from_peer_id) = from_replicas[0]; let (to_shard_id, to_peer_id) = to_replicas[0]; // validate source & target peers exist validate_peer_exists(to_peer_id)?; validate_peer_exists(from_peer_id)?; // submit operation to consensus dispatcher .submit_collection_meta_op( CollectionMetaOperations::TransferShard( collection_name, Start(ShardTransfer { shard_id: from_shard_id, to_shard_id: Some(to_shard_id), from: from_peer_id, to: to_peer_id, sync: true, method: Some(ShardTransferMethod::StreamRecords), filter, }), ), access, wait_timeout, ) .await } ClusterOperations::AbortTransfer(AbortTransferOperation { abort_transfer }) => { let transfer = ShardTransferKey { shard_id: abort_transfer.shard_id, to_shard_id: abort_transfer.to_shard_id, to: abort_transfer.to_peer_id, from: abort_transfer.from_peer_id, }; if !collection.check_transfer_exists(&transfer).await { return Err(StorageError::NotFound { description: format!( "Shard transfer {} -> {} for collection {}:{} does not exist", transfer.from, transfer.to, collection_name, transfer.shard_id ), }); } dispatcher .submit_collection_meta_op( CollectionMetaOperations::TransferShard( collection_name, Abort { transfer, reason: "user request".to_string(), }, ), access, wait_timeout, ) .await } ClusterOperations::DropReplica(DropReplicaOperation { drop_replica }) => { if !collection.contains_shard(drop_replica.shard_id).await { return Err(StorageError::BadRequest { description: format!( "Shard {} of {} does not exist", drop_replica.shard_id, collection_name ), }); }; validate_peer_exists(drop_replica.peer_id)?; let mut update_operation = UpdateCollectionOperation::new_empty(collection_name); update_operation.set_shard_replica_changes(vec![replica_set::Change::Remove( drop_replica.shard_id, drop_replica.peer_id, )]); dispatcher .submit_collection_meta_op( CollectionMetaOperations::UpdateCollection(update_operation), access, wait_timeout, ) .await } ClusterOperations::CreateShardingKey(create_sharding_key_op) => { let create_sharding_key = create_sharding_key_op.create_sharding_key; // Validate that: // - proper sharding method is used // - key does not exist yet // // If placement suggested: // - Peers exist let state = collection.state().await; match state.config.params.sharding_method.unwrap_or_default() { ShardingMethod::Auto => { return Err(StorageError::bad_request( "Shard Key cannot be created with Auto sharding method", )); } ShardingMethod::Custom => {} } let shard_number = create_sharding_key .shards_number .unwrap_or(state.config.params.shard_number) .get() as usize; let replication_factor = create_sharding_key .replication_factor .unwrap_or(state.config.params.replication_factor) .get() as usize; if let Some(initial_state) = create_sharding_key.initial_state { match initial_state { replica_set_state::ReplicaState::Active | replica_set_state::ReplicaState::Partial => {} _ => { return Err(StorageError::bad_request(format!( "Initial state cannot be {initial_state:?}, only Active or Partial are allowed", ))); } } } let shard_keys_mapping = state.shards_key_mapping; if shard_keys_mapping.contains_key(&create_sharding_key.shard_key) { return Err(StorageError::BadRequest { description: format!( "Sharding key {} already exists for collection {}", create_sharding_key.shard_key, collection_name ), }); } let peers_pool: Vec<_> = if let Some(placement) = create_sharding_key.placement { if placement.is_empty() { return Err(StorageError::BadRequest { description: format!( "Sharding key {} placement cannot be empty. If you want to use random placement, do not specify placement", create_sharding_key.shard_key ), }); } for peer_id in placement.iter().copied() { validate_peer_exists(peer_id)?; } placement } else { get_all_peer_ids() }; let exact_placement = generate_even_placement(peers_pool, shard_number, replication_factor); dispatcher .submit_collection_meta_op( CollectionMetaOperations::CreateShardKey(CreateShardKey { collection_name, shard_key: create_sharding_key.shard_key, placement: exact_placement, initial_state: create_sharding_key.initial_state, }), access, wait_timeout, ) .await } ClusterOperations::DropShardingKey(drop_sharding_key_op) => { let drop_sharding_key = drop_sharding_key_op.drop_sharding_key; // Validate that: // - proper sharding method is used // - key does exist let state = collection.state().await; match state.config.params.sharding_method.unwrap_or_default() { ShardingMethod::Auto => { return Err(StorageError::bad_request( "Shard Key cannot be created with Auto sharding method", )); } ShardingMethod::Custom => {} } let shard_keys_mapping = state.shards_key_mapping; if !shard_keys_mapping.contains_key(&drop_sharding_key.shard_key) { return Err(StorageError::BadRequest { description: format!( "Sharding key {} does not exist for collection {collection_name}", drop_sharding_key.shard_key, ), }); } dispatcher .submit_collection_meta_op( CollectionMetaOperations::DropShardKey(DropShardKey { collection_name, shard_key: drop_sharding_key.shard_key, }), access, wait_timeout, ) .await } ClusterOperations::RestartTransfer(RestartTransferOperation { restart_transfer }) => { // TODO(reshading): Deduplicate resharding operations handling? let RestartTransfer { shard_id, to_shard_id, from_peer_id, to_peer_id, method, } = restart_transfer; let transfer_key = ShardTransferKey { shard_id, to_shard_id, to: to_peer_id, from: from_peer_id, }; if !collection.check_transfer_exists(&transfer_key).await { return Err(StorageError::NotFound { description: format!( "Shard transfer {} -> {} for collection {}:{} does not exist", transfer_key.from, transfer_key.to, collection_name, transfer_key.shard_id ), }); } dispatcher .submit_collection_meta_op( CollectionMetaOperations::TransferShard( collection_name, ShardTransferOperations::Restart(ShardTransferRestart { shard_id, to_shard_id, to: to_peer_id, from: from_peer_id, method, }), ), access, wait_timeout, ) .await } ClusterOperations::StartResharding(op) => { let StartResharding { uuid, direction, peer_id, shard_key, } = op.start_resharding; if !dispatcher.is_resharding_enabled() { return Err(StorageError::bad_request( "resharding is only supported in Qdrant Cloud", )); } // Assign random UUID if not specified by user before processing operation on all peers let uuid = uuid.unwrap_or_else(Uuid::new_v4); let collection_state = collection.state().await; if let Some(shard_key) = &shard_key && !collection_state.shards_key_mapping.contains_key(shard_key) { return Err(StorageError::bad_request(format!( "sharding key {shard_key} does not exist for collection {collection_name}", ))); } let shard_id = match (direction, shard_key.as_ref()) { // When scaling up, just pick the next shard ID (ReshardingDirection::Up, _) => { collection_state .shards .keys() .copied() .max() .expect("collection must contain shards") + 1 } // When scaling down without shard keys, pick the last shard ID (ReshardingDirection::Down, None) => collection_state .shards .keys() .copied() .max() .expect("collection must contain shards"), // When scaling down with shard keys, pick the last shard ID of that key (ReshardingDirection::Down, Some(shard_key)) => collection_state .shards_key_mapping .get(shard_key) .expect("specified shard key must exist") .iter() .copied() .max() .expect("collection must contain shards"), }; let peer_id = match (peer_id, direction) { // Select user specified peer, but make sure it exists (Some(peer_id), _) => { validate_peer_exists(peer_id)?; peer_id } // When scaling up, select peer with least number of shards for this collection (None, ReshardingDirection::Up) => { let mut shards_on_peers = collection_state .shards .values() .flat_map(|shard_info| shard_info.replicas.keys()) .fold(HashMap::new(), |mut counts, peer_id| { *counts.entry(*peer_id).or_insert(0) += 1; counts }); for peer_id in get_all_peer_ids() { // Add registered peers not holding any shard yet shards_on_peers.entry(peer_id).or_insert(0); } shards_on_peers .into_iter() .min_by_key(|(_, count)| *count) .map(|(peer_id, _)| peer_id) .expect("expected at least one peer") } // When scaling down, select random peer that contains the shard we're dropping // Other peers work, but are less efficient due to remote operations (None, ReshardingDirection::Down) => collection_state .shards .get(&shard_id) .expect("select shard ID must always exist in collection state") .replicas .keys() .choose(&mut rand::rng()) .copied() .unwrap(), }; if let Some(resharding) = &collection_state.resharding { return Err(StorageError::bad_request(format!( "resharding {resharding:?} is already in progress \ for collection {collection_name}" ))); } dispatcher .submit_collection_meta_op( CollectionMetaOperations::Resharding( collection_name.clone(), ReshardingOperation::Start(ReshardKey { uuid, direction, peer_id, shard_id, shard_key, }), ), access, wait_timeout, ) .await } ClusterOperations::AbortResharding(_) => { // TODO(reshading): Deduplicate resharding operations handling? let Some(state) = collection.resharding_state().await else { return Err(StorageError::bad_request(format!( "resharding is not in progress for collection {collection_name}" ))); }; dispatcher .submit_collection_meta_op( CollectionMetaOperations::Resharding( collection_name.clone(), ReshardingOperation::Abort(ReshardKey { uuid: state.uuid, direction: state.direction, peer_id: state.peer_id, shard_id: state.shard_id, shard_key: state.shard_key.clone(), }), ), access, wait_timeout, ) .await } ClusterOperations::FinishResharding(_) => { // TODO(resharding): Deduplicate resharding operations handling? let Some(state) = collection.resharding_state().await else { return Err(StorageError::bad_request(format!( "resharding is not in progress for collection {collection_name}" ))); }; dispatcher .submit_collection_meta_op( CollectionMetaOperations::Resharding( collection_name.clone(), ReshardingOperation::Finish(state.key()), ), access, wait_timeout, ) .await } ClusterOperations::FinishMigratingPoints(op) => { // TODO(resharding): Deduplicate resharding operations handling? let Some(state) = collection.resharding_state().await else { return Err(StorageError::bad_request(format!( "resharding is not in progress for collection {collection_name}" ))); }; let op = op.finish_migrating_points; let shard_id = match (op.shard_id, state.direction) { (Some(shard_id), _) => shard_id, (None, ReshardingDirection::Up) => state.shard_id, (None, ReshardingDirection::Down) => { return Err(StorageError::bad_request( "shard ID must be specified when resharding down", )); } }; let peer_id = match (op.peer_id, state.direction) { (Some(peer_id), _) => peer_id, (None, ReshardingDirection::Up) => state.peer_id, (None, ReshardingDirection::Down) => { return Err(StorageError::bad_request( "peer ID must be specified when resharding down", )); } }; let from_state = match state.direction { ReshardingDirection::Up => replica_set_state::ReplicaState::Resharding, ReshardingDirection::Down => replica_set_state::ReplicaState::ReshardingScaleDown, }; dispatcher .submit_collection_meta_op( CollectionMetaOperations::SetShardReplicaState(SetShardReplicaState { collection_name: collection_name.clone(), shard_id, peer_id, state: replica_set_state::ReplicaState::Active, from_state: Some(from_state), }), access, wait_timeout, ) .await } ClusterOperations::CommitReadHashRing(_) => { // TODO(reshading): Deduplicate resharding operations handling? let Some(state) = collection.resharding_state().await else { return Err(StorageError::bad_request(format!( "resharding is not in progress for collection {collection_name}" ))); }; // TODO(resharding): Add precondition checks? dispatcher .submit_collection_meta_op(
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
true
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/mod.rs
src/common/mod.rs
pub mod auth; pub mod collections; pub mod debugger; pub mod error_reporting; pub mod health; pub mod helpers; pub mod http_client; pub mod inference; pub mod metrics; pub mod pyroscope_state; pub mod query; pub mod snapshots; pub mod stacktrace; pub mod strict_mode; pub mod strings; pub mod telemetry; pub mod telemetry_ops; pub mod telemetry_reporting; pub mod update;
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/snapshots.rs
src/common/snapshots.rs
use std::sync::Arc; use collection::collection::Collection; use collection::common::sha_256; use collection::common::snapshot_stream::SnapshotStream; use collection::operations::snapshot_ops::{ ShardSnapshotLocation, SnapshotDescription, SnapshotPriority, }; use collection::operations::verification::VerificationPass; use collection::shards::replica_set::replica_set_state::ReplicaState; use collection::shards::replica_set::snapshots::RecoveryType; use collection::shards::shard::ShardId; use common::tempfile_ext::MaybeTempPath; use segment::data_types::manifest::SnapshotManifest; use storage::content_manager::errors::StorageError; use storage::content_manager::snapshots; use storage::content_manager::toc::TableOfContent; use storage::dispatcher::Dispatcher; use storage::rbac::{Access, AccessRequirements}; use tokio::sync::OwnedRwLockWriteGuard; use super::http_client::HttpClient; /// # Cancel safety /// /// This function is cancel safe. pub async fn create_shard_snapshot( toc: Arc<TableOfContent>, access: Access, collection_name: String, shard_id: ShardId, ) -> Result<SnapshotDescription, StorageError> { let collection_pass = access .check_collection_access(&collection_name, AccessRequirements::new().write().extras())?; let collection = toc.get_collection(&collection_pass).await?; let _telemetry_scope_guard = toc .snapshot_telemetry_collector(&collection_name) .running_snapshots .measure_scope(); let snapshot = collection .create_shard_snapshot(shard_id, &toc.optional_temp_or_snapshot_temp_path()?) .await?; Ok(snapshot) } /// # Cancel safety /// /// This function is cancel safe. pub async fn stream_shard_snapshot( toc: Arc<TableOfContent>, access: Access, collection_name: String, shard_id: ShardId, manifest: Option<SnapshotManifest>, ) -> Result<SnapshotStream, StorageError> { let collection_pass = access .check_collection_access(&collection_name, AccessRequirements::new().write().extras())?; let collection = toc.get_collection(&collection_pass).await?; let _telemetry_scope_guard = toc .snapshot_telemetry_collector(&collection_name) .running_snapshots .measure_scope(); if let Some(old_manifest) = &manifest { let current_manifest = collection.get_partial_snapshot_manifest(shard_id).await?; // If `old_manifest` is *exactly* the same, as `current_manifest`, return specialized error // instead of creating partial snapshot. // // Snapshot manifest format is flexible, so it *might* be possible that manifests are *not* // exactly the same, but resulting partial snapshot will still be "empty", but: // - it should *not* happen in practice currently // - we intentionally use exact equality as the most "conservative" comparison, just in case if old_manifest == &current_manifest { return Err(StorageError::EmptyPartialSnapshot { shard_id }); } } let snapshot_stream = toc .get_collection(&collection_pass) .await? .stream_shard_snapshot( shard_id, manifest, &toc.optional_temp_or_snapshot_temp_path()?, ) .await?; Ok(snapshot_stream) } /// # Cancel safety /// /// This function is cancel safe. pub async fn list_shard_snapshots( toc: Arc<TableOfContent>, access: Access, collection_name: String, shard_id: ShardId, ) -> Result<Vec<SnapshotDescription>, StorageError> { let collection_pass = access.check_collection_access(&collection_name, AccessRequirements::new().extras())?; let collection = toc.get_collection(&collection_pass).await?; let snapshots = collection.list_shard_snapshots(shard_id).await?; Ok(snapshots) } /// # Cancel safety /// /// This function is cancel safe. pub async fn delete_shard_snapshot( toc: Arc<TableOfContent>, access: Access, collection_name: String, shard_id: ShardId, snapshot_name: String, ) -> Result<(), StorageError> { let collection_pass = access .check_collection_access(&collection_name, AccessRequirements::new().write().extras())?; let collection = toc.get_collection(&collection_pass).await?; let snapshot_manager = collection.get_snapshots_storage_manager()?; let snapshot_path = collection .shards_holder() .read() .await .get_shard_snapshot_path(collection.snapshots_path(), shard_id, &snapshot_name) .await?; tokio::spawn(async move { snapshot_manager.delete_snapshot(&snapshot_path).await }).await??; Ok(()) } /// # Cancel safety /// /// This function is cancel safe. #[allow(clippy::too_many_arguments)] pub async fn recover_shard_snapshot( toc: Arc<TableOfContent>, access: Access, collection_name: String, shard_id: ShardId, snapshot_location: ShardSnapshotLocation, snapshot_priority: SnapshotPriority, checksum: Option<String>, client: HttpClient, api_key: Option<String>, ) -> Result<(), StorageError> { let collection_pass = access .check_global_access(AccessRequirements::new().manage())? .issue_pass(&collection_name) .into_static(); // - `recover_shard_snapshot_impl` is *not* cancel safe // - but the task is *spawned* on the runtime and won't be cancelled, if request is cancelled cancel::future::spawn_cancel_on_drop(async move |cancel| { let cancel_safe = async { let collection = toc.get_collection(&collection_pass).await?; collection.assert_shard_exists(shard_id).await?; let download_dir = toc.optional_temp_or_snapshot_temp_path()?; let snapshot_path = match snapshot_location { ShardSnapshotLocation::Url(url) => { if !matches!(url.scheme(), "http" | "https") { let description = format!( "Invalid snapshot URL {url}: URLs with {} scheme are not supported", url.scheme(), ); return Err(StorageError::bad_input(description)); } let client = client.client(api_key.as_deref())?; snapshots::download::download_snapshot(&client, url, &download_dir).await? } ShardSnapshotLocation::Path(snapshot_file_name) => { let snapshot_path = collection .shards_holder() .read() .await .get_shard_snapshot_path( collection.snapshots_path(), shard_id, &snapshot_file_name, ) .await?; collection .get_snapshots_storage_manager()? .get_snapshot_file(&snapshot_path, &download_dir) .await? } }; if let Some(checksum) = checksum { let snapshot_checksum = sha_256::hash_file(&snapshot_path).await?; if !sha_256::hashes_equal(&snapshot_checksum, &checksum) { return Err(StorageError::bad_input(format!( "Snapshot checksum mismatch: expected {checksum}, got {snapshot_checksum}" ))); } } Ok((collection, snapshot_path)) }; let (collection, snapshot_path) = cancel::future::cancel_on_token(cancel.clone(), cancel_safe).await??; // `recover_shard_snapshot_impl` is *not* cancel safe recover_shard_snapshot_impl( &toc, &collection, shard_id, snapshot_path, snapshot_priority, RecoveryType::Full, cancel, ) .await }) .await??; Ok(()) } /// # Cancel safety /// /// This function is *not* cancel safe. pub async fn recover_shard_snapshot_impl( toc: &TableOfContent, collection: &Collection, shard: ShardId, snapshot_path: MaybeTempPath, priority: SnapshotPriority, recovery_type: RecoveryType, cancel: cancel::CancellationToken, ) -> Result<(), StorageError> { let _recover_tracker_guard = toc .snapshot_telemetry_collector(collection.name()) .running_snapshot_recovery .measure_scope(); // `Collection::restore_shard_snapshot` and `activate_shard` calls *have to* be executed as a // single transaction // // It is *possible* to make this function to be cancel safe, but it is *extremely tedious* to do so // TODO: `Collection::restore_shard_snapshot` *is* cancel-safe, but `recover_shard_snapshot_impl` is *not* cancel-safe (yet) collection .restore_shard_snapshot( shard, snapshot_path, recovery_type, toc.this_peer_id, toc.is_distributed(), &toc.optional_temp_or_snapshot_temp_path()?, cancel, ) .await? .await?; let state = collection.state().await; let shard_info = state.shards.get(&shard).unwrap(); // TODO: Handle `unwrap`?.. // TODO: Unify (and de-duplicate) "recovered shard state notification" logic in `_do_recover_from_snapshot` with this one! let other_active_replicas: Vec<_> = shard_info .replicas .iter() .map(|(&peer, &state)| (peer, state)) .filter(|&(peer, state)| { // Check if there are *other* active replicas, after recovering shard snapshot. // This should include `ReshardingScaleDown` replicas. let is_active = matches!( state, ReplicaState::Active | ReplicaState::ReshardingScaleDown ); peer != toc.this_peer_id && is_active }) .collect(); if other_active_replicas.is_empty() || recovery_type.is_partial() { snapshots::recover::activate_shard(toc, collection, toc.this_peer_id, &shard).await?; } else { match priority { SnapshotPriority::NoSync => { snapshots::recover::activate_shard(toc, collection, toc.this_peer_id, &shard) .await?; } SnapshotPriority::Snapshot => { snapshots::recover::activate_shard(toc, collection, toc.this_peer_id, &shard) .await?; for &(peer, _) in other_active_replicas.iter() { toc.send_set_replica_state_proposal( collection.name().to_string(), peer, shard, ReplicaState::Dead, None, )?; } } SnapshotPriority::Replica => { toc.send_set_replica_state_proposal( collection.name().to_string(), toc.this_peer_id, shard, ReplicaState::Dead, None, )?; } // `ShardTransfer` is only used during snapshot *shard transfer*. // State transitions are performed as part of shard transfer *later*, so this simply does *nothing*. SnapshotPriority::ShardTransfer => (), } } Ok(()) } pub async fn try_take_partial_snapshot_recovery_lock( dispatcher: &Dispatcher, collection_name: &str, shard_id: ShardId, access: &Access, pass: &VerificationPass, ) -> Result<Option<OwnedRwLockWriteGuard<()>>, StorageError> { let collection_pass = access .check_global_access(AccessRequirements::new().manage())? .issue_pass(collection_name); let recovery_lock = dispatcher .toc(access, pass) .get_collection(&collection_pass) .await? .try_take_partial_snapshot_recovery_lock(shard_id, RecoveryType::Partial) .await?; Ok(recovery_lock) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/health.rs
src/common/health.rs
use std::collections::HashSet; use std::future::{self, Future}; use std::sync::Arc; use std::sync::atomic::{self, AtomicBool}; use std::time::Duration; use std::{panic, thread}; use api::grpc::qdrant::qdrant_internal_client::QdrantInternalClient; use api::grpc::qdrant::{GetConsensusCommitRequest, GetConsensusCommitResponse}; use api::grpc::transport_channel_pool::{self, TransportChannelPool}; use collection::shards::CollectionId; use collection::shards::shard::ShardId; use common::defaults; use futures::stream::FuturesUnordered; use futures::{FutureExt as _, StreamExt as _, TryStreamExt as _}; use itertools::Itertools; use storage::content_manager::consensus_manager::ConsensusStateRef; use storage::content_manager::toc::TableOfContent; use storage::rbac::Access; use tokio::{runtime, sync, time}; const READY_CHECK_TIMEOUT: Duration = Duration::from_millis(500); const GET_CONSENSUS_COMMITS_RETRIES: usize = 2; /// Structure used to process health checks like `/readyz` endpoints. pub struct HealthChecker { // The state of the health checker. // Once set to `true`, it should not change back to `false`. // Initially set to `false`. is_ready: Arc<AtomicBool>, // The signal that notifies that state has changed. // Comes from the health checker task. is_ready_signal: Arc<sync::Notify>, // Signal to the health checker task, that the API was called. // Used to drive the health checker task and avoid constant polling. check_ready_signal: Arc<sync::Notify>, _cancel: cancel::DropGuard, } impl HealthChecker { pub fn spawn( toc: Arc<TableOfContent>, consensus_state: ConsensusStateRef, runtime: &runtime::Handle, wait_for_bootstrap: bool, ) -> Self { let task = Task { toc, consensus_state, is_ready: Default::default(), is_ready_signal: Default::default(), check_ready_signal: Default::default(), cancel: Default::default(), wait_for_bootstrap, }; let health_checker = Self { is_ready: task.is_ready.clone(), is_ready_signal: task.is_ready_signal.clone(), check_ready_signal: task.check_ready_signal.clone(), _cancel: task.cancel.clone().drop_guard(), }; let task = runtime.spawn(task.exec()); drop(task); // drop `JoinFuture` explicitly to make clippy happy health_checker } pub async fn check_ready(&self) -> bool { if self.is_ready() { return true; } self.notify_task(); self.wait_ready().await } pub fn is_ready(&self) -> bool { self.is_ready.load(atomic::Ordering::Relaxed) } pub fn notify_task(&self) { self.check_ready_signal.notify_one(); } async fn wait_ready(&self) -> bool { let is_ready_signal = self.is_ready_signal.notified(); if self.is_ready() { return true; } time::timeout(READY_CHECK_TIMEOUT, is_ready_signal) .await .is_ok() } } pub struct Task { toc: Arc<TableOfContent>, consensus_state: ConsensusStateRef, // Shared state with the health checker // Once set to `true`, it should not change back to `false`. is_ready: Arc<AtomicBool>, // Used to notify the health checker service that the state has changed. is_ready_signal: Arc<sync::Notify>, // Driver signal for the health checker task // Once received, the task should proceed with an attempt to check the state. // Usually comes from the API call, but can be triggered by the task itself. check_ready_signal: Arc<sync::Notify>, cancel: cancel::CancellationToken, wait_for_bootstrap: bool, } impl Task { pub async fn exec(self) { while let Err(err) = self.exec_catch_unwind().await { let message = common::panic::downcast_str(&err).unwrap_or(""); let separator = if !message.is_empty() { ": " } else { "" }; log::error!("HealthChecker task panicked, retrying{separator}{message}",); } } async fn exec_catch_unwind(&self) -> thread::Result<()> { panic::AssertUnwindSafe(self.exec_cancel()) .catch_unwind() .await } async fn exec_cancel(&self) { let _ = cancel::future::cancel_on_token(self.cancel.clone(), self.exec_impl()).await; } async fn exec_impl(&self) { // Wait until node joins cluster for the first time // // If this is a new deployment and `--bootstrap` CLI parameter was specified... if self.wait_for_bootstrap { // Check if this is the only node in the cluster while self.consensus_state.peer_count() <= 1 { // If cluster is empty, make another attempt to check // after we receive another call to `/readyz` // // Wait for `/readyz` signal self.check_ready_signal.notified().await; } } // Artificial simulate signal from `/readyz` endpoint // as if it was already called by the user. // This allows to check the happy path without waiting for the first call. self.check_ready_signal.notify_one(); // Get estimate of current cluster commit so we can wait for it let Some(mut cluster_commit_index) = self.cluster_commit_index(true).await else { self.set_ready(); return; }; // Wait until local peer has reached cluster commit loop { while self.commit_index() < cluster_commit_index { // Wait for `/readyz` signal self.check_ready_signal.notified().await; // Ensure we're not the only peer left if self.consensus_state.peer_count() <= 1 { self.set_ready(); return; } } match self.cluster_commit_index(false).await { // If cluster commit is still the same, we caught up and we're done Some(new_index) if cluster_commit_index == new_index => break, // Cluster commit is newer, update it and wait again Some(new_index) => cluster_commit_index = new_index, // Failed to get cluster commit, assume we're done None => break, } } // Collect "unhealthy" shards list let mut unhealthy_shards = self.unhealthy_shards().await; // Check if all shards are "healthy"... while !unhealthy_shards.is_empty() { // If not: // // - Wait for `/readyz` signal self.check_ready_signal.notified().await; // - Refresh "unhealthy" shards list let current_unhealthy_shards = self.unhealthy_shards().await; // - Check if any shards "healed" since last check unhealthy_shards.retain(|shard| current_unhealthy_shards.contains(shard)); } self.set_ready(); } /// Get the highest consensus commit across cluster peers /// /// If `one_peer` is true the first fetched commit is returned. It may not necessarily be the /// latest commit. async fn cluster_commit_index(&self, one_peer: bool) -> Option<u64> { // Wait for `/readyz` signal self.check_ready_signal.notified().await; // Check if there is only 1 node in the cluster if self.consensus_state.peer_count() <= 1 { return None; } // Get *cluster* commit index let peer_address_by_id = self.consensus_state.peer_address_by_id(); let transport_channel_pool = &self.toc.get_channel_service().channel_pool; let this_peer_id = self.toc.this_peer_id; let this_peer_uri = peer_address_by_id.get(&this_peer_id); let mut requests = peer_address_by_id .values() // Do not get the current commit from ourselves .filter(|&uri| Some(uri) != this_peer_uri) // Historic peers might use the same URLs as our current peers, request each URI once .unique() .map(|uri| get_consensus_commit(transport_channel_pool, uri)) .collect::<FuturesUnordered<_>>() .inspect_err(|err| log::error!("GetConsensusCommit request failed: {err}")) .filter_map(|res| future::ready(res.ok())); // Raft commits consensus operation, after majority of nodes persisted it. // // This means, if we check the majority of nodes (e.g., `total nodes / 2 + 1`), at least one // of these nodes will *always* have an up-to-date commit index. And so, the highest commit // index among majority of nodes *is* the cluster commit index. // // Our current node *is* one of the cluster nodes, so it's enough to query `total nodes / 2` // *additional* nodes, to get cluster commit index. // // The check goes like this: // - Either at least one of the "additional" nodes return a *higher* commit index, which // means our node is *not* up-to-date, and we have to wait to reach this commit index // - Or *all* of them return *lower* commit index, which means current node is *already* // up-to-date, and `/readyz` check will pass to the next step // // Example: // // Total nodes: 2 // Required: 2 / 2 = 1 // // Total nodes: 3 // Required: 3 / 2 = 1 // // Total nodes: 4 // Required: 4 / 2 = 2 // // Total nodes: 5 // Required: 5 / 2 = 2 let sufficient_commit_indices_count = if !one_peer { peer_address_by_id.len() / 2 } else { 1 }; // *Wait* for `total nodex / 2` successful responses... let mut commit_indices: Vec<_> = (&mut requests) .take(sufficient_commit_indices_count) .collect() .await; // ...and also collect any additional responses, that we might have *already* received while let Ok(Some(resp)) = time::timeout(Duration::ZERO, requests.next()).await { commit_indices.push(resp); } // Find the maximum commit index among all responses. // // Note, that we progress even if most (or even *all*) requests failed (e.g., because all // other nodes are unavailable or they don't support `GetConsensusCommit` gRPC API). // // So this check is not 100% reliable and can give a false-positive result! let cluster_commit_index = commit_indices .into_iter() .map(|resp| resp.into_inner().commit) .max() .unwrap_or(0); Some(cluster_commit_index as _) } fn commit_index(&self) -> u64 { // TODO: Blocking call in async context!? self.consensus_state .persistent .read() .last_applied_entry() .unwrap_or(0) } /// List shards that are unhealthy, which may undergo automatic recovery. /// /// Shards in resharding state are not considered unhealthy and are excluded here. /// They require an external driver to make them active or to drop them. async fn unhealthy_shards(&self) -> HashSet<Shard> { let this_peer_id = self.toc.this_peer_id; let collections = self .toc .all_collections(&Access::full("For health check")) .await; let mut unhealthy_shards = HashSet::new(); for collection_pass in &collections { let state = match self.toc.get_collection(collection_pass).await { Ok(collection) => collection.state().await, Err(_) => continue, }; for (&shard, info) in state.shards.iter() { let Some(state) = info.replicas.get(&this_peer_id) else { continue; }; if state.is_healthy() { continue; } unhealthy_shards.insert(Shard::new(collection_pass.name(), shard)); } } unhealthy_shards } fn set_ready(&self) { self.is_ready.store(true, atomic::Ordering::Relaxed); self.is_ready_signal.notify_waiters(); } } fn get_consensus_commit<'a>( transport_channel_pool: &'a TransportChannelPool, uri: &'a tonic::transport::Uri, ) -> impl Future<Output = GetConsensusCommitResult> + 'a { transport_channel_pool.with_channel_timeout( uri, |channel| async { let mut client = QdrantInternalClient::new(channel); let mut request = tonic::Request::new(GetConsensusCommitRequest {}); request.set_timeout(defaults::CONSENSUS_META_OP_WAIT); client.get_consensus_commit(request).await }, Some(defaults::CONSENSUS_META_OP_WAIT), GET_CONSENSUS_COMMITS_RETRIES, ) } type GetConsensusCommitResult = Result< tonic::Response<GetConsensusCommitResponse>, transport_channel_pool::RequestError<tonic::Status>, >; #[derive(Clone, Debug, Eq, PartialEq, Hash)] struct Shard { collection: CollectionId, shard: ShardId, } impl Shard { pub fn new(collection: impl Into<CollectionId>, shard: ShardId) -> Self { Self { collection: collection.into(), shard, } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/stacktrace.rs
src/common/stacktrace.rs
use schemars::JsonSchema; use serde::{Deserialize, Serialize}; #[derive(Deserialize, Serialize, JsonSchema, Debug)] struct StackTraceSymbol { name: Option<String>, file: Option<String>, line: Option<u32>, } #[derive(Deserialize, Serialize, JsonSchema, Debug)] struct StackTraceFrame { symbols: Vec<StackTraceSymbol>, } impl StackTraceFrame { #[allow(dead_code)] pub fn render(&self) -> String { let mut result = String::new(); for symbol in &self.symbols { let symbol_string = format!( "{}:{} - {} ", symbol.file.as_deref().unwrap_or_default(), symbol.line.unwrap_or_default(), symbol.name.as_deref().unwrap_or_default(), ); result.push_str(&symbol_string); } result } } #[derive(Deserialize, Serialize, JsonSchema, Debug)] pub struct ThreadStackTrace { id: u32, name: String, frames: Vec<String>, } #[derive(Deserialize, Serialize, JsonSchema, Debug)] pub struct StackTrace { threads: Vec<ThreadStackTrace>, } pub fn get_stack_trace() -> StackTrace { #[cfg(not(all(target_os = "linux", feature = "stacktrace")))] { StackTrace { threads: vec![] } } #[cfg(all(target_os = "linux", feature = "stacktrace"))] { let exe = std::env::current_exe().unwrap(); let trace = rstack_self::trace(std::process::Command::new(exe).arg("--stacktrace")).unwrap(); StackTrace { threads: trace .threads() .iter() .map(|thread| ThreadStackTrace { id: thread.id(), name: thread.name().to_string(), frames: thread .frames() .iter() .map(|frame| { let frame = StackTraceFrame { symbols: frame .symbols() .iter() .map(|symbol| StackTraceSymbol { name: symbol.name().map(|name| name.to_string()), file: symbol.file().map(|file| { file.to_str().unwrap_or_default().to_string() }), line: symbol.line(), }) .collect(), }; frame.render() }) .collect(), }) .collect(), } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/strings.rs
src/common/strings.rs
/// Constant-time equality for String types #[inline] pub fn ct_eq(lhs: impl AsRef<str>, rhs: impl AsRef<str>) -> bool { constant_time_eq::constant_time_eq(lhs.as_ref().as_bytes(), rhs.as_ref().as_bytes()) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/query.rs
src/common/query.rs
use std::time::Duration; use api::rest::SearchGroupsRequestInternal; use collection::collection::distance_matrix::*; use collection::common::batching::batch_requests; use collection::grouping::group_by::GroupRequest; use collection::operations::consistency_params::ReadConsistency; use collection::operations::shard_selector_internal::ShardSelectorInternal; use collection::operations::types::*; use collection::operations::universal_query::collection_query::*; use common::counter::hardware_accumulator::HwMeasurementAcc; use segment::types::ScoredPoint; use shard::retrieve::record_internal::RecordInternal; use shard::search::CoreSearchRequestBatch; use storage::content_manager::errors::StorageError; use storage::content_manager::toc::TableOfContent; use storage::rbac::Access; #[allow(clippy::too_many_arguments)] pub async fn do_core_search_points( toc: &TableOfContent, collection_name: &str, request: CoreSearchRequest, read_consistency: Option<ReadConsistency>, shard_selection: ShardSelectorInternal, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<Vec<ScoredPoint>, StorageError> { let batch_res = do_core_search_batch_points( toc, collection_name, CoreSearchRequestBatch { searches: vec![request], }, read_consistency, shard_selection, access, timeout, hw_measurement_acc, ) .await?; batch_res .into_iter() .next() .ok_or_else(|| StorageError::service_error("Empty search result")) } pub async fn do_search_batch_points( toc: &TableOfContent, collection_name: &str, requests: Vec<(CoreSearchRequest, ShardSelectorInternal)>, read_consistency: Option<ReadConsistency>, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<Vec<Vec<ScoredPoint>>, StorageError> { let requests = batch_requests::< (CoreSearchRequest, ShardSelectorInternal), ShardSelectorInternal, Vec<CoreSearchRequest>, Vec<_>, >( requests, |(_, shard_selector)| shard_selector, |(request, _), core_reqs| { core_reqs.push(request); Ok(()) }, |shard_selector, core_requests, res| { if core_requests.is_empty() { return Ok(()); } let core_batch = CoreSearchRequestBatch { searches: core_requests, }; let req = toc.core_search_batch( collection_name, core_batch, read_consistency, shard_selector, access.clone(), timeout, hw_measurement_acc.clone(), ); res.push(req); Ok(()) }, )?; let results = futures::future::try_join_all(requests).await?; let flatten_results: Vec<Vec<_>> = results.into_iter().flatten().collect(); Ok(flatten_results) } #[allow(clippy::too_many_arguments)] pub async fn do_core_search_batch_points( toc: &TableOfContent, collection_name: &str, request: CoreSearchRequestBatch, read_consistency: Option<ReadConsistency>, shard_selection: ShardSelectorInternal, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<Vec<Vec<ScoredPoint>>, StorageError> { toc.core_search_batch( collection_name, request, read_consistency, shard_selection, access, timeout, hw_measurement_acc, ) .await } #[allow(clippy::too_many_arguments)] pub async fn do_search_point_groups( toc: &TableOfContent, collection_name: &str, request: SearchGroupsRequestInternal, read_consistency: Option<ReadConsistency>, shard_selection: ShardSelectorInternal, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<GroupsResult, StorageError> { toc.group( collection_name, GroupRequest::from(request), read_consistency, shard_selection, access, timeout, hw_measurement_acc, ) .await } #[allow(clippy::too_many_arguments)] pub async fn do_recommend_point_groups( toc: &TableOfContent, collection_name: &str, request: RecommendGroupsRequestInternal, read_consistency: Option<ReadConsistency>, shard_selection: ShardSelectorInternal, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<GroupsResult, StorageError> { toc.group( collection_name, GroupRequest::from(request), read_consistency, shard_selection, access, timeout, hw_measurement_acc, ) .await } pub async fn do_discover_batch_points( toc: &TableOfContent, collection_name: &str, request: DiscoverRequestBatch, read_consistency: Option<ReadConsistency>, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<Vec<Vec<ScoredPoint>>, StorageError> { let requests = request .searches .into_iter() .map(|req| { let shard_selector = match req.shard_key { None => ShardSelectorInternal::All, Some(shard_key) => ShardSelectorInternal::from(shard_key), }; (req.discover_request, shard_selector) }) .collect(); toc.discover_batch( collection_name, requests, read_consistency, access, timeout, hw_measurement_acc, ) .await } #[allow(clippy::too_many_arguments)] pub async fn do_count_points( toc: &TableOfContent, collection_name: &str, request: CountRequestInternal, read_consistency: Option<ReadConsistency>, timeout: Option<Duration>, shard_selection: ShardSelectorInternal, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<CountResult, StorageError> { toc.count( collection_name, request, read_consistency, timeout, shard_selection, access, hw_measurement_acc, ) .await } #[allow(clippy::too_many_arguments)] pub async fn do_get_points( toc: &TableOfContent, collection_name: &str, request: PointRequestInternal, read_consistency: Option<ReadConsistency>, timeout: Option<Duration>, shard_selection: ShardSelectorInternal, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<Vec<RecordInternal>, StorageError> { toc.retrieve( collection_name, request, read_consistency, timeout, shard_selection, access, hw_measurement_acc, ) .await } #[allow(clippy::too_many_arguments)] pub async fn do_scroll_points( toc: &TableOfContent, collection_name: &str, request: ScrollRequestInternal, read_consistency: Option<ReadConsistency>, timeout: Option<Duration>, shard_selection: ShardSelectorInternal, access: Access, hw_measurement_acc: HwMeasurementAcc, ) -> Result<ScrollResult, StorageError> { toc.scroll( collection_name, request, read_consistency, timeout, shard_selection, access, hw_measurement_acc, ) .await } #[allow(clippy::too_many_arguments)] pub async fn do_query_points( toc: &TableOfContent, collection_name: &str, request: CollectionQueryRequest, read_consistency: Option<ReadConsistency>, shard_selection: ShardSelectorInternal, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<Vec<ScoredPoint>, StorageError> { let requests = vec![(request, shard_selection)]; let batch_res = toc .query_batch( collection_name, requests, read_consistency, access, timeout, hw_measurement_acc, ) .await?; batch_res .into_iter() .next() .ok_or_else(|| StorageError::service_error("Empty query result")) } #[allow(clippy::too_many_arguments)] pub async fn do_query_batch_points( toc: &TableOfContent, collection_name: &str, requests: Vec<(CollectionQueryRequest, ShardSelectorInternal)>, read_consistency: Option<ReadConsistency>, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<Vec<Vec<ScoredPoint>>, StorageError> { toc.query_batch( collection_name, requests, read_consistency, access, timeout, hw_measurement_acc, ) .await } #[allow(clippy::too_many_arguments)] pub async fn do_query_point_groups( toc: &TableOfContent, collection_name: &str, request: CollectionQueryGroupsRequest, read_consistency: Option<ReadConsistency>, shard_selection: ShardSelectorInternal, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<GroupsResult, StorageError> { toc.group( collection_name, GroupRequest::from(request), read_consistency, shard_selection, access, timeout, hw_measurement_acc, ) .await } #[allow(clippy::too_many_arguments)] pub async fn do_search_points_matrix( toc: &TableOfContent, collection_name: &str, request: CollectionSearchMatrixRequest, read_consistency: Option<ReadConsistency>, shard_selection: ShardSelectorInternal, access: Access, timeout: Option<Duration>, hw_measurement_acc: HwMeasurementAcc, ) -> Result<CollectionSearchMatrixResponse, StorageError> { toc.search_points_matrix( collection_name, request, read_consistency, shard_selection, access, timeout, hw_measurement_acc, ) .await }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/metrics.rs
src/common/metrics.rs
use std::collections::HashMap; use api::rest::models::HardwareUsage; use collection::shards::replica_set::replica_set_state::ReplicaState; use itertools::Itertools; use prometheus::TextEncoder; use prometheus::proto::{Counter, Gauge, LabelPair, Metric, MetricFamily, MetricType}; use segment::common::operation_time_statistics::OperationDurationStatistics; use shard::PeerId; use storage::types::ConsensusThreadStatus; use super::telemetry_ops::hardware::HardwareTelemetry; use crate::common::telemetry::TelemetryData; use crate::common::telemetry_ops::app_telemetry::{AppBuildTelemetry, AppFeaturesTelemetry}; use crate::common::telemetry_ops::cluster_telemetry::{ClusterStatusTelemetry, ClusterTelemetry}; use crate::common::telemetry_ops::collections_telemetry::{ CollectionTelemetryEnum, CollectionsTelemetry, }; use crate::common::telemetry_ops::memory_telemetry::MemoryTelemetry; use crate::common::telemetry_ops::requests_telemetry::{ GrpcTelemetry, RequestsTelemetry, WebApiTelemetry, }; /// Whitelist for REST endpoints in metrics output. /// /// Contains selection of search, recommend, scroll and upsert endpoints. /// /// This array *must* be sorted. const REST_ENDPOINT_WHITELIST: &[&str] = &[ "/collections/{name}/index", "/collections/{name}/points", "/collections/{name}/points/batch", "/collections/{name}/points/count", "/collections/{name}/points/delete", "/collections/{name}/points/discover", "/collections/{name}/points/discover/batch", "/collections/{name}/points/facet", "/collections/{name}/points/payload", "/collections/{name}/points/payload/clear", "/collections/{name}/points/payload/delete", "/collections/{name}/points/query", "/collections/{name}/points/query/batch", "/collections/{name}/points/query/groups", "/collections/{name}/points/recommend", "/collections/{name}/points/recommend/batch", "/collections/{name}/points/recommend/groups", "/collections/{name}/points/scroll", "/collections/{name}/points/search", "/collections/{name}/points/search/batch", "/collections/{name}/points/search/groups", "/collections/{name}/points/search/matrix/offsets", "/collections/{name}/points/search/matrix/pairs", "/collections/{name}/points/vectors", "/collections/{name}/points/vectors/delete", ]; /// Whitelist for GRPC endpoints in metrics output. /// /// Contains selection of search, recommend, scroll and upsert endpoints. /// /// This array *must* be sorted. const GRPC_ENDPOINT_WHITELIST: &[&str] = &[ "/qdrant.Points/ClearPayload", "/qdrant.Points/Count", "/qdrant.Points/Delete", "/qdrant.Points/DeletePayload", "/qdrant.Points/Discover", "/qdrant.Points/DiscoverBatch", "/qdrant.Points/Facet", "/qdrant.Points/Get", "/qdrant.Points/OverwritePayload", "/qdrant.Points/Query", "/qdrant.Points/QueryBatch", "/qdrant.Points/QueryGroups", "/qdrant.Points/Recommend", "/qdrant.Points/RecommendBatch", "/qdrant.Points/RecommendGroups", "/qdrant.Points/Scroll", "/qdrant.Points/Search", "/qdrant.Points/SearchBatch", "/qdrant.Points/SearchGroups", "/qdrant.Points/SetPayload", "/qdrant.Points/UpdateBatch", "/qdrant.Points/UpdateVectors", "/qdrant.Points/Upsert", ]; /// For REST requests, only report timings when having this HTTP response status. const REST_TIMINGS_FOR_STATUS: u16 = 200; /// Encapsulates metrics data in Prometheus format. pub struct MetricsData { metrics: Vec<MetricFamily>, } impl MetricsData { pub fn format_metrics(&self) -> String { TextEncoder::new().encode_to_string(&self.metrics).unwrap() } /// Creates a new `MetricsData` from telemetry data and an optional prefix for metrics names. pub fn new_from_telemetry(telemetry_data: TelemetryData, prefix: Option<&str>) -> Self { let mut metrics = MetricsData::empty(); telemetry_data.add_metrics(&mut metrics, prefix); metrics } /// Adds the given `metrics_family` to the `MetricsData` collection, if `Some`. fn push_metric(&mut self, metric_family: Option<MetricFamily>) { if let Some(metric_family) = metric_family { self.metrics.push(metric_family); } } /// Creates an empty collection of Prometheus metric families `MetricsData`. /// This should only be used when you explicitly need an empty collection to gather metrics. /// /// In most cases, you should use [`MetricsData::new_from_telemetry`] to initialize new metrics data. fn empty() -> Self { Self { metrics: vec![] } } } trait MetricsProvider { /// Add metrics definitions for this. fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>); } impl MetricsProvider for TelemetryData { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { self.app.add_metrics(metrics, prefix); let this_peer_id = self.cluster.as_ref().and_then(|i| i.this_peer_id()); self.collections.add_metrics(metrics, prefix, this_peer_id); if let Some(cluster) = &self.cluster { cluster.add_metrics(metrics, prefix); } if let Some(requests) = &self.requests { requests.add_metrics(metrics, prefix); } if let Some(hardware) = &self.hardware { hardware.add_metrics(metrics, prefix); } if let Some(mem) = &self.memory { mem.add_metrics(metrics, prefix); } #[cfg(target_os = "linux")] match procfs_metrics::ProcFsMetrics::collect() { Ok(procfs_provider) => procfs_provider.add_metrics(metrics, prefix), Err(err) => log::warn!("Error reading procfs infos: {err:?}"), }; } } impl MetricsProvider for AppBuildTelemetry { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { metrics.push_metric(metric_family( "app_info", "information about qdrant server", MetricType::GAUGE, vec![gauge( 1.0, &[("name", &self.name), ("version", &self.version)], )], prefix, )); self.features .iter() .for_each(|f| f.add_metrics(metrics, prefix)); } } impl MetricsProvider for AppFeaturesTelemetry { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { metrics.push_metric(metric_family( "app_status_recovery_mode", "features enabled in qdrant server", MetricType::GAUGE, vec![gauge(if self.recovery_mode { 1.0 } else { 0.0 }, &[])], prefix, )) } } impl CollectionsTelemetry { fn add_metrics( &self, metrics: &mut MetricsData, prefix: Option<&str>, peer_id: Option<PeerId>, ) { metrics.push_metric(metric_family( "collections_total", "number of collections", MetricType::GAUGE, vec![gauge(self.number_of_collections as f64, &[])], prefix, )); let num_collections = self.collections.as_ref().map_or(0, |c| c.len()); // Optimizers let mut total_optimizations_running = Vec::with_capacity(num_collections); // Min/Max/Expected/Active replicas over all shards. let mut total_min_active_replicas = usize::MAX; let mut total_max_active_replicas = 0; // Points per collection let mut points_per_collection = Vec::with_capacity(num_collections); // Vectors excluded from index-only requests. let mut indexed_only_excluded = Vec::with_capacity(num_collections); let mut total_dead_replicas = 0; // Snapshot metrics let mut snapshots_creation_running = Vec::with_capacity(num_collections); let mut snapshots_recovery_running = Vec::with_capacity(num_collections); let mut snapshots_created_total = Vec::with_capacity(num_collections); let mut vector_count_by_name = Vec::with_capacity(num_collections); // Shard transfers let mut shard_transfers_in = Vec::with_capacity(num_collections); let mut shard_transfers_out = Vec::with_capacity(num_collections); for collection in self.collections.iter().flatten() { let collection = match collection { CollectionTelemetryEnum::Full(collection_telemetry) => collection_telemetry, CollectionTelemetryEnum::Aggregated(_) => { continue; } }; total_optimizations_running.push(gauge( collection.count_optimizers_running() as f64, &[("id", &collection.id)], )); let min_max_active_replicas = collection .shards .iter() .flatten() // While resharding up, some (shard) replica sets may still be incomplete during // the resharding process. In that case we don't want to consider these replica // sets at all in the active replica calculation. This is fine because searches nor // updates don't depend on them being available yet. // // More specifically: // - in stage 2 (migrate points) of resharding up we don't rely on the replica // to be available yet. In this stage, these replicas will have the `Resharding` // state. // - in stage 3 (replicate) of resharding up we activate the the replica and // replicate to match the configuration replication factor. From this point on we // do rely on the replica to be available. Now one replica will be `Active`, and // the other replicas will be in a transfer state. No replica will have `Resharding` // state. // // So, during stage 2 of resharding up we don't want to adjust the minimum number // of active replicas downwards. During stage 3 we do want it to affect the minimum // available replica number. It will be 1 for some time until replication transfers // complete. // // To ignore a (shard) replica set that is in stage 2 of resharding up, we simply // check if any of it's replicas is in `Resharding` state. .filter(|shard| { !shard .replicate_states .values() .any(|i| matches!(i, ReplicaState::Resharding)) }) .map(|shard| { shard .replicate_states .values() // While resharding down, all the replicas that we keep will get the // `ReshardingScaleDown` state for a period of time. We simply consider // these replicas to be active. The `is_active` function already accounts // this. .filter(|state| state.is_active()) .count() }) .minmax(); let min_max_active_replicas = match min_max_active_replicas { itertools::MinMaxResult::NoElements => None, itertools::MinMaxResult::OneElement(one) => Some((one, one)), itertools::MinMaxResult::MinMax(min, max) => Some((min, max)), }; if let Some((min, max)) = min_max_active_replicas { total_min_active_replicas = total_min_active_replicas.min(min); total_max_active_replicas = total_max_active_replicas.max(max); } points_per_collection.push(gauge( collection.count_points() as f64, &[("id", &collection.id)], )); for (vec_name, count) in collection.count_points_per_vector() { vector_count_by_name.push(gauge( count as f64, &[("collection", &collection.id), ("vector", &vec_name)], )) } let points_excluded_from_index_only = collection .shards .iter() .flatten() .filter_map(|shard| shard.local.as_ref()) .filter_map(|local| local.indexed_only_excluded_vectors.as_ref()) .flatten() .fold( HashMap::<&str, usize>::default(), |mut acc, (name, vector_size)| { *acc.entry(name).or_insert(0) += vector_size; acc }, ); for (name, vector_size) in points_excluded_from_index_only { indexed_only_excluded.push(gauge( vector_size as f64, &[("id", &collection.id), ("vector", name)], )) } total_dead_replicas += collection .shards .iter() .flatten() .filter(|i| i.replicate_states.values().any(|state| !state.is_active())) .count(); // Shard Transfers let mut incoming_transfers = 0; let mut outgoing_transfers = 0; if let Some(this_peer_id) = peer_id { for transfer in collection.transfers.iter().flatten() { if transfer.to == this_peer_id { incoming_transfers += 1; } if transfer.from == this_peer_id { outgoing_transfers += 1; } } } shard_transfers_in.push(gauge( f64::from(incoming_transfers), &[("id", &collection.id)], )); shard_transfers_out.push(gauge( f64::from(outgoing_transfers), &[("id", &collection.id)], )); } for snapshot_telemetry in self.snapshots.iter().flatten() { let id = &snapshot_telemetry.id; snapshots_recovery_running.push(gauge( snapshot_telemetry .running_snapshot_recovery .unwrap_or_default() as f64, &[("id", id)], )); snapshots_creation_running.push(gauge( snapshot_telemetry.running_snapshots.unwrap_or_default() as f64, &[("id", id)], )); snapshots_created_total.push(counter( snapshot_telemetry .total_snapshot_creations .unwrap_or_default() as f64, &[("id", id)], )); } let vector_count = vector_count_by_name .iter() .map(|m| m.get_gauge().get_value()) .sum::<f64>() // The sum of an empty f64 iterator returns `-0`. Since a negative // number of vectors is impossible, taking the absolute value is always safe. .abs(); metrics.push_metric(metric_family( "collections_vector_total", "total number of vectors in all collections", MetricType::GAUGE, vec![gauge(vector_count, &[])], prefix, )); metrics.push_metric(metric_family( "collection_vectors", "amount of vectors grouped by vector name", MetricType::GAUGE, vector_count_by_name, prefix, )); metrics.push_metric(metric_family( "collection_indexed_only_excluded_points", "amount of points excluded in indexed_only requests", MetricType::GAUGE, indexed_only_excluded, prefix, )); let total_min_active_replicas = if total_min_active_replicas == usize::MAX { 0 } else { total_min_active_replicas }; metrics.push_metric(metric_family( "collection_active_replicas_min", "minimum number of active replicas across all shards", MetricType::GAUGE, vec![gauge(total_min_active_replicas as f64, &[])], prefix, )); metrics.push_metric(metric_family( "collection_active_replicas_max", "maximum number of active replicas across all shards", MetricType::GAUGE, vec![gauge(total_max_active_replicas as f64, &[])], prefix, )); metrics.push_metric(metric_family( "collection_running_optimizations", "number of currently running optimization tasks per collection", MetricType::GAUGE, total_optimizations_running, prefix, )); metrics.push_metric(metric_family( "collection_points", "approximate amount of points per collection", MetricType::GAUGE, points_per_collection, prefix, )); metrics.push_metric(metric_family( "collection_dead_replicas", "total amount of shard replicas in non-active state", MetricType::GAUGE, vec![gauge(total_dead_replicas as f64, &[])], prefix, )); metrics.push_metric(metric_family( "snapshot_creation_running", "amount of snapshot creations that are currently running", MetricType::GAUGE, snapshots_creation_running, prefix, )); metrics.push_metric(metric_family( "snapshot_recovery_running", "amount of snapshot recovery operations currently running", MetricType::GAUGE, snapshots_recovery_running, prefix, )); metrics.push_metric(metric_family( "snapshot_created_total", "total amount of snapshots created", MetricType::COUNTER, snapshots_created_total, prefix, )); metrics.push_metric(metric_family( "collection_shard_transfer_incoming", "incoming shard transfers currently running", MetricType::GAUGE, shard_transfers_in, prefix, )); metrics.push_metric(metric_family( "collection_shard_transfer_outgoing", "outgoing shard transfers currently running", MetricType::GAUGE, shard_transfers_out, prefix, )); } } impl MetricsProvider for ClusterTelemetry { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { let ClusterTelemetry { enabled, status, config: _, peers: _, peer_metadata: _, metadata: _, } = self; metrics.push_metric(metric_family( "cluster_enabled", "is cluster support enabled", MetricType::GAUGE, vec![gauge(if *enabled { 1.0 } else { 0.0 }, &[])], prefix, )); if let Some(status) = status { status.add_metrics(metrics, prefix); } } } impl MetricsProvider for ClusterStatusTelemetry { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { metrics.push_metric(metric_family( "cluster_peers_total", "total number of cluster peers", MetricType::GAUGE, vec![gauge(self.number_of_peers as f64, &[])], prefix, )); metrics.push_metric(metric_family( "cluster_term", "current cluster term", MetricType::COUNTER, vec![counter(self.term as f64, &[])], prefix, )); if let Some(ref peer_id) = self.peer_id.map(|p| p.to_string()) { metrics.push_metric(metric_family( "cluster_commit", "index of last committed (finalized) operation cluster peer is aware of", MetricType::COUNTER, vec![counter(self.commit as f64, &[("peer_id", peer_id)])], prefix, )); metrics.push_metric(metric_family( "cluster_pending_operations_total", "total number of pending operations for cluster peer", MetricType::GAUGE, vec![gauge(self.pending_operations as f64, &[])], prefix, )); metrics.push_metric(metric_family( "cluster_voter", "is cluster peer a voter or learner", MetricType::GAUGE, vec![gauge(if self.is_voter { 1.0 } else { 0.0 }, &[])], prefix, )); } // Initialize all states so that every state has a zeroed metric by default. let mut state_working = 0.0; let mut state_stopped = 0.0; match &self.consensus_thread_status { ConsensusThreadStatus::Working { last_update } => { let timestamp = last_update.timestamp(); metrics.push_metric(metric_family( "cluster_last_update_timestamp_seconds", "unix timestamp of last update", MetricType::COUNTER, vec![counter(timestamp as f64, &[])], prefix, )); state_working = 1.0; } ConsensusThreadStatus::Stopped | ConsensusThreadStatus::StoppedWithErr { err: _ } => { state_stopped = 1.0; } } let working_states = vec![ gauge(state_working, &[("state", "working")]), gauge(state_stopped, &[("state", "stopped")]), ]; metrics.push_metric(metric_family( "cluster_working_state", "working state of the cluster", MetricType::GAUGE, working_states, prefix, )); } } impl MetricsProvider for RequestsTelemetry { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { self.rest.add_metrics(metrics, prefix); self.grpc.add_metrics(metrics, prefix); } } impl MetricsProvider for WebApiTelemetry { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { let mut builder = OperationDurationMetricsBuilder::default(); for (endpoint, responses) in &self.responses { let Some((method, endpoint)) = endpoint.split_once(' ') else { continue; }; // Endpoint must be whitelisted if REST_ENDPOINT_WHITELIST.binary_search(&endpoint).is_err() { continue; } for (status, stats) in responses { builder.add( stats, &[ ("method", method), ("endpoint", endpoint), ("status", &status.to_string()), ], *status == REST_TIMINGS_FOR_STATUS, ); } } builder.build(prefix, "rest", metrics); } } impl MetricsProvider for GrpcTelemetry { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { let mut builder = OperationDurationMetricsBuilder::default(); for (endpoint, stats) in &self.responses { // Endpoint must be whitelisted if GRPC_ENDPOINT_WHITELIST .binary_search(&endpoint.as_str()) .is_err() { continue; } builder.add(stats, &[("endpoint", endpoint.as_str())], true); } builder.build(prefix, "grpc", metrics); } } impl MetricsProvider for MemoryTelemetry { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { metrics.push_metric(metric_family( "memory_active_bytes", "Total number of bytes in active pages allocated by the application", MetricType::GAUGE, vec![gauge(self.active_bytes as f64, &[])], prefix, )); metrics.push_metric(metric_family( "memory_allocated_bytes", "Total number of bytes allocated by the application", MetricType::GAUGE, vec![gauge(self.allocated_bytes as f64, &[])], prefix, )); metrics.push_metric(metric_family( "memory_metadata_bytes", "Total number of bytes dedicated to metadata", MetricType::GAUGE, vec![gauge(self.metadata_bytes as f64, &[])], prefix, )); metrics.push_metric(metric_family( "memory_resident_bytes", "Maximum number of bytes in physically resident data pages mapped", MetricType::GAUGE, vec![gauge(self.resident_bytes as f64, &[])], prefix, )); metrics.push_metric(metric_family( "memory_retained_bytes", "Total number of bytes in virtual memory mappings", MetricType::GAUGE, vec![gauge(self.retained_bytes as f64, &[])], prefix, )); } } impl HardwareTelemetry { // Helper function to create counter metrics of a single Hw type, like cpu. fn make_metric_counters<F: Fn(&HardwareUsage) -> usize>(&self, f: F) -> Vec<Metric> { self.collection_data .iter() .map(|(collection_id, hw_usage)| counter(f(hw_usage) as f64, &[("id", collection_id)])) .collect() } } impl MetricsProvider for HardwareTelemetry { fn add_metrics(&self, metrics: &mut MetricsData, prefix: Option<&str>) { // MetricType::COUNTER requires non-empty collection data. if self.collection_data.is_empty() { return; } // Keep a dummy type decomposition of HwUsage here to enforce coverage of new fields in metrics. // This gets optimized away by the compiler: https://godbolt.org/z/9cMTzcYr4 let HardwareUsage { cpu: _, payload_io_read: _, payload_io_write: _, payload_index_io_read: _, payload_index_io_write: _, vector_io_read: _, vector_io_write: _, } = HardwareUsage::default(); metrics.push_metric(metric_family( "collection_hardware_metric_cpu", "CPU measurements of a collection", MetricType::COUNTER, self.make_metric_counters(|hw| hw.cpu), prefix, )); metrics.push_metric(metric_family( "collection_hardware_metric_payload_io_read", "Total IO payload read metrics of a collection", MetricType::COUNTER, self.make_metric_counters(|hw| hw.payload_io_read), prefix, )); metrics.push_metric(metric_family( "collection_hardware_metric_payload_index_io_read", "Total IO payload index read metrics of a collection", MetricType::COUNTER, self.make_metric_counters(|hw| hw.payload_index_io_read), prefix, )); metrics.push_metric(metric_family( "collection_hardware_metric_payload_index_io_write", "Total IO payload index write metrics of a collection", MetricType::COUNTER, self.make_metric_counters(|hw| hw.payload_index_io_write), prefix, )); metrics.push_metric(metric_family( "collection_hardware_metric_payload_io_write", "Total IO payload write metrics of a collection", MetricType::COUNTER, self.make_metric_counters(|hw| hw.payload_io_write), prefix, )); metrics.push_metric(metric_family( "collection_hardware_metric_vector_io_read", "Total IO vector read metrics of a collection", MetricType::COUNTER, self.make_metric_counters(|hw| hw.vector_io_read), prefix, )); metrics.push_metric(metric_family( "collection_hardware_metric_vector_io_write", "Total IO vector write metrics of a collection", MetricType::COUNTER, self.make_metric_counters(|hw| hw.vector_io_write), prefix, )); } } /// A helper struct to build a vector of [`MetricFamily`] out of a collection of /// [`OperationDurationStatistics`]. #[derive(Default)] struct OperationDurationMetricsBuilder { total: Vec<Metric>, fail_total: Vec<Metric>, avg_secs: Vec<Metric>, min_secs: Vec<Metric>, max_secs: Vec<Metric>, duration_histogram_secs: Vec<Metric>, } impl OperationDurationMetricsBuilder { /// Add metrics for the provided statistics. /// If `add_timings` is `false`, only the total and fail_total counters will be added. pub fn add( &mut self, stat: &OperationDurationStatistics, labels: &[(&str, &str)], add_timings: bool, ) { self.total.push(counter(stat.count as f64, labels)); self.fail_total .push(counter(stat.fail_count.unwrap_or_default() as f64, labels)); if !add_timings { return; } self.avg_secs.push(gauge( f64::from(stat.avg_duration_micros.unwrap_or(0.0)) / 1_000_000.0, labels, )); self.min_secs.push(gauge( f64::from(stat.min_duration_micros.unwrap_or(0.0)) / 1_000_000.0, labels, )); self.max_secs.push(gauge( f64::from(stat.max_duration_micros.unwrap_or(0.0)) / 1_000_000.0, labels, )); self.duration_histogram_secs.push(histogram( stat.count as u64, stat.total_duration_micros.unwrap_or(0) as f64 / 1_000_000.0, &stat .duration_micros_histogram .iter() .map(|&(b, c)| (f64::from(b) / 1_000_000.0, c as u64)) .collect::<Vec<_>>(), labels, )); } /// Build metrics and add them to the provided vector. pub fn build(self, global_prefix: Option<&str>, prefix: &str, metrics: &mut MetricsData) { let prefix = format!("{}{prefix}_", global_prefix.unwrap_or("")); metrics.push_metric(metric_family( "responses_total", "total number of responses", MetricType::COUNTER, self.total, Some(&prefix), )); metrics.push_metric(metric_family( "responses_fail_total", "total number of failed responses", MetricType::COUNTER, self.fail_total, Some(&prefix), )); metrics.push_metric(metric_family( "responses_avg_duration_seconds", "average response duration", MetricType::GAUGE, self.avg_secs, Some(&prefix), )); metrics.push_metric(metric_family( "responses_min_duration_seconds", "minimum response duration", MetricType::GAUGE, self.min_secs, Some(&prefix), )); metrics.push_metric(metric_family( "responses_max_duration_seconds", "maximum response duration", MetricType::GAUGE, self.max_secs, Some(&prefix), )); metrics.push_metric(metric_family( "responses_duration_seconds", "response duration histogram", MetricType::HISTOGRAM, self.duration_histogram_secs, Some(&prefix), )); } } fn metric_family( name: &str, help: &str, r#type: MetricType, metrics: Vec<Metric>, prefix: Option<&str>, ) -> Option<MetricFamily> { // We can't create a new `MetricsFamily` without metrics. if metrics.is_empty() { return None; } let mut metric_family = MetricFamily::default(); let name_with_prefix = prefix .map(|prefix| format!("{prefix}{name}")) .unwrap_or_else(|| name.to_string()); metric_family.set_name(name_with_prefix); metric_family.set_help(help.into()); metric_family.set_field_type(r#type); metric_family.set_metric(metrics);
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
true
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/debugger.rs
src/common/debugger.rs
use std::sync::Arc; use parking_lot::Mutex; use schemars::JsonSchema; use serde::{Deserialize, Serialize}; use crate::common::pyroscope_state::pyro::PyroscopeState; use crate::settings::Settings; #[derive(Serialize, JsonSchema, Debug, Deserialize, Clone)] pub struct PyroscopeConfig { pub url: String, pub identifier: String, pub user: Option<String>, pub password: Option<String>, pub sampling_rate: Option<u32>, } #[derive(Default, Debug, Serialize, JsonSchema, Deserialize, Clone)] pub struct DebuggerConfig { pub pyroscope: Option<PyroscopeConfig>, } #[derive(Debug, Serialize, JsonSchema, Deserialize, Clone)] #[serde(rename_all = "snake_case")] pub enum DebugConfigPatch { Pyroscope(Option<PyroscopeConfig>), } pub struct DebuggerState { #[cfg_attr(not(target_os = "linux"), allow(dead_code))] pub pyroscope: Arc<Mutex<Option<PyroscopeState>>>, } impl DebuggerState { pub fn from_settings(settings: &Settings) -> Self { let pyroscope_config = settings.debugger.pyroscope.clone(); Self { pyroscope: Arc::new(Mutex::new(PyroscopeState::from_config(pyroscope_config))), } } #[cfg_attr(not(target_os = "linux"), allow(clippy::unused_self))] pub fn get_config(&self) -> DebuggerConfig { let pyroscope_config = { #[cfg(target_os = "linux")] { let pyroscope_state_guard = self.pyroscope.lock(); pyroscope_state_guard.as_ref().map(|s| s.config.clone()) } #[cfg(not(target_os = "linux"))] { None } }; DebuggerConfig { pyroscope: pyroscope_config, } } #[cfg_attr(not(target_os = "linux"), allow(clippy::unused_self))] pub fn apply_config_patch(&self, patch: DebugConfigPatch) -> bool { #[cfg(target_os = "linux")] { match patch { DebugConfigPatch::Pyroscope(new_config) => { let mut pyroscope_guard = self.pyroscope.lock(); if let Some(pyroscope_state) = pyroscope_guard.as_mut() { let stopped = pyroscope_state.stop_agent(); if !stopped { return false; } } if let Some(new_config) = new_config { *pyroscope_guard = PyroscopeState::from_config(Some(new_config)); } true } } } #[cfg(not(target_os = "linux"))] { let _ = patch; // Ignore new_config on non-linux OS false } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/strict_mode.rs
src/common/strict_mode.rs
use std::sync::Arc; use collection::operations::verification::StrictModeVerification; use storage::content_manager::collection_verification::{ check_strict_mode, check_strict_mode_batch, check_strict_mode_toc, check_strict_mode_toc_batch, }; use storage::content_manager::errors::StorageError; use storage::content_manager::toc::TableOfContent; use storage::dispatcher::Dispatcher; use storage::rbac::Access; /// Trait for different ways of providing something with `toc` that may do additional checks eg. for Strict mode. pub trait CheckedTocProvider { async fn check_strict_mode( &self, request: &impl StrictModeVerification, collection_name: &str, timeout: Option<usize>, access: &Access, ) -> Result<&Arc<TableOfContent>, StorageError>; async fn check_strict_mode_batch<I, R>( &self, requests: &[I], conv: impl Fn(&I) -> &R, collection_name: &str, timeout: Option<usize>, access: &Access, ) -> Result<&Arc<TableOfContent>, StorageError> where R: StrictModeVerification; } /// Simple provider for TableOfContent that doesn't do any checks. #[derive(Clone)] pub struct UncheckedTocProvider<'a> { toc: &'a Arc<TableOfContent>, } impl<'a> UncheckedTocProvider<'a> { pub fn new_unchecked(toc: &'a Arc<TableOfContent>) -> Self { Self { toc } } } impl CheckedTocProvider for UncheckedTocProvider<'_> { async fn check_strict_mode( &self, _request: &impl StrictModeVerification, _collection_name: &str, _timeout: Option<usize>, _access: &Access, ) -> Result<&Arc<TableOfContent>, StorageError> { // No checks here Ok(self.toc) } async fn check_strict_mode_batch<I, R>( &self, _requests: &[I], _conv: impl Fn(&I) -> &R, _collection_name: &str, _timeout: Option<usize>, _access: &Access, ) -> Result<&Arc<TableOfContent>, StorageError> where R: StrictModeVerification, { // No checks here Ok(self.toc) } } /// Provider for TableOfContent that requires Strict mode to be checked. #[derive(Clone)] pub struct StrictModeCheckedTocProvider<'a> { dispatcher: &'a Dispatcher, } impl<'a> StrictModeCheckedTocProvider<'a> { pub fn new(dispatcher: &'a Dispatcher) -> Self { Self { dispatcher } } } impl CheckedTocProvider for StrictModeCheckedTocProvider<'_> { async fn check_strict_mode( &self, request: &impl StrictModeVerification, collection_name: &str, timeout: Option<usize>, access: &Access, ) -> Result<&Arc<TableOfContent>, StorageError> { let pass = check_strict_mode(request, timeout, collection_name, self.dispatcher, access).await?; Ok(self.dispatcher.toc(access, &pass)) } async fn check_strict_mode_batch<I, R>( &self, requests: &[I], conv: impl Fn(&I) -> &R, collection_name: &str, timeout: Option<usize>, access: &Access, ) -> Result<&Arc<TableOfContent>, StorageError> where R: StrictModeVerification, { let pass = check_strict_mode_batch( requests.iter().map(conv), timeout, collection_name, self.dispatcher, access, ) .await?; Ok(self.dispatcher.toc(access, &pass)) } } /// TableOfContent "Provider" for internal API. The `CheckedTocProvider` usually is designed to provide TOC by checking strictmode. /// However this is not possible in the internal API as it requires `Dispatcher` which we don't have in this case. /// /// Note: Only use this if you only have access to `TableOfContent` and not `Dispatcher`! pub(crate) struct StrictModeCheckedInternalTocProvider<'a> { toc: &'a Arc<TableOfContent>, } impl<'a> StrictModeCheckedInternalTocProvider<'a> { pub fn new(toc: &'a Arc<TableOfContent>) -> Self { Self { toc } } } impl CheckedTocProvider for StrictModeCheckedInternalTocProvider<'_> { async fn check_strict_mode( &self, request: &impl StrictModeVerification, collection_name: &str, timeout: Option<usize>, access: &Access, ) -> Result<&Arc<TableOfContent>, StorageError> { check_strict_mode_toc(request, timeout, collection_name, self.toc, access).await?; Ok(self.toc) } async fn check_strict_mode_batch<I, R>( &self, requests: &[I], conv: impl Fn(&I) -> &R, collection_name: &str, timeout: Option<usize>, access: &Access, ) -> Result<&Arc<TableOfContent>, StorageError> where R: StrictModeVerification, { check_strict_mode_toc_batch( requests.iter().map(conv), timeout, collection_name, self.toc, access, ) .await?; Ok(self.toc) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/auth/jwt_parser.rs
src/common/auth/jwt_parser.rs
use jsonwebtoken::errors::ErrorKind; use jsonwebtoken::{Algorithm, DecodingKey, Validation, decode}; use validator::Validate; use super::AuthError; use super::claims::Claims; #[derive(Clone)] pub struct JwtParser { key: DecodingKey, validation: Validation, } impl JwtParser { const ALGORITHM: Algorithm = Algorithm::HS256; pub fn new(secret: &str) -> Self { let key = DecodingKey::from_secret(secret.as_bytes()); let mut validation = Validation::new(Self::ALGORITHM); // Qdrant server is the only audience validation.validate_aud = false; // Expiration time leeway to account for clock skew validation.leeway = 30; // All claims are optional validation.required_spec_claims = Default::default(); JwtParser { key, validation } } /// Decode the token and return the claims, this already validates the `exp` claim with some leeway. /// Returns None when the token doesn't look like a JWT. pub fn decode(&self, token: &str) -> Option<Result<Claims, AuthError>> { let claims = match decode::<Claims>(token, &self.key, &self.validation) { Ok(token_data) => token_data.claims, Err(e) => { return match e.kind() { ErrorKind::ExpiredSignature | ErrorKind::InvalidSignature => { Some(Err(AuthError::Forbidden(e.to_string()))) } _ => None, }; } }; if let Err(e) = claims.validate() { return Some(Err(AuthError::Unauthorized(e.to_string()))); } Some(Ok(claims)) } } #[cfg(test)] mod tests { use serde_json::json; use storage::rbac::{ Access, CollectionAccess, CollectionAccessList, CollectionAccessMode, GlobalAccessMode, }; use super::*; pub fn create_token(claims: &Claims) -> String { use jsonwebtoken::{EncodingKey, Header, encode}; let key = EncodingKey::from_secret("secret".as_ref()); let header = Header::new(JwtParser::ALGORITHM); encode(&header, claims, &key).unwrap() } #[test] fn test_jwt_parser() { let exp = std::time::SystemTime::now() .duration_since(std::time::UNIX_EPOCH) .expect("Time went backwards") .as_secs(); let claims = Claims { sub: None, exp: Some(exp), access: Access::Collection(CollectionAccessList(vec![CollectionAccess { collection: "collection".to_string(), access: CollectionAccessMode::ReadWrite, #[expect(deprecated)] payload: None, }])), value_exists: None, }; let token = create_token(&claims); let secret = "secret"; let parser = JwtParser::new(secret); let decoded_claims = parser.decode(&token).unwrap().unwrap(); assert_eq!(claims, decoded_claims); } #[test] fn test_jwt_parser_with_deprecated_payloads() { let exp = std::time::SystemTime::now() .duration_since(std::time::UNIX_EPOCH) .expect("Time went backwards") .as_secs(); let claims = Claims { sub: None, exp: Some(exp), access: Access::Collection(CollectionAccessList(vec![CollectionAccess { collection: "collection".to_string(), access: CollectionAccessMode::ReadWrite, #[expect(deprecated)] payload: Some(json!({ "field1": "value", "field2": 42, "field3": true, })), }])), value_exists: None, }; let token = create_token(&claims); let secret = "secret"; let parser = JwtParser::new(secret); assert!(parser.decode(&token).unwrap().is_err()); // Validation should fail due to PayloadConstraint } #[test] fn test_exp_validation() { let exp = std::time::SystemTime::now() .duration_since(std::time::UNIX_EPOCH) .expect("Time went backwards") .as_secs() - 31; // 31 seconds in the past, bigger than the 30 seconds leeway let mut claims = Claims { sub: None, exp: Some(exp), access: Access::Global(GlobalAccessMode::Read), value_exists: None, }; let token = create_token(&claims); let secret = "secret"; let parser = JwtParser::new(secret); assert!(matches!( parser.decode(&token), Some(Err(AuthError::Forbidden(_))) )); // Remove the exp claim and it should work claims.exp = None; let token = create_token(&claims); let decoded_claims = parser.decode(&token).unwrap().unwrap(); assert_eq!(claims, decoded_claims); } #[test] fn test_invalid_token() { let claims = Claims { sub: None, exp: None, access: Access::Global(GlobalAccessMode::Read), value_exists: None, }; let token = create_token(&claims); assert!(matches!( JwtParser::new("wrong-secret").decode(&token), Some(Err(AuthError::Forbidden(_))) )); assert!(JwtParser::new("secret").decode("foo.bar.baz").is_none()); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/auth/mod.rs
src/common/auth/mod.rs
use std::sync::Arc; use collection::operations::shard_selector_internal::ShardSelectorInternal; use collection::operations::types::ScrollRequestInternal; use common::counter::hardware_accumulator::HwMeasurementAcc; use segment::types::{WithPayloadInterface, WithVector}; use storage::content_manager::errors::StorageError; use storage::content_manager::toc::TableOfContent; use storage::rbac::Access; use self::claims::{Claims, ValueExists}; use self::jwt_parser::JwtParser; use super::strings::ct_eq; use crate::common::inference::token::InferenceToken; use crate::settings::ServiceConfig; pub mod claims; pub mod jwt_parser; pub const HTTP_HEADER_API_KEY: &str = "api-key"; /// The API keys used for auth #[derive(Clone)] pub struct AuthKeys { /// A key allowing Read or Write operations read_write: Option<String>, /// A key allowing Read operations read_only: Option<String>, /// A JWT parser, based on the read_write key jwt_parser: Option<JwtParser>, /// Table of content, needed to do stateful validation of JWT toc: Arc<TableOfContent>, } #[derive(Debug)] pub enum AuthError { Unauthorized(String), Forbidden(String), StorageError(StorageError), } impl AuthKeys { fn get_jwt_parser(service_config: &ServiceConfig) -> Option<JwtParser> { if service_config.jwt_rbac.unwrap_or_default() { service_config .api_key .as_ref() .map(|secret| JwtParser::new(secret)) } else { None } } /// Defines the auth scheme given the service config /// /// Returns None if no scheme is specified. pub fn try_create(service_config: &ServiceConfig, toc: Arc<TableOfContent>) -> Option<Self> { match ( service_config.api_key.clone(), service_config.read_only_api_key.clone(), ) { (None, None) => None, (read_write, read_only) => Some(Self { read_write, read_only, jwt_parser: Self::get_jwt_parser(service_config), toc, }), } } /// Validate that the specified request is allowed for given keys. pub async fn validate_request<'a>( &self, get_header: impl Fn(&'a str) -> Option<&'a str>, ) -> Result<(Access, InferenceToken), AuthError> { let Some(key) = get_header(HTTP_HEADER_API_KEY) .or_else(|| get_header("authorization").and_then(|v| v.strip_prefix("Bearer "))) else { return Err(AuthError::Unauthorized( "Must provide an API key or an Authorization bearer token".to_string(), )); }; if self.can_write(key) { return Ok(( Access::full("Read-write access by key"), InferenceToken(None), )); } if self.can_read(key) { return Ok(( Access::full_ro("Read-only access by key"), InferenceToken(None), )); } if let Some(claims) = self.jwt_parser.as_ref().and_then(|p| p.decode(key)) { let Claims { sub, exp: _, // already validated on decoding access, value_exists, } = claims?; if let Some(value_exists) = value_exists { self.validate_value_exists(&value_exists).await?; } return Ok((access, InferenceToken(sub))); } Err(AuthError::Unauthorized( "Invalid API key or JWT".to_string(), )) } async fn validate_value_exists(&self, value_exists: &ValueExists) -> Result<(), AuthError> { let scroll_req = ScrollRequestInternal { offset: None, limit: Some(1), filter: Some(value_exists.to_filter()), with_payload: Some(WithPayloadInterface::Bool(false)), with_vector: WithVector::Bool(false), order_by: None, }; let res = self .toc .scroll( value_exists.get_collection(), scroll_req, None, None, // no timeout ShardSelectorInternal::All, Access::full("JWT stateful validation"), HwMeasurementAcc::disposable(), ) .await .map_err(|e| match e { StorageError::NotFound { .. } => { AuthError::Forbidden("Invalid JWT, stateful validation failed".to_string()) } _ => AuthError::StorageError(e), })?; if res.points.is_empty() { return Err(AuthError::Unauthorized( "Invalid JWT, stateful validation failed".to_string(), )); }; Ok(()) } /// Check if a key is allowed to read #[inline] fn can_read(&self, key: &str) -> bool { self.read_only .as_ref() .is_some_and(|ro_key| ct_eq(ro_key, key)) } /// Check if a key is allowed to write #[inline] fn can_write(&self, key: &str) -> bool { self.read_write .as_ref() .is_some_and(|rw_key| ct_eq(rw_key, key)) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/auth/claims.rs
src/common/auth/claims.rs
use segment::json_path::JsonPath; use segment::types::{Condition, FieldCondition, Filter, Match, ValueVariants}; use serde::{Deserialize, Serialize}; use storage::rbac::Access; use validator::{Validate, ValidationErrors}; #[derive(Serialize, Deserialize, PartialEq, Clone, Debug)] pub struct Claims { /// The subject ID; can be a subscription ID, cluster ID, or user ID pub sub: Option<String>, /// Expiration time (seconds since UNIX epoch) pub exp: Option<u64>, #[serde(default = "default_access")] pub access: Access, /// Validate this token by looking for a value inside a collection. pub value_exists: Option<ValueExists>, } #[derive(Serialize, Deserialize, PartialEq, Clone, Debug)] pub struct KeyValuePair { key: JsonPath, value: ValueVariants, } impl KeyValuePair { pub fn to_condition(&self) -> Condition { Condition::Field(FieldCondition::new_match( self.key.clone(), Match::new_value(self.value.clone()), )) } } #[derive(Serialize, Deserialize, PartialEq, Clone, Debug)] pub struct ValueExists { collection: String, matches: Vec<KeyValuePair>, } fn default_access() -> Access { Access::full("Give full access when the access field is not present") } impl ValueExists { pub fn get_collection(&self) -> &str { &self.collection } pub fn to_filter(&self) -> Filter { let conditions = self .matches .iter() .map(|pair| pair.to_condition()) .collect(); Filter { should: None, min_should: None, must: Some(conditions), must_not: None, } } } impl Validate for Claims { fn validate(&self) -> Result<(), ValidationErrors> { ValidationErrors::merge_all(Ok(()), "access", self.access.validate()) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/query_requests_grpc.rs
src/common/inference/query_requests_grpc.rs
use api::conversions::json::json_path_from_proto; use api::grpc::qdrant::RecommendInput; use api::grpc::qdrant::query::Variant; use api::grpc::{InferenceUsage, qdrant as grpc}; use api::rest::{self, LookupLocation, RecommendStrategy}; use collection::operations::universal_query::collection_query::{ CollectionPrefetch, CollectionQueryGroupsRequest, CollectionQueryRequest, Mmr, NearestWithMmr, Query, VectorInputInternal, VectorQuery, }; use collection::operations::universal_query::formula::FormulaInternal; use collection::operations::universal_query::shard_query::{FusionInternal, SampleInternal}; use ordered_float::OrderedFloat; use segment::data_types::order_by::OrderBy; use segment::data_types::vectors::{DEFAULT_VECTOR_NAME, MultiDenseVectorInternal, VectorInternal}; use segment::types::{Filter, PointIdType, SearchParams}; use segment::vector_storage::query::{ContextPair, ContextQuery, DiscoveryQuery, RecoQuery}; use tonic::Status; use crate::common::inference::batch_processing_grpc::{ BatchAccumGrpc, collect_prefetch, collect_query, }; use crate::common::inference::infer_processing::BatchAccumInferred; use crate::common::inference::params::InferenceParams; use crate::common::inference::service::{InferenceData, InferenceType}; /// ToDo: this function is supposed to call an inference endpoint internally pub async fn convert_query_point_groups_from_grpc( query: grpc::QueryPointGroups, inference_params: InferenceParams, ) -> Result<(CollectionQueryGroupsRequest, InferenceUsage), Status> { let grpc::QueryPointGroups { collection_name: _, prefetch, query, using, filter, params, score_threshold, with_payload, with_vectors, lookup_from, limit, group_size, group_by, with_lookup, read_consistency: _, timeout: _, shard_key_selector: _, } = query; let mut batch = BatchAccumGrpc::new(); if let Some(q) = &query { collect_query(q, &mut batch)?; } for p in &prefetch { collect_prefetch(p, &mut batch)?; } let BatchAccumGrpc { objects } = batch; let (inferred, usage) = BatchAccumInferred::from_objects(objects, InferenceType::Search, inference_params) .await .map_err(|e| Status::internal(format!("Inference error: {e}")))?; let query = if let Some(q) = query { Some(convert_query_with_inferred(q, &inferred)?) } else { None }; let prefetch = prefetch .into_iter() .map(|p| convert_prefetch_with_inferred(p, &inferred)) .collect::<Result<Vec<_>, _>>()?; let request = CollectionQueryGroupsRequest { prefetch, query, using: using.unwrap_or_else(|| DEFAULT_VECTOR_NAME.to_owned()), filter: filter.map(TryFrom::try_from).transpose()?, score_threshold, with_vector: with_vectors .map(From::from) .unwrap_or(CollectionQueryRequest::DEFAULT_WITH_VECTOR), with_payload: with_payload .map(TryFrom::try_from) .transpose()? .unwrap_or(CollectionQueryRequest::DEFAULT_WITH_PAYLOAD), lookup_from: lookup_from.map(LookupLocation::try_from).transpose()?, group_by: json_path_from_proto(&group_by)?, group_size: group_size .map(|s| s as usize) .unwrap_or(CollectionQueryRequest::DEFAULT_GROUP_SIZE), limit: limit .map(|l| l as usize) .unwrap_or(CollectionQueryRequest::DEFAULT_LIMIT), params: params.map(From::from), with_lookup: with_lookup.map(TryFrom::try_from).transpose()?, }; Ok((request, usage.unwrap_or_default().into())) } /// ToDo: this function is supposed to call an inference endpoint internally pub async fn convert_query_points_from_grpc( query: grpc::QueryPoints, inference_params: InferenceParams, ) -> Result<(CollectionQueryRequest, InferenceUsage), Status> { let grpc::QueryPoints { collection_name: _, prefetch, query, using, filter, params, score_threshold, limit, offset, with_payload, with_vectors, read_consistency: _, shard_key_selector: _, lookup_from, timeout: _, } = query; let mut batch = BatchAccumGrpc::new(); if let Some(q) = &query { collect_query(q, &mut batch)?; } for p in &prefetch { collect_prefetch(p, &mut batch)?; } let BatchAccumGrpc { objects } = batch; let (inferred, usage) = BatchAccumInferred::from_objects(objects, InferenceType::Search, inference_params) .await .map_err(|e| Status::internal(format!("Inference error: {e}")))?; let prefetch = prefetch .into_iter() .map(|p| convert_prefetch_with_inferred(p, &inferred)) .collect::<Result<Vec<_>, _>>()?; let query = query .map(|q| convert_query_with_inferred(q, &inferred)) .transpose()?; Ok(( CollectionQueryRequest { prefetch, query, using: using.unwrap_or_else(|| DEFAULT_VECTOR_NAME.to_owned()), filter: filter.map(TryFrom::try_from).transpose()?, score_threshold, limit: limit .map(|l| l as usize) .unwrap_or(CollectionQueryRequest::DEFAULT_LIMIT), offset: offset .map(|o| o as usize) .unwrap_or(CollectionQueryRequest::DEFAULT_OFFSET), params: params.map(From::from), with_vector: with_vectors .map(From::from) .unwrap_or(CollectionQueryRequest::DEFAULT_WITH_VECTOR), with_payload: with_payload .map(TryFrom::try_from) .transpose()? .unwrap_or(CollectionQueryRequest::DEFAULT_WITH_PAYLOAD), lookup_from: lookup_from.map(LookupLocation::try_from).transpose()?, }, usage.unwrap_or_default().into(), )) } fn convert_prefetch_with_inferred( prefetch: grpc::PrefetchQuery, inferred: &BatchAccumInferred, ) -> Result<CollectionPrefetch, Status> { let grpc::PrefetchQuery { prefetch, query, using, filter, params, score_threshold, limit, lookup_from, } = prefetch; let nested_prefetches = prefetch .into_iter() .map(|p| convert_prefetch_with_inferred(p, inferred)) .collect::<Result<Vec<_>, _>>()?; let query = query .map(|q| convert_query_with_inferred(q, inferred)) .transpose()?; Ok(CollectionPrefetch { prefetch: nested_prefetches, query, using: using.unwrap_or_else(|| DEFAULT_VECTOR_NAME.to_owned()), filter: filter.map(Filter::try_from).transpose()?, score_threshold: score_threshold.map(OrderedFloat), limit: limit .map(|l| l as usize) .unwrap_or(CollectionQueryRequest::DEFAULT_LIMIT), params: params.map(SearchParams::from), lookup_from: lookup_from.map(LookupLocation::try_from).transpose()?, }) } fn convert_query_with_inferred( query: grpc::Query, inferred: &BatchAccumInferred, ) -> Result<Query, Status> { let variant = query .variant .ok_or_else(|| Status::invalid_argument("Query variant is missing"))?; let query = match variant { Variant::Nearest(nearest) => { let vector = convert_vector_input_with_inferred(nearest, inferred)?; Query::Vector(VectorQuery::Nearest(vector)) } Variant::Recommend(recommend) => { let RecommendInput { positive, negative, strategy, } = recommend; let positives = positive .into_iter() .map(|v| convert_vector_input_with_inferred(v, inferred)) .collect::<Result<Vec<_>, _>>()?; let negatives = negative .into_iter() .map(|v| convert_vector_input_with_inferred(v, inferred)) .collect::<Result<Vec<_>, _>>()?; let reco_query = RecoQuery::new(positives, negatives); let strategy = strategy .and_then(|x| grpc::RecommendStrategy::try_from(x).ok()) .map(RecommendStrategy::from) .unwrap_or_default(); match strategy { RecommendStrategy::AverageVector => { Query::Vector(VectorQuery::RecommendAverageVector(reco_query)) } RecommendStrategy::BestScore => { Query::Vector(VectorQuery::RecommendBestScore(reco_query)) } RecommendStrategy::SumScores => { Query::Vector(VectorQuery::RecommendSumScores(reco_query)) } } } Variant::Discover(discover) => { let grpc::DiscoverInput { target, context } = discover; let target = target .map(|t| convert_vector_input_with_inferred(t, inferred)) .transpose()? .ok_or_else(|| Status::invalid_argument("DiscoverInput target is missing"))?; let grpc::ContextInput { pairs } = context .ok_or_else(|| Status::invalid_argument("DiscoverInput context is missing"))?; let context = pairs .into_iter() .map(|pair| context_pair_from_grpc_with_inferred(pair, inferred)) .collect::<Result<_, _>>()?; Query::Vector(VectorQuery::Discover(DiscoveryQuery::new(target, context))) } Variant::Context(context) => { let context_query = context_query_from_grpc_with_inferred(context, inferred)?; Query::Vector(VectorQuery::Context(context_query)) } Variant::OrderBy(order_by) => Query::OrderBy(OrderBy::try_from(order_by)?), Variant::Fusion(fusion) => Query::Fusion(FusionInternal::try_from(fusion)?), Variant::Rrf(rrf) => Query::Fusion(FusionInternal::try_from(rrf)?), Variant::Formula(formula) => Query::Formula(FormulaInternal::try_from(formula)?), Variant::Sample(sample) => Query::Sample(SampleInternal::try_from(sample)?), Variant::NearestWithMmr(grpc::NearestInputWithMmr { nearest, mmr }) => { let nearest = nearest.ok_or_else(|| Status::invalid_argument("nearest vector is missing"))?; let nearest = convert_vector_input_with_inferred(nearest, inferred)?; let mmr = mmr.ok_or_else(|| Status::invalid_argument("mmr is missing"))?; let grpc::Mmr { diversity, candidates_limit, } = mmr; let mmr = Mmr { diversity, candidates_limit: candidates_limit.map(|x| x as usize), }; Query::Vector(VectorQuery::NearestWithMmr(NearestWithMmr { nearest, mmr })) } }; Ok(query) } fn convert_vector_input_with_inferred( vector: grpc::VectorInput, inferred: &BatchAccumInferred, ) -> Result<VectorInputInternal, Status> { use api::grpc::qdrant::vector_input::Variant; let variant = vector .variant .ok_or_else(|| Status::invalid_argument("VectorInput variant is missing"))?; match variant { Variant::Id(id) => Ok(VectorInputInternal::Id(PointIdType::try_from(id)?)), Variant::Dense(dense) => Ok(VectorInputInternal::Vector(VectorInternal::Dense( From::from(dense), ))), Variant::Sparse(sparse) => Ok(VectorInputInternal::Vector(VectorInternal::Sparse( From::from(sparse), ))), Variant::MultiDense(multi_dense) => Ok(VectorInputInternal::Vector( VectorInternal::MultiDense(MultiDenseVectorInternal::from(multi_dense)), )), Variant::Document(doc) => { let doc: rest::Document = doc .try_into() .map_err(|e| Status::internal(format!("Document conversion error: {e}")))?; let data = InferenceData::Document(doc); let vector = inferred .get_vector(&data) .ok_or_else(|| Status::internal("Missing inferred vector for document"))?; Ok(VectorInputInternal::Vector(VectorInternal::from( vector.clone(), ))) } Variant::Image(img) => { let img: rest::Image = img .try_into() .map_err(|e| Status::internal(format!("Image conversion error: {e}",)))?; let data = InferenceData::Image(img); let vector = inferred .get_vector(&data) .ok_or_else(|| Status::internal("Missing inferred vector for image"))?; Ok(VectorInputInternal::Vector(VectorInternal::from( vector.clone(), ))) } Variant::Object(obj) => { let obj: rest::InferenceObject = obj .try_into() .map_err(|e| Status::internal(format!("Object conversion error: {e}")))?; let data = InferenceData::Object(obj); let vector = inferred .get_vector(&data) .ok_or_else(|| Status::internal("Missing inferred vector for object"))?; Ok(VectorInputInternal::Vector(VectorInternal::from( vector.clone(), ))) } } } fn context_query_from_grpc_with_inferred( value: grpc::ContextInput, inferred: &BatchAccumInferred, ) -> Result<ContextQuery<VectorInputInternal>, Status> { let grpc::ContextInput { pairs } = value; Ok(ContextQuery { pairs: pairs .into_iter() .map(|pair| context_pair_from_grpc_with_inferred(pair, inferred)) .collect::<Result<_, _>>()?, }) } fn context_pair_from_grpc_with_inferred( value: grpc::ContextInputPair, inferred: &BatchAccumInferred, ) -> Result<ContextPair<VectorInputInternal>, Status> { let grpc::ContextInputPair { positive, negative } = value; let positive = positive.ok_or_else(|| Status::invalid_argument("ContextPair positive is missing"))?; let negative = negative.ok_or_else(|| Status::invalid_argument("ContextPair negative is missing"))?; Ok(ContextPair { positive: convert_vector_input_with_inferred(positive, inferred)?, negative: convert_vector_input_with_inferred(negative, inferred)?, }) } #[cfg(test)] mod tests { use std::collections::HashMap; use api::grpc::qdrant::Value; use api::grpc::qdrant::value::Kind; use api::grpc::qdrant::vector_input::Variant; use collection::operations::point_ops::VectorPersisted; use super::*; fn create_test_document() -> api::grpc::qdrant::Document { api::grpc::qdrant::Document { text: "test".to_string(), model: "test-model".to_string(), options: HashMap::new(), } } fn create_test_image() -> api::grpc::qdrant::Image { api::grpc::qdrant::Image { image: Some(Value { kind: Some(Kind::StringValue("test.jpg".to_string())), }), model: "test-model".to_string(), options: HashMap::new(), } } fn create_test_object() -> api::grpc::qdrant::InferenceObject { api::grpc::qdrant::InferenceObject { object: Some(Value { kind: Some(Kind::StringValue("test".to_string())), }), model: "test-model".to_string(), options: HashMap::new(), } } fn create_test_inferred_batch() -> BatchAccumInferred { let mut objects = HashMap::new(); let grpc_doc = create_test_document(); let grpc_img = create_test_image(); let grpc_obj = create_test_object(); let doc: rest::Document = grpc_doc.try_into().unwrap(); let img: rest::Image = grpc_img.try_into().unwrap(); let obj: rest::InferenceObject = grpc_obj.try_into().unwrap(); let doc_data = InferenceData::Document(doc); let img_data = InferenceData::Image(img); let obj_data = InferenceData::Object(obj); let dense_vector = vec![1.0, 2.0, 3.0]; let vector_persisted = VectorPersisted::Dense(dense_vector); objects.insert(doc_data, vector_persisted.clone()); objects.insert(img_data, vector_persisted.clone()); objects.insert(obj_data, vector_persisted); BatchAccumInferred { objects } } #[test] fn test_convert_vector_input_with_inferred_dense() { let inferred = create_test_inferred_batch(); let vector = grpc::VectorInput { variant: Some(Variant::Dense(grpc::DenseVector { data: vec![1.0, 2.0, 3.0], })), }; let result = convert_vector_input_with_inferred(vector, &inferred).unwrap(); match result { VectorInputInternal::Vector(VectorInternal::Dense(values)) => { assert_eq!(values, vec![1.0, 2.0, 3.0]); } _ => panic!("Expected dense vector"), } } #[test] fn test_convert_vector_input_with_inferred_document() { let inferred = create_test_inferred_batch(); let doc = create_test_document(); let vector = grpc::VectorInput { variant: Some(Variant::Document(doc)), }; let result = convert_vector_input_with_inferred(vector, &inferred).unwrap(); match result { VectorInputInternal::Vector(VectorInternal::Dense(values)) => { assert_eq!(values, vec![1.0, 2.0, 3.0]); } _ => panic!("Expected dense vector from inference"), } } #[test] fn test_convert_vector_input_missing_variant() { let inferred = create_test_inferred_batch(); let vector = grpc::VectorInput { variant: None }; let result = convert_vector_input_with_inferred(vector, &inferred); assert!(result.is_err()); assert!(result.unwrap_err().message().contains("variant is missing")); } #[test] fn test_context_pair_from_grpc_with_inferred() { let inferred = create_test_inferred_batch(); let pair = grpc::ContextInputPair { positive: Some(grpc::VectorInput { variant: Some(Variant::Dense(grpc::DenseVector { data: vec![1.0, 2.0, 3.0], })), }), negative: Some(grpc::VectorInput { variant: Some(Variant::Document(create_test_document())), }), }; let result = context_pair_from_grpc_with_inferred(pair, &inferred).unwrap(); match (result.positive, result.negative) { ( VectorInputInternal::Vector(VectorInternal::Dense(pos)), VectorInputInternal::Vector(VectorInternal::Dense(neg)), ) => { assert_eq!(pos, vec![1.0, 2.0, 3.0]); assert_eq!(neg, vec![1.0, 2.0, 3.0]); } _ => panic!("Expected dense vectors"), } } #[test] fn test_context_pair_missing_vectors() { let inferred = create_test_inferred_batch(); let pair = grpc::ContextInputPair { positive: None, negative: Some(grpc::VectorInput { variant: Some(Variant::Document(create_test_document())), }), }; let result = context_pair_from_grpc_with_inferred(pair, &inferred); assert!(result.is_err()); assert!( result .unwrap_err() .message() .contains("positive is missing"), ); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/config.rs
src/common/inference/config.rs
use serde::{Deserialize, Serialize}; #[derive(Debug, Clone, Serialize, Deserialize, Default)] pub struct InferenceConfig { pub address: Option<String>, pub timeout: Option<u64>, pub token: Option<String>, } impl InferenceConfig { pub fn new(address: Option<String>) -> Self { Self { address, timeout: None, token: None, } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/update_requests.rs
src/common/inference/update_requests.rs
use std::collections::HashMap; use api::rest::models::InferenceUsage; use api::rest::{Batch, BatchVectorStruct, PointStruct, PointVectors, Vector, VectorStruct}; use collection::operations::point_ops::{ BatchPersisted, BatchVectorStructPersisted, PointStructPersisted, VectorPersisted, VectorStructPersisted, }; use collection::operations::vector_ops::PointVectorsPersisted; use storage::content_manager::errors::StorageError; use crate::common::inference::batch_processing::BatchAccum; use crate::common::inference::infer_processing::BatchAccumInferred; use crate::common::inference::params::InferenceParams; use crate::common::inference::service::{InferenceData, InferenceType}; pub async fn convert_point_struct( point_structs: Vec<PointStruct>, inference_type: InferenceType, inference_params: InferenceParams, ) -> Result<(Vec<PointStructPersisted>, Option<InferenceUsage>), StorageError> { let mut batch_accum = BatchAccum::new(); for point_struct in &point_structs { match &point_struct.vector { VectorStruct::Named(named) => { for vector in named.values() { match vector { Vector::Document(doc) => { batch_accum.add(InferenceData::Document(doc.clone())) } Vector::Image(img) => batch_accum.add(InferenceData::Image(img.clone())), Vector::Object(obj) => batch_accum.add(InferenceData::Object(obj.clone())), Vector::Dense(_) | Vector::Sparse(_) | Vector::MultiDense(_) => {} } } } VectorStruct::Document(doc) => batch_accum.add(InferenceData::Document(doc.clone())), VectorStruct::Image(img) => batch_accum.add(InferenceData::Image(img.clone())), VectorStruct::Object(obj) => batch_accum.add(InferenceData::Object(obj.clone())), VectorStruct::MultiDense(_) | VectorStruct::Single(_) => {} } } let (inferred, usage) = if !batch_accum.objects.is_empty() { let (inferred_data, usage) = BatchAccumInferred::from_batch_accum(batch_accum, inference_type, &inference_params) .await?; (Some(inferred_data), usage) } else { (None, None) }; let mut converted_points: Vec<PointStructPersisted> = Vec::new(); for point_struct in point_structs { let PointStruct { id, vector, payload, } = point_struct; let converted_vector_struct = match vector { VectorStruct::Single(single) => VectorStructPersisted::Single(single), VectorStruct::MultiDense(multi) => VectorStructPersisted::MultiDense(multi), VectorStruct::Named(named) => { let mut named_vectors = HashMap::new(); for (name, vector_data) in named { let converted_vector = match &inferred { Some(inferred) => convert_vector_with_inferred(vector_data, inferred)?, None => match vector_data { Vector::Dense(dense) => VectorPersisted::Dense(dense), Vector::Sparse(sparse) => VectorPersisted::Sparse(sparse), Vector::MultiDense(multi) => VectorPersisted::MultiDense(multi), Vector::Document(_) | Vector::Image(_) | Vector::Object(_) => { return Err(StorageError::inference_error( "Inference required but service returned no results for named vector", )); } }, }; named_vectors.insert(name, converted_vector); } VectorStructPersisted::Named(named_vectors) } VectorStruct::Document(doc) => { let vector_data = match &inferred { Some(inferred) => { convert_vector_with_inferred(Vector::Document(doc), inferred)? } None => { return Err(StorageError::inference_error( "Inference required for document but service returned no results", )); } }; match vector_data { VectorPersisted::Dense(dense) => VectorStructPersisted::Single(dense), VectorPersisted::Sparse(_) => { return Err(StorageError::bad_request( "Sparse vector from document inference should be named", )); } VectorPersisted::MultiDense(multi) => VectorStructPersisted::MultiDense(multi), } } VectorStruct::Image(img) => { let vector_data = match &inferred { Some(inferred) => convert_vector_with_inferred(Vector::Image(img), inferred)?, None => { return Err(StorageError::inference_error( "Inference required for image but service returned no results", )); } }; match vector_data { VectorPersisted::Dense(dense) => VectorStructPersisted::Single(dense), VectorPersisted::Sparse(_) => { return Err(StorageError::bad_request( "Sparse vector from image inference should be named", )); } VectorPersisted::MultiDense(multi) => VectorStructPersisted::MultiDense(multi), } } VectorStruct::Object(obj) => { let vector_data = match &inferred { Some(inferred) => convert_vector_with_inferred(Vector::Object(obj), inferred)?, None => { return Err(StorageError::inference_error( "Inference required for object but service returned no results", )); } }; match vector_data { VectorPersisted::Dense(dense) => VectorStructPersisted::Single(dense), VectorPersisted::Sparse(_) => { return Err(StorageError::bad_request( "Sparse vector from object inference should be named", )); } VectorPersisted::MultiDense(multi) => VectorStructPersisted::MultiDense(multi), } } }; let converted = PointStructPersisted { id, vector: converted_vector_struct, payload, }; converted_points.push(converted); } Ok((converted_points, usage)) } pub async fn convert_batch( batch: Batch, inference_params: InferenceParams, ) -> Result<(BatchPersisted, Option<InferenceUsage>), StorageError> { let Batch { ids, vectors, payloads, } = batch; let mut inference_usage = InferenceUsage::default(); let batch_persisted = match vectors { BatchVectorStruct::Single(single) => BatchVectorStructPersisted::Single(single), BatchVectorStruct::MultiDense(multi) => BatchVectorStructPersisted::MultiDense(multi), BatchVectorStruct::Named(named) => { let mut named_vectors = HashMap::new(); for (name, vecs) in named { let (converted_vectors, batch_usage) = convert_vectors(vecs, InferenceType::Update, inference_params.clone()).await?; inference_usage.merge_opt(batch_usage); named_vectors.insert(name, converted_vectors); } BatchVectorStructPersisted::Named(named_vectors) } BatchVectorStruct::Document(_) => { return Err(StorageError::inference_error( "Direct Document processing is not supported in top-level batch vectors. Use named vectors.", )); } BatchVectorStruct::Image(_) => { return Err(StorageError::inference_error( "Direct Image processing is not supported in top-level batch vectors. Use named vectors.", )); } BatchVectorStruct::Object(_) => { return Err(StorageError::inference_error( "Direct Object processing is not supported in top-level batch vectors. Use named vectors.", )); } }; let batch_persisted = BatchPersisted { ids, vectors: batch_persisted, payloads, }; Ok((batch_persisted, inference_usage.into_non_empty())) } pub async fn convert_point_vectors( point_vectors_list: Vec<PointVectors>, inference_type: InferenceType, inference_params: InferenceParams, ) -> Result<(Vec<PointVectorsPersisted>, Option<InferenceUsage>), StorageError> { let mut batch_accum = BatchAccum::new(); for point_vectors in &point_vectors_list { if let VectorStruct::Named(named) = &point_vectors.vector { for vector in named.values() { match vector { Vector::Document(doc) => batch_accum.add(InferenceData::Document(doc.clone())), Vector::Image(img) => batch_accum.add(InferenceData::Image(img.clone())), Vector::Object(obj) => batch_accum.add(InferenceData::Object(obj.clone())), Vector::Dense(_) | Vector::Sparse(_) | Vector::MultiDense(_) => {} } } } } let mut inference_usage = InferenceUsage::default(); let inferred = if !batch_accum.objects.is_empty() { let (inferred_data, usage) = BatchAccumInferred::from_batch_accum(batch_accum, inference_type, &inference_params) .await?; inference_usage.merge_opt(usage); Some(inferred_data) } else { None }; let mut converted_point_vectors: Vec<PointVectorsPersisted> = Vec::new(); for point_vectors in point_vectors_list { let PointVectors { id, vector } = point_vectors; let converted_vector = match vector { VectorStruct::Single(dense) => VectorStructPersisted::Single(dense), VectorStruct::MultiDense(multi) => VectorStructPersisted::MultiDense(multi), VectorStruct::Named(named) => { let mut converted = HashMap::new(); for (name, vec) in named { let converted_vec = match &inferred { Some(inferred) => convert_vector_with_inferred(vec, inferred)?, None => match vec { Vector::Dense(dense) => VectorPersisted::Dense(dense), Vector::Sparse(sparse) => VectorPersisted::Sparse(sparse), Vector::MultiDense(multi) => VectorPersisted::MultiDense(multi), Vector::Document(_) | Vector::Image(_) | Vector::Object(_) => { return Err(StorageError::inference_error( "Inference required for named vector in PointVectors but no results", )); } }, }; converted.insert(name, converted_vec); } VectorStructPersisted::Named(converted) } VectorStruct::Document(_) => { return Err(StorageError::inference_error( "Direct Document processing not supported for PointVectors. Use named vectors.", )); } VectorStruct::Image(_) => { return Err(StorageError::inference_error( "Direct Image processing not supported for PointVectors. Use named vectors.", )); } VectorStruct::Object(_) => { return Err(StorageError::inference_error( "Direct Object processing not supported for PointVectors. Use named vectors.", )); } }; converted_point_vectors.push(PointVectorsPersisted { id, vector: converted_vector, }); } Ok((converted_point_vectors, inference_usage.into_non_empty())) } fn convert_point_struct_with_inferred( point_structs: Vec<PointStruct>, inferred: &BatchAccumInferred, ) -> Result<Vec<PointStructPersisted>, StorageError> { point_structs .into_iter() .map(|point_struct| { let PointStruct { id, vector, payload, } = point_struct; let converted_vector_struct = match vector { VectorStruct::Single(single) => VectorStructPersisted::Single(single), VectorStruct::MultiDense(multi) => VectorStructPersisted::MultiDense(multi), VectorStruct::Named(named) => { let mut named_vectors = HashMap::new(); for (name, vector_data) in named { let converted_vector = convert_vector_with_inferred(vector_data, inferred)?; named_vectors.insert(name, converted_vector); } VectorStructPersisted::Named(named_vectors) } VectorStruct::Document(doc) => { let vector_data = convert_vector_with_inferred(Vector::Document(doc), inferred)?; match vector_data { VectorPersisted::Dense(dense) => VectorStructPersisted::Single(dense), VectorPersisted::Sparse(_) => { return Err(StorageError::bad_request( "Sparse vector from document inference must be named", )); } VectorPersisted::MultiDense(multi) => { VectorStructPersisted::MultiDense(multi) } } } VectorStruct::Image(img) => { let vector_data = convert_vector_with_inferred(Vector::Image(img), inferred)?; match vector_data { VectorPersisted::Dense(dense) => VectorStructPersisted::Single(dense), VectorPersisted::Sparse(_) => { return Err(StorageError::bad_request( "Sparse vector from image inference must be named", )); } VectorPersisted::MultiDense(multi) => { VectorStructPersisted::MultiDense(multi) } } } VectorStruct::Object(obj) => { let vector_data = convert_vector_with_inferred(Vector::Object(obj), inferred)?; match vector_data { VectorPersisted::Dense(dense) => VectorStructPersisted::Single(dense), VectorPersisted::Sparse(_) => { return Err(StorageError::bad_request( "Sparse vector from object inference must be named", )); } VectorPersisted::MultiDense(multi) => { VectorStructPersisted::MultiDense(multi) } } } }; Ok(PointStructPersisted { id, vector: converted_vector_struct, payload, }) }) .collect() } pub async fn convert_vectors( vectors: Vec<Vector>, inference_type: InferenceType, inference_params: InferenceParams, ) -> Result<(Vec<VectorPersisted>, Option<InferenceUsage>), StorageError> { let mut batch_accum = BatchAccum::new(); for vector in &vectors { match vector { Vector::Document(doc) => batch_accum.add(InferenceData::Document(doc.clone())), Vector::Image(img) => batch_accum.add(InferenceData::Image(img.clone())), Vector::Object(obj) => batch_accum.add(InferenceData::Object(obj.clone())), Vector::Dense(_) | Vector::Sparse(_) | Vector::MultiDense(_) => {} } } let mut inference_usage = InferenceUsage::default(); let inferred = if !batch_accum.objects.is_empty() { let (inferred_data, usage) = BatchAccumInferred::from_batch_accum(batch_accum, inference_type, &inference_params) .await?; inference_usage.merge_opt(usage); Some(inferred_data) } else { None }; let converted_vectors: Result<Vec<VectorPersisted>, StorageError> = vectors .into_iter() .map(|vector_data| match &inferred { Some(inferred) => convert_vector_with_inferred(vector_data, inferred), None => match vector_data { Vector::Dense(dense) => Ok(VectorPersisted::Dense(dense)), Vector::Sparse(sparse) => Ok(VectorPersisted::Sparse(sparse)), Vector::MultiDense(multi) => Ok(VectorPersisted::MultiDense(multi)), Vector::Document(_) | Vector::Image(_) | Vector::Object(_) => { Err(StorageError::inference_error( "Inference required but no inference service results available", )) } }, }) .collect(); converted_vectors.map(|vecs| (vecs, inference_usage.into_non_empty())) } fn convert_vector_with_inferred( vector: Vector, inferred: &BatchAccumInferred, ) -> Result<VectorPersisted, StorageError> { match vector { Vector::Dense(dense) => Ok(VectorPersisted::Dense(dense)), Vector::Sparse(sparse) => Ok(VectorPersisted::Sparse(sparse)), Vector::MultiDense(multi) => Ok(VectorPersisted::MultiDense(multi)), Vector::Document(doc) => { let data = InferenceData::Document(doc); inferred.get_vector(&data).cloned().ok_or_else(|| { StorageError::inference_error("Missing inferred vector for document") }) } Vector::Image(img) => { let data = InferenceData::Image(img); inferred .get_vector(&data) .cloned() .ok_or_else(|| StorageError::inference_error("Missing inferred vector for image")) } Vector::Object(obj) => { let data = InferenceData::Object(obj); inferred .get_vector(&data) .cloned() .ok_or_else(|| StorageError::inference_error("Missing inferred vector for object")) } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/batch_processing.rs
src/common/inference/batch_processing.rs
use std::collections::HashSet; use api::rest::{ ContextInput, ContextPair, DiscoverInput, Prefetch, Query, QueryGroupsRequestInternal, QueryInterface, QueryRequestInternal, RecommendInput, VectorInput, }; use super::service::{InferenceData, InferenceInput, InferenceRequest}; pub struct BatchAccum { pub(crate) objects: HashSet<InferenceData>, } impl BatchAccum { pub fn new() -> Self { Self { objects: HashSet::new(), } } pub fn add(&mut self, data: InferenceData) { self.objects.insert(data); } pub fn extend(&mut self, other: BatchAccum) { self.objects.extend(other.objects); } pub fn is_empty(&self) -> bool { self.objects.is_empty() } } impl From<&BatchAccum> for InferenceRequest { fn from(batch: &BatchAccum) -> Self { Self { inputs: batch .objects .iter() .cloned() .map(InferenceInput::from) .collect(), inference: None, token: None, } } } fn collect_vector_input(vector: &VectorInput, batch: &mut BatchAccum) { match vector { VectorInput::Document(doc) => batch.add(InferenceData::Document(doc.clone())), VectorInput::Image(img) => batch.add(InferenceData::Image(img.clone())), VectorInput::Object(obj) => batch.add(InferenceData::Object(obj.clone())), // types that are not supported in the Inference Service VectorInput::DenseVector(_) => {} VectorInput::SparseVector(_) => {} VectorInput::MultiDenseVector(_) => {} VectorInput::Id(_) => {} } } fn collect_context_pair(pair: &ContextPair, batch: &mut BatchAccum) { collect_vector_input(&pair.positive, batch); collect_vector_input(&pair.negative, batch); } fn collect_discover_input(discover: &DiscoverInput, batch: &mut BatchAccum) { collect_vector_input(&discover.target, batch); if let Some(context) = &discover.context { for pair in context { collect_context_pair(pair, batch); } } } fn collect_recommend_input(recommend: &RecommendInput, batch: &mut BatchAccum) { if let Some(positive) = &recommend.positive { for vector in positive { collect_vector_input(vector, batch); } } if let Some(negative) = &recommend.negative { for vector in negative { collect_vector_input(vector, batch); } } } fn collect_query(query: &Query, batch: &mut BatchAccum) { match query { Query::Nearest(nearest) => collect_vector_input(&nearest.nearest, batch), Query::Recommend(recommend) => collect_recommend_input(&recommend.recommend, batch), Query::Discover(discover) => collect_discover_input(&discover.discover, batch), Query::Context(context) => { if let ContextInput(Some(pairs)) = &context.context { for pair in pairs { collect_context_pair(pair, batch); } } } Query::OrderBy(_) | Query::Fusion(_) | Query::Rrf(_) | Query::Formula(_) | Query::Sample(_) => {} } } fn collect_query_interface(query: &QueryInterface, batch: &mut BatchAccum) { match query { QueryInterface::Nearest(vector) => collect_vector_input(vector, batch), QueryInterface::Query(query) => collect_query(query, batch), } } fn collect_prefetch(prefetch: &Prefetch, batch: &mut BatchAccum) { let Prefetch { prefetch, query, using: _, filter: _, params: _, score_threshold: _, limit: _, lookup_from: _, } = prefetch; if let Some(query) = query { collect_query_interface(query, batch); } if let Some(prefetches) = prefetch { for p in prefetches { collect_prefetch(p, batch); } } } pub fn collect_query_groups_request(request: &QueryGroupsRequestInternal) -> BatchAccum { let mut batch = BatchAccum::new(); let QueryGroupsRequestInternal { query, prefetch, using: _, filter: _, params: _, score_threshold: _, with_vector: _, with_payload: _, lookup_from: _, group_request: _, } = request; if let Some(query) = query { collect_query_interface(query, &mut batch); } if let Some(prefetches) = prefetch { for prefetch in prefetches { collect_prefetch(prefetch, &mut batch); } } batch } pub fn collect_query_request(request: &QueryRequestInternal) -> BatchAccum { let mut batch = BatchAccum::new(); let QueryRequestInternal { prefetch, query, using: _, filter: _, score_threshold: _, params: _, limit: _, offset: _, with_vector: _, with_payload: _, lookup_from: _, } = request; if let Some(query) = query { collect_query_interface(query, &mut batch); } if let Some(prefetches) = prefetch { for prefetch in prefetches { collect_prefetch(prefetch, &mut batch); } } batch } #[cfg(test)] mod tests { use api::rest::QueryBaseGroupRequest; use api::rest::schema::{DiscoverQuery, Document, Image, InferenceObject, NearestQuery}; use serde_json::json; use super::*; fn create_test_document(text: &str) -> Document { Document { text: text.to_string(), model: "test-model".to_string(), options: Default::default(), } } fn create_test_image(url: &str) -> Image { Image { image: json!({"data": url.to_string()}), model: "test-model".to_string(), options: Default::default(), } } fn create_test_object(data: &str) -> InferenceObject { InferenceObject { object: json!({"data": data}), model: "test-model".to_string(), options: Default::default(), } } #[test] fn test_batch_accum_basic() { let mut batch = BatchAccum::new(); assert!(batch.objects.is_empty()); let doc = InferenceData::Document(create_test_document("test")); batch.add(doc.clone()); assert_eq!(batch.objects.len(), 1); batch.add(doc); assert_eq!(batch.objects.len(), 1); } #[test] fn test_batch_accum_extend() { let mut batch1 = BatchAccum::new(); let mut batch2 = BatchAccum::new(); let doc1 = InferenceData::Document(create_test_document("test1")); let doc2 = InferenceData::Document(create_test_document("test2")); batch1.add(doc1); batch2.add(doc2); batch1.extend(batch2); assert_eq!(batch1.objects.len(), 2); } #[test] fn test_deduplication() { let mut batch = BatchAccum::new(); let doc1 = InferenceData::Document(create_test_document("same")); let doc2 = InferenceData::Document(create_test_document("same")); batch.add(doc1); batch.add(doc2); assert_eq!(batch.objects.len(), 1); } #[test] fn test_collect_vector_input() { let mut batch = BatchAccum::new(); let doc_input = VectorInput::Document(create_test_document("test")); let img_input = VectorInput::Image(create_test_image("test.jpg")); let obj_input = VectorInput::Object(create_test_object("test")); collect_vector_input(&doc_input, &mut batch); collect_vector_input(&img_input, &mut batch); collect_vector_input(&obj_input, &mut batch); assert_eq!(batch.objects.len(), 3); } #[test] fn test_collect_prefetch() { let prefetch = Prefetch { query: Some(QueryInterface::Nearest(VectorInput::Document( create_test_document("test"), ))), prefetch: Some(vec![Prefetch { query: Some(QueryInterface::Nearest(VectorInput::Image( create_test_image("nested.jpg"), ))), prefetch: None, using: None, filter: None, params: None, score_threshold: None, limit: None, lookup_from: None, }]), using: None, filter: None, params: None, score_threshold: None, limit: None, lookup_from: None, }; let mut batch = BatchAccum::new(); collect_prefetch(&prefetch, &mut batch); assert_eq!(batch.objects.len(), 2); } #[test] fn test_collect_query_groups_request() { let request = QueryGroupsRequestInternal { query: Some(QueryInterface::Query(Query::Nearest(NearestQuery { nearest: VectorInput::Document(create_test_document("test")), mmr: None, }))), prefetch: Some(vec![Prefetch { query: Some(QueryInterface::Query(Query::Discover(DiscoverQuery { discover: DiscoverInput { target: VectorInput::Image(create_test_image("test.jpg")), context: Some(vec![ContextPair { positive: VectorInput::Document(create_test_document("pos")), negative: VectorInput::Image(create_test_image("neg.jpg")), }]), }, }))), prefetch: None, using: None, filter: None, params: None, score_threshold: None, limit: None, lookup_from: None, }]), using: None, filter: None, params: None, score_threshold: None, with_vector: None, with_payload: None, lookup_from: None, group_request: QueryBaseGroupRequest { group_by: "test".parse().unwrap(), group_size: None, limit: None, with_lookup: None, }, }; let batch = collect_query_groups_request(&request); assert_eq!(batch.objects.len(), 4); } #[test] fn test_different_model_same_content() { let mut batch = BatchAccum::new(); let mut doc1 = create_test_document("same"); let mut doc2 = create_test_document("same"); doc1.model = "model1".to_string(); doc2.model = "model2".to_string(); batch.add(InferenceData::Document(doc1)); batch.add(InferenceData::Document(doc2)); assert_eq!(batch.objects.len(), 2); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/local_model.rs
src/common/inference/local_model.rs
use collection::operations::point_ops::VectorPersisted; use storage::content_manager::errors::StorageError; use super::bm25::Bm25; use super::service::{InferenceInput, InferenceType}; use crate::common::inference::inference_input::InferenceDataType; enum LocalModelName { Bm25, } impl LocalModelName { fn from_str(model_name: &str) -> Option<Self> { match model_name.to_lowercase().as_str() { "qdrant/bm25" => Some(LocalModelName::Bm25), "bm25" => Some(LocalModelName::Bm25), _ => None, } } } /// Run inference with only local models. /// /// # Panics /// Panics if one inference input did not target a local model. pub fn infer_local( inference_inputs: Vec<InferenceInput>, inference_type: InferenceType, ) -> Result<Vec<VectorPersisted>, StorageError> { let mut out = Vec::with_capacity(inference_inputs.len()); for input in inference_inputs { let InferenceInput { data, data_type, model, options, } = input; let Some(model_name) = LocalModelName::from_str(&model) else { unreachable!( "Non local model has been passed to infer_local(). This can happen if a newly added model wasn't added to infer_local()" ) }; // Validate it is text match data_type { InferenceDataType::Text => {} InferenceDataType::Image | InferenceDataType::Object => { return Err(StorageError::bad_input(format!( "Only text input is supported for {model}." ))); } }; let input_str = data.as_str().ok_or_else(|| { StorageError::bad_input(format!("Only text input is supported for {model}.")) })?; let embedding = match model_name { LocalModelName::Bm25 => { let bm25_config = InferenceInput::parse_bm25_config(options)?; let bm25 = Bm25::new(bm25_config); match inference_type { InferenceType::Update => bm25.doc_embed(input_str), InferenceType::Search => bm25.search_embed(input_str), } } }; out.push(embedding); } Ok(out) } /// Returns `true` if the provided `model_name` targets a local model. Local models /// are models that are handled by Qdrant and are not forwarded to a remote inference service. pub fn is_local_model(model_name: &str) -> bool { let model_name = LocalModelName::from_str(model_name); model_name.is_some() }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/inference_input.rs
src/common/inference/inference_input.rs
use std::collections::HashMap; use api::rest::{Bm25Config, Document, DocumentOptions, Image, InferenceObject}; use serde::de::IntoDeserializer; use serde::{Deserialize, Serialize}; use serde_json::Value; use storage::content_manager::errors::StorageError; use super::service::InferenceData; #[derive(Debug, Serialize, Clone)] pub struct InferenceInput { pub data: Value, pub data_type: InferenceDataType, pub model: String, pub options: Option<HashMap<String, Value>>, } impl InferenceInput { /// Attempts to parse the input's options into a local model config. pub fn parse_bm25_config( options: Option<HashMap<String, Value>>, ) -> Result<Bm25Config, StorageError> { let options = options.unwrap_or_default(); Bm25Config::deserialize(options.into_deserializer()) .map_err(|err| StorageError::bad_input(format!("Invalid BM25 config: {err:#?}"))) } } #[derive(Debug, Serialize, Clone)] #[serde(rename_all = "snake_case")] pub enum InferenceDataType { Text, Image, Object, } impl From<InferenceData> for InferenceInput { fn from(value: InferenceData) -> Self { match value { InferenceData::Document(doc) => { let Document { text, model, options, } = doc; InferenceInput { data: Value::String(text), data_type: InferenceDataType::Text, model, options: options.map(DocumentOptions::into_options), } } InferenceData::Image(img) => { let Image { image, model, options, } = img; InferenceInput { data: image, data_type: InferenceDataType::Image, model, options: options.options, } } InferenceData::Object(obj) => { let InferenceObject { object, model, options, } = obj; InferenceInput { data: object, data_type: InferenceDataType::Object, model, options: options.options, } } } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/query_requests_rest.rs
src/common/inference/query_requests_rest.rs
use api::rest::models::InferenceUsage; use api::rest::schema as rest; use collection::lookup::WithLookup; use collection::operations::universal_query::collection_query::{ CollectionPrefetch, CollectionQueryGroupsRequest, CollectionQueryRequest, Mmr, NearestWithMmr, Query, VectorInputInternal, VectorQuery, }; use collection::operations::universal_query::formula::FormulaInternal; use collection::operations::universal_query::shard_query::{FusionInternal, SampleInternal}; use ordered_float::OrderedFloat; use segment::data_types::order_by::OrderBy; use segment::data_types::vectors::{DEFAULT_VECTOR_NAME, MultiDenseVectorInternal, VectorInternal}; use segment::vector_storage::query::{ContextPair, ContextQuery, DiscoveryQuery, RecoQuery}; use storage::content_manager::errors::StorageError; use crate::common::inference::batch_processing::{ collect_query_groups_request, collect_query_request, }; use crate::common::inference::infer_processing::BatchAccumInferred; use crate::common::inference::params::InferenceParams; use crate::common::inference::service::{InferenceData, InferenceType}; pub struct CollectionQueryRequestWithUsage { pub request: CollectionQueryRequest, pub usage: Option<InferenceUsage>, } pub struct CollectionQueryGroupsRequestWithUsage { pub request: CollectionQueryGroupsRequest, pub usage: Option<InferenceUsage>, } pub async fn convert_query_groups_request_from_rest( request: rest::QueryGroupsRequestInternal, inference_params: InferenceParams, ) -> Result<CollectionQueryGroupsRequestWithUsage, StorageError> { let batch = collect_query_groups_request(&request); let rest::QueryGroupsRequestInternal { prefetch, query, using, filter, score_threshold, params, with_vector, with_payload, lookup_from, group_request, } = request; let (inferred, usage) = BatchAccumInferred::from_batch_accum(batch, InferenceType::Search, &inference_params) .await?; let query = query .map(|q| convert_query_with_inferred(q, &inferred)) .transpose()?; let prefetch = prefetch .map(|prefetches| { prefetches .into_iter() .map(|p| convert_prefetch_with_inferred(p, &inferred)) .collect::<Result<Vec<_>, _>>() }) .transpose()? .unwrap_or_default(); let collection_query_groups_request = CollectionQueryGroupsRequest { prefetch, query, using: using.unwrap_or_else(|| DEFAULT_VECTOR_NAME.to_owned()), filter, score_threshold, params, with_vector: with_vector.unwrap_or(CollectionQueryRequest::DEFAULT_WITH_VECTOR), with_payload: with_payload.unwrap_or(CollectionQueryRequest::DEFAULT_WITH_PAYLOAD), lookup_from, limit: group_request .limit .unwrap_or(CollectionQueryRequest::DEFAULT_LIMIT), group_by: group_request.group_by, group_size: group_request .group_size .unwrap_or(CollectionQueryRequest::DEFAULT_GROUP_SIZE), with_lookup: group_request.with_lookup.map(WithLookup::from), }; Ok(CollectionQueryGroupsRequestWithUsage { request: collection_query_groups_request, usage, }) } pub async fn convert_query_request_from_rest( request: rest::QueryRequestInternal, inference_params: &InferenceParams, ) -> Result<CollectionQueryRequestWithUsage, StorageError> { let batch = collect_query_request(&request); let (inferred, usage) = BatchAccumInferred::from_batch_accum(batch, InferenceType::Search, inference_params) .await?; let rest::QueryRequestInternal { prefetch, query, using, filter, score_threshold, params, limit, offset, with_vector, with_payload, lookup_from, } = request; let prefetch = prefetch .map(|prefetches| { prefetches .into_iter() .map(|p| convert_prefetch_with_inferred(p, &inferred)) .collect::<Result<Vec<_>, _>>() }) .transpose()? .unwrap_or_default(); let query = query .map(|q| convert_query_with_inferred(q, &inferred)) .transpose()?; let collection_query_request = CollectionQueryRequest { prefetch, query, using: using.unwrap_or_else(|| DEFAULT_VECTOR_NAME.to_owned()), filter, score_threshold, limit: limit.unwrap_or(CollectionQueryRequest::DEFAULT_LIMIT), offset: offset.unwrap_or(CollectionQueryRequest::DEFAULT_OFFSET), params, with_vector: with_vector.unwrap_or(CollectionQueryRequest::DEFAULT_WITH_VECTOR), with_payload: with_payload.unwrap_or(CollectionQueryRequest::DEFAULT_WITH_PAYLOAD), lookup_from, }; Ok(CollectionQueryRequestWithUsage { request: collection_query_request, usage, }) } fn convert_vector_input_with_inferred( vector: rest::VectorInput, inferred: &BatchAccumInferred, ) -> Result<VectorInputInternal, StorageError> { match vector { rest::VectorInput::Id(id) => Ok(VectorInputInternal::Id(id)), rest::VectorInput::DenseVector(dense) => { Ok(VectorInputInternal::Vector(VectorInternal::Dense(dense))) } rest::VectorInput::SparseVector(sparse) => { Ok(VectorInputInternal::Vector(VectorInternal::Sparse(sparse))) } rest::VectorInput::MultiDenseVector(multi_dense) => Ok(VectorInputInternal::Vector( VectorInternal::MultiDense(MultiDenseVectorInternal::new_unchecked(multi_dense)), )), rest::VectorInput::Document(doc) => { let data = InferenceData::Document(doc); let vector = inferred.get_vector(&data).ok_or_else(|| { StorageError::inference_error("Missing inferred vector for document") })?; Ok(VectorInputInternal::Vector(VectorInternal::from( vector.clone(), ))) } rest::VectorInput::Image(img) => { let data = InferenceData::Image(img); let vector = inferred.get_vector(&data).ok_or_else(|| { StorageError::inference_error("Missing inferred vector for image") })?; Ok(VectorInputInternal::Vector(VectorInternal::from( vector.clone(), ))) } rest::VectorInput::Object(obj) => { let data = InferenceData::Object(obj); let vector = inferred.get_vector(&data).ok_or_else(|| { StorageError::inference_error("Missing inferred vector for object") })?; Ok(VectorInputInternal::Vector(VectorInternal::from( vector.clone(), ))) } } } fn convert_query_with_inferred( query: rest::QueryInterface, inferred: &BatchAccumInferred, ) -> Result<Query, StorageError> { let query = rest::Query::from(query); match query { rest::Query::Nearest(rest::NearestQuery { nearest, mmr }) => { let vector = convert_vector_input_with_inferred(nearest, inferred)?; if let Some(mmr) = mmr { let mmr = Mmr { diversity: mmr.diversity, candidates_limit: mmr.candidates_limit, }; Ok(Query::Vector(VectorQuery::NearestWithMmr(NearestWithMmr { nearest: vector, mmr, }))) } else { Ok(Query::Vector(VectorQuery::Nearest(vector))) } } rest::Query::Recommend(recommend) => { let rest::RecommendInput { positive, negative, strategy, } = recommend.recommend; let positives = positive .into_iter() .flatten() .map(|v| convert_vector_input_with_inferred(v, inferred)) .collect::<Result<Vec<_>, _>>()?; let negatives = negative .into_iter() .flatten() .map(|v| convert_vector_input_with_inferred(v, inferred)) .collect::<Result<Vec<_>, _>>()?; let reco_query = RecoQuery::new(positives, negatives); match strategy.unwrap_or_default() { rest::RecommendStrategy::AverageVector => Ok(Query::Vector( VectorQuery::RecommendAverageVector(reco_query), )), rest::RecommendStrategy::BestScore => { Ok(Query::Vector(VectorQuery::RecommendBestScore(reco_query))) } rest::RecommendStrategy::SumScores => { Ok(Query::Vector(VectorQuery::RecommendSumScores(reco_query))) } } } rest::Query::Discover(discover) => { let rest::DiscoverInput { target, context } = discover.discover; let target = convert_vector_input_with_inferred(target, inferred)?; let context = context .into_iter() .flatten() .map(|pair| context_pair_from_rest_with_inferred(pair, inferred)) .collect::<Result<Vec<_>, _>>()?; Ok(Query::Vector(VectorQuery::Discover(DiscoveryQuery::new( target, context, )))) } rest::Query::Context(context) => { let rest::ContextInput(context) = context.context; let context = context .into_iter() .flatten() .map(|pair| context_pair_from_rest_with_inferred(pair, inferred)) .collect::<Result<Vec<_>, _>>()?; Ok(Query::Vector(VectorQuery::Context(ContextQuery::new( context, )))) } rest::Query::OrderBy(order_by) => Ok(Query::OrderBy(OrderBy::from(order_by.order_by))), rest::Query::Fusion(fusion) => Ok(Query::Fusion(FusionInternal::from(fusion.fusion))), rest::Query::Rrf(rrf) => Ok(Query::Fusion(FusionInternal::from(rrf.rrf))), rest::Query::Formula(formula) => Ok(Query::Formula(FormulaInternal::from(formula))), rest::Query::Sample(sample) => Ok(Query::Sample(SampleInternal::from(sample.sample))), } } fn convert_prefetch_with_inferred( prefetch: rest::Prefetch, inferred: &BatchAccumInferred, ) -> Result<CollectionPrefetch, StorageError> { let rest::Prefetch { prefetch, query, using, filter, score_threshold, params, limit, lookup_from, } = prefetch; let query = query .map(|q| convert_query_with_inferred(q, inferred)) .transpose()?; let nested_prefetches = prefetch .map(|prefetches| { prefetches .into_iter() .map(|p| convert_prefetch_with_inferred(p, inferred)) .collect::<Result<Vec<_>, _>>() }) .transpose()? .unwrap_or_default(); Ok(CollectionPrefetch { prefetch: nested_prefetches, query, using: using.unwrap_or_else(|| DEFAULT_VECTOR_NAME.to_owned()), filter, score_threshold: score_threshold.map(OrderedFloat), limit: limit.unwrap_or(CollectionQueryRequest::DEFAULT_LIMIT), params, lookup_from, }) } fn context_pair_from_rest_with_inferred( value: rest::ContextPair, inferred: &BatchAccumInferred, ) -> Result<ContextPair<VectorInputInternal>, StorageError> { let rest::ContextPair { positive, negative } = value; Ok(ContextPair { positive: convert_vector_input_with_inferred(positive, inferred)?, negative: convert_vector_input_with_inferred(negative, inferred)?, }) } #[cfg(test)] mod tests { use std::collections::HashMap; use api::rest::schema::{Document, Image, InferenceObject, NearestQuery}; use collection::operations::point_ops::VectorPersisted; use serde_json::json; use super::*; fn create_test_document(text: &str) -> Document { Document { text: text.to_string(), model: "test-model".to_string(), options: Default::default(), } } fn create_test_image(url: &str) -> Image { Image { image: json!({"data": url.to_string()}), model: "test-model".to_string(), options: Default::default(), } } fn create_test_object(data: &str) -> InferenceObject { InferenceObject { object: json!({"data": data}), model: "test-model".to_string(), options: Default::default(), } } fn create_test_inferred_batch() -> BatchAccumInferred { let mut objects = HashMap::new(); let doc = InferenceData::Document(create_test_document("test")); let img = InferenceData::Image(create_test_image("test.jpg")); let obj = InferenceData::Object(create_test_object("test")); let dense_vector = vec![1.0, 2.0, 3.0]; let vector_persisted = VectorPersisted::Dense(dense_vector); objects.insert(doc, vector_persisted.clone()); objects.insert(img, vector_persisted.clone()); objects.insert(obj, vector_persisted); BatchAccumInferred { objects } } #[test] fn test_convert_vector_input_with_inferred_dense() { let inferred = create_test_inferred_batch(); let vector = rest::VectorInput::DenseVector(vec![1.0, 2.0, 3.0]); let result = convert_vector_input_with_inferred(vector, &inferred).unwrap(); match result { VectorInputInternal::Vector(VectorInternal::Dense(values)) => { assert_eq!(values, vec![1.0, 2.0, 3.0]); } _ => panic!("Expected dense vector"), } } #[test] fn test_convert_vector_input_with_inferred_document() { let inferred = create_test_inferred_batch(); let doc = create_test_document("test"); let vector = rest::VectorInput::Document(doc); let result = convert_vector_input_with_inferred(vector, &inferred).unwrap(); match result { VectorInputInternal::Vector(VectorInternal::Dense(values)) => { assert_eq!(values, vec![1.0, 2.0, 3.0]); } _ => panic!("Expected dense vector from inference"), } } #[test] fn test_convert_vector_input_with_inferred_missing() { let inferred = create_test_inferred_batch(); let doc = create_test_document("missing"); let vector = rest::VectorInput::Document(doc); let result = convert_vector_input_with_inferred(vector, &inferred); assert!(result.is_err()); assert!( result .unwrap_err() .to_string() .contains("Missing inferred vector"), ); } #[test] fn test_context_pair_from_rest_with_inferred() { let inferred = create_test_inferred_batch(); let pair = rest::ContextPair { positive: rest::VectorInput::DenseVector(vec![1.0, 2.0, 3.0]), negative: rest::VectorInput::Document(create_test_document("test")), }; let result = context_pair_from_rest_with_inferred(pair, &inferred).unwrap(); match (result.positive, result.negative) { ( VectorInputInternal::Vector(VectorInternal::Dense(pos)), VectorInputInternal::Vector(VectorInternal::Dense(neg)), ) => { assert_eq!(pos, vec![1.0, 2.0, 3.0]); assert_eq!(neg, vec![1.0, 2.0, 3.0]); } _ => panic!("Expected dense vectors"), } } #[test] fn test_convert_query_with_inferred_nearest() { let inferred = create_test_inferred_batch(); let nearest = NearestQuery { nearest: rest::VectorInput::Document(create_test_document("test")), mmr: None, }; let query = rest::QueryInterface::Query(rest::Query::Nearest(nearest)); let result = convert_query_with_inferred(query, &inferred).unwrap(); match result { Query::Vector(VectorQuery::Nearest(vector)) => match vector { VectorInputInternal::Vector(VectorInternal::Dense(values)) => { assert_eq!(values, vec![1.0, 2.0, 3.0]); } _ => panic!("Expected dense vector"), }, _ => panic!("Expected nearest query"), } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/params.rs
src/common/inference/params.rs
use crate::common::inference::token::InferenceToken; #[derive(Debug, Clone, PartialEq, Default)] pub struct InferenceParams { pub token: InferenceToken, pub timeout: Option<std::time::Duration>, } impl InferenceParams { pub fn new(token: impl Into<InferenceToken>, timeout: Option<std::time::Duration>) -> Self { Self { token: token.into(), timeout, } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/bm25.rs
src/common/inference/bm25.rs
use std::borrow::Cow; use std::collections::{BTreeMap, HashMap}; use std::str::FromStr; use std::sync::Arc; use api::rest::{Bm25Config, TextPreprocessingConfig}; use collection::operations::point_ops::VectorPersisted; use itertools::Itertools; use murmur3::murmur3_32_of_slice; use segment::data_types::index::{Language, StopwordsInterface}; use segment::index::field_index::full_text_index::stop_words::StopwordsFilter; use segment::index::field_index::full_text_index::tokenizers::{ Stemmer, Tokenizer, TokensProcessor, }; const DEFAULT_LANGUAGE: &str = "english"; /// Bm25 implementation #[derive(Debug)] pub struct Bm25 { config: Bm25Config, tokenizer: Tokenizer, } impl Bm25 { pub fn new(mut config: Bm25Config) -> Self { let tokenizer_conf = std::mem::take(&mut config.text_preprocessing_config); let tokens_processor = new_token_processor_from_config(tokenizer_conf); let tokenizer = Tokenizer::new(config.tokenizer, tokens_processor); Self { config, tokenizer } } /// Tokenizes the `input` with the configured tokenizer options. fn tokenize<'b>(&'b self, input: &'b str) -> Vec<Cow<'b, str>> { let mut out = vec![]; self.tokenizer.tokenize_query(input, |i| out.push(i)); out } /// Embeds the given input using the Bm25 algorithm and configured options/hyperparameters. pub fn search_embed(&self, input: &str) -> VectorPersisted { let tokens = self.tokenize(input); if tokens.is_empty() { return VectorPersisted::empty_sparse(); } let indices: Vec<u32> = tokens .into_iter() .map(|token| Self::compute_token_id(&token)) .unique() .collect(); let values: Vec<f32> = vec![1.0; indices.len()]; VectorPersisted::new_sparse(indices, values) } /// Embeds the given input using the Bm25 algorithm and configured options/hyperparameters. pub fn doc_embed(&self, input: &str) -> VectorPersisted { let tokens = self.tokenize(input); if tokens.is_empty() { return VectorPersisted::empty_sparse(); } let tf_map = self.term_frequency(&tokens); let (indices, values): (Vec<u32>, Vec<f32>) = tf_map.into_iter().unzip(); VectorPersisted::new_sparse(indices, values) } fn term_frequency(&self, tokens: &[Cow<str>]) -> BTreeMap<u32, f32> { let mut tf_map = BTreeMap::new(); let doc_len = tokens.len() as f64; let mut counter: HashMap<&str, u32> = HashMap::new(); tokens .iter() .for_each(|token| *counter.entry(token.as_ref()).or_insert(0) += 1); let k = self.config.k.into_inner(); let b = self.config.b.into_inner(); let avg_len = self.config.avg_len.into_inner(); for (token, count) in &counter { let token_id = Self::compute_token_id(token); let num_occurrences = f64::from(*count); let mut tf = num_occurrences * (k + 1.0); tf /= k.mul_add(1.0 - b + b * doc_len / avg_len, num_occurrences); tf_map.insert(token_id, tf as f32); } tf_map } fn compute_token_id(token: &str) -> u32 { (murmur3_32_of_slice(token.as_bytes(), 0) as i32).unsigned_abs() } } fn new_token_processor_from_config(value: TextPreprocessingConfig) -> TokensProcessor { let TextPreprocessingConfig { language, lowercase, ascii_folding, stopwords, stemmer, min_token_len, max_token_len, } = value; let lowercase = lowercase.unwrap_or(true); let ascii_folding = ascii_folding.unwrap_or(false); let language = language.unwrap_or_else(|| DEFAULT_LANGUAGE.to_string()); let stemmer = match stemmer { None => Stemmer::try_default_from_language(&language), Some(stemmer_algorithm) => Some(Stemmer::from_algorithm(&stemmer_algorithm)), }; let stopwords_config = match stopwords { None => { // Try to create from the language Language::from_str(&language) .ok() .map(StopwordsInterface::Language) } Some(stopwords_interface) => Some(stopwords_interface), }; TokensProcessor::new( lowercase, ascii_folding, Arc::new(StopwordsFilter::new(&stopwords_config, lowercase)), stemmer, min_token_len, max_token_len, ) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/service.rs
src/common/inference/service.rs
use std::fmt::Display; use std::hash::Hash; use std::sync::Arc; use std::time::{Duration, SystemTime}; use actix_web::http::header::HttpDate; use api::rest::models::InferenceUsage; use api::rest::{Document, Image, InferenceObject}; use collection::operations::point_ops::VectorPersisted; use common::defaults::APP_USER_AGENT; use itertools::{Either, Itertools}; use parking_lot::RwLock; use reqwest::Client; use serde::{Deserialize, Serialize}; use storage::content_manager::errors::StorageError; pub use super::inference_input::InferenceInput; use super::local_model; use crate::common::inference::config::InferenceConfig; use crate::common::inference::params::InferenceParams; #[derive(Debug, Serialize, Default, Clone, Copy)] #[serde(rename_all = "lowercase")] pub enum InferenceType { #[default] Update, Search, } impl Display for InferenceType { fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { write!(f, "{}", format!("{self:?}").to_lowercase()) } } #[derive(Debug, Serialize)] pub struct InferenceRequest { pub(crate) inputs: Vec<InferenceInput>, pub(crate) inference: Option<InferenceType>, #[serde(default)] pub(crate) token: Option<String>, } #[derive(Debug, Deserialize)] #[cfg_attr(test, derive(Serialize))] pub struct InferenceResponse { pub embeddings: Vec<VectorPersisted>, pub usage: Option<InferenceUsage>, } #[derive(Debug, Serialize, Deserialize, Clone, PartialEq, Eq, Hash)] pub enum InferenceData { Document(Document), Image(Image), Object(InferenceObject), } #[derive(Debug, Deserialize)] struct InferenceError { pub error: String, } impl InferenceData { pub(crate) fn type_name(&self) -> &'static str { match self { InferenceData::Document(_) => "document", InferenceData::Image(_) => "image", InferenceData::Object(_) => "object", } } } pub struct InferenceService { pub(crate) config: InferenceConfig, pub(crate) client: Client, } static INFERENCE_SERVICE: RwLock<Option<Arc<InferenceService>>> = RwLock::new(None); /// We assume that the inference provider will handle timeouts itself, if /// not provided by the user or configured. But we need ensurance, that we don't /// wait forever for a response. static DEFAULT_INFERENCE_TIMEOUT_SECS: u64 = 10 * 60; // 10 minutes impl InferenceService { pub fn new(config: Option<InferenceConfig>) -> Self { let config = config.unwrap_or_default(); let InferenceConfig { address: _, timeout, token: _, } = &config; let timeout = timeout.unwrap_or(DEFAULT_INFERENCE_TIMEOUT_SECS); let client_builder = Client::builder() .user_agent(APP_USER_AGENT.as_str()) .timeout(Duration::from_secs(timeout)); Self { config, client: client_builder .build() .expect("Invalid timeout value for HTTP client"), } } pub fn init_global(config: Option<InferenceConfig>) -> Result<(), StorageError> { let mut inference_service = INFERENCE_SERVICE.write(); let service = Self::new(config); if !service.is_address_valid() { return Err(StorageError::service_error( "Cannot initialize InferenceService: address is required but not provided or empty in config", )); } *inference_service = Some(Arc::new(service)); Ok(()) } pub fn get_global() -> Option<Arc<InferenceService>> { INFERENCE_SERVICE.read().as_ref().cloned() } pub(crate) fn validate(&self) -> Result<(), StorageError> { if !self.is_address_valid() { return Err(StorageError::service_error( "InferenceService configuration error: address is missing or empty", )); } Ok(()) } pub async fn infer( &self, inference_inputs: Vec<InferenceInput>, inference_type: InferenceType, inference_params: InferenceParams, ) -> Result<InferenceResponse, StorageError> { let ( (local_inference_inputs, local_inference_positions), (remote_inference_inputs, remote_inference_positions), ): ((Vec<_>, Vec<_>), (Vec<_>, Vec<_>)) = inference_inputs .into_iter() // Keep track of the input's positions so we can properly merge them together later. .enumerate() .partition_map(|(pos, input)| { // Check if input is targeting a local model or the configured remote server. if local_model::is_local_model(&input.model) { Either::Left((input, pos)) } else { Either::Right((input, pos)) } }); // Run inference on local models let local_model_results = local_model::infer_local(local_inference_inputs, inference_type)?; // Early return with the local model's results if no other inference_inputs were passed. // If local models is also empty, we automatically return an empty response here. if remote_inference_inputs.is_empty() { return Ok(InferenceResponse { embeddings: local_model_results, usage: None, // No usage since everything was processed locally. }); } let remote_result = self .infer_remote(remote_inference_inputs, inference_type, inference_params) .await?; Ok(Self::merge_local_and_remote_result( local_model_results, local_inference_positions, remote_result, remote_inference_positions, )) } async fn infer_remote( &self, inference_inputs: Vec<InferenceInput>, inference_type: InferenceType, inference_params: InferenceParams, ) -> Result<InferenceResponse, StorageError> { // Assume that either: // - User doesn't have access to generating random JWT tokens (like in serverless) // - Inference server checks validity of the tokens. let InferenceParams { token: inference_token, timeout, } = inference_params; let token = inference_token.0.or_else(|| self.config.token.clone()); let Some(url) = self.config.address.as_ref() else { return Err(StorageError::service_error( "InferenceService URL not configured - please provide valid address in config", )); }; let request_body = InferenceRequest { inputs: inference_inputs, inference: Some(inference_type), token, }; let request = self.client.post(url); let request = if let Some(timeout) = timeout { request.timeout(timeout) } else { request }; let response = request.json(&request_body).send().await; let (response_body, status, retry_after) = match response { Ok(response) => { let status = response.status(); let retry_after = Self::parse_retry_after(response.headers()); match response.text().await { Ok(body) => (body, status, retry_after), Err(err) => { return Err(StorageError::service_error(format!( "Failed to read inference response body: {err}" ))); } } } Err(error) => { if let Some(status) = error.status() { (error.to_string(), status, None) } else { return Err(StorageError::service_error(format!( "Failed to send inference request: {error}" ))); } } }; Self::handle_inference_response(status, &response_body, retry_after) } fn merge_local_and_remote_result( local_results: Vec<VectorPersisted>, local_pos: Vec<usize>, remote_res: InferenceResponse, remote_pos: Vec<usize>, ) -> InferenceResponse { // Skip merging with local results if we only have inference results from remote. if local_results.is_empty() { return remote_res; } // Merge remote results and local results together in the exact same order they have been passed. let merged = merge_position_items( local_results, local_pos, remote_res.embeddings, remote_pos, ) .expect( "Expected local results and remote items being contiguous. This is an internal bug!", ); InferenceResponse { embeddings: merged, usage: remote_res.usage, // Only account for usage of remote. } } fn parse_retry_after(headers: &reqwest::header::HeaderMap) -> Option<Duration> { headers .get(reqwest::header::RETRY_AFTER) .and_then(|value| value.to_str().ok()) .and_then(|value| { // Check if the value is a valid duration in seconds if let Ok(seconds) = value.parse::<u64>() { return Some(Duration::from_secs(seconds)); } // Check if the value is a valid Date if let Ok(http_date) = value.parse::<HttpDate>() { let ts = SystemTime::from(http_date); return ts .duration_since(SystemTime::now()) .ok() .map(|d| d.max(Duration::ZERO)); } None }) } pub(crate) fn handle_inference_response( status: reqwest::StatusCode, response_body: &str, retry_after: Option<Duration>, ) -> Result<InferenceResponse, StorageError> { match status { reqwest::StatusCode::OK => { serde_json::from_str(response_body) .map_err(|e| { StorageError::service_error(format!( "Failed to parse successful inference response: {e}. Response body: {response_body}", )) }) } reqwest::StatusCode::BAD_REQUEST => { // Try to extract error description from the response body, if it is a valid JSON let parsed_body: Result<InferenceError, _> = serde_json::from_str(response_body); match parsed_body { Ok(InferenceError { error }) => { Err(StorageError::bad_request(format!( "Inference request validation failed: {error}", ))) } Err(_) => { Err(StorageError::bad_request(format!( "Invalid inference request: {response_body}", ))) } } } status @ (reqwest::StatusCode::UNAUTHORIZED | reqwest::StatusCode::FORBIDDEN) => { Err(StorageError::service_error(format!( "Authentication failed for inference service ({status}): {response_body}", ))) } status @ reqwest::StatusCode::TOO_MANY_REQUESTS => { Err(StorageError::rate_limit_exceeded( format!("Too many requests for inference service ({status}): {response_body}"), retry_after, )) } status @ (reqwest::StatusCode::INTERNAL_SERVER_ERROR | reqwest::StatusCode::SERVICE_UNAVAILABLE | reqwest::StatusCode::GATEWAY_TIMEOUT) => Err(StorageError::service_error(format!( "Inference service error ({status}): {response_body}", ))), _ => { if status.is_server_error() { Err(StorageError::service_error(format!( "Inference service error ({status}): {response_body}", ))) } else if status.is_client_error() { Err(StorageError::bad_request(format!( "Inference can't process request ({status}): {response_body}", ))) } else { Err(StorageError::service_error(format!( "Unexpected inference error ({status}): {response_body}", ))) } } } } fn is_address_valid(&self) -> bool { self.config.address.is_none() // In BM25 we don't need an address so we allow InferenceService to have an empty address. || self.config.address.as_ref().is_some_and(|i| !i.is_empty()) } } /// 2-way merge of lists with `PositionItems`. Also checks for skipped items and returns `None` in case an item is left out. fn merge_position_items<I>( left: impl IntoIterator<Item = I>, left_pos: Vec<usize>, right: impl IntoIterator<Item = I>, right_pos: Vec<usize>, ) -> Option<Vec<I>> { let left_iter = left.into_iter().zip(left_pos); let right_iter = right.into_iter().zip(right_pos); let mut i = 0; // Check that we cover all items and don't skip any. left_iter .merge_by(right_iter, |l: &(I, usize), r: &(I, usize)| l.1 < r.1) .map(|item| { if item.1 == i { i += 1; Some(item.0) } else { None } }) .collect::<Option<Vec<_>>>() } #[cfg(test)] mod test { use std::collections::HashMap; use api::rest::Bm25Config; use rand::rngs::StdRng; use rand::seq::SliceRandom; use rand::{Rng, SeedableRng}; use serde_json::{Value, json}; use super::*; use crate::common::inference::bm25::Bm25; use crate::common::inference::inference_input::InferenceDataType; const BM25_LOCAL_MODEL_NAME: &str = "bm25"; #[test] fn test_merge_position_items() { let (left, right): ((Vec<_>, Vec<_>), (Vec<_>, Vec<_>)) = (0..1000).map(|i| (i, i)).partition(|i| i.0 % 7 == 0); let merged = merge_position_items(left.0, left.1, right.0, right.1); assert_eq!(merged, Some((0..1000).collect::<Vec<_>>())); } #[test] fn test_merge_position_items_fail() { let (left, mut right): ((Vec<_>, Vec<_>), (Vec<_>, Vec<_>)) = (0..1000).map(|i| (i, i)).partition(|i| i.0 % 7 == 0); right.0.remove(5); right.1.remove(5); let merged = merge_position_items(left.0, left.1, right.0, right.1); // We were missing an item and therefore expect `None`. assert_eq!(merged, None); } #[tokio::test] async fn test_bm25_end_to_end() { let mut rng = StdRng::seed_from_u64(42); // Test without any BM25 let only_inference_inputs: Vec<_> = (0..rng.random_range(30..100)) .map(|_| make_normal_inference_input("this is some input", &mut rng)) .collect(); let res = run_inference_with_mocked_remote(only_inference_inputs.clone()).await; check_inference_response(only_inference_inputs, res); // Test with only BM25 let only_bm25_inputs: Vec<_> = (0..rng.random_range(30..100)) .map(|_| make_bm25_inference_input("this is some input")) .collect(); let res = run_inference_with_mocked_remote(only_bm25_inputs.clone()).await; check_inference_response(only_bm25_inputs, res); // Test BM25 and inference mixed. let mut inputs: Vec<InferenceInput> = vec![]; inputs.extend( (0..rng.random_range(30..100)).map(|_| make_bm25_inference_input("this is some input")), ); inputs.extend( (0..rng.random_range(30..100)) .map(|_| make_normal_inference_input("this is some input", &mut rng)), ); inputs.shuffle(&mut rng); let res = run_inference_with_mocked_remote(inputs.clone()).await; check_inference_response(inputs, res); } fn make_normal_inference_input(input: &str, rand: &mut StdRng) -> InferenceInput { let options = if rand.random_bool(0.3) { let mut opts = HashMap::default(); let value = rand.random_iter::<char>().take(10).collect::<String>(); // Test utf8 opts.insert("some-key".to_string(), Value::String(value)); Some(opts) } else { None }; InferenceInput { data: Value::String(input.to_string()), data_type: InferenceDataType::Text, model: "anyModel".to_string(), options, } } fn make_bm25_inference_input(input: &str) -> InferenceInput { let bm25_config = Bm25Config::default(); let options: HashMap<String, Value> = serde_json::from_str(&serde_json::to_string(&bm25_config).unwrap()).unwrap(); InferenceInput { data: Value::String(input.to_string()), data_type: InferenceDataType::Text, model: BM25_LOCAL_MODEL_NAME.to_string(), options: Some(options), } } fn check_inference_response(inputs: Vec<InferenceInput>, response: InferenceResponse) { assert_eq!(inputs.len(), response.embeddings.len()); for (idx, (input, response)) in inputs.into_iter().zip(response.embeddings).enumerate() { if input.model == BM25_LOCAL_MODEL_NAME { // In our test-setup, only BM25 returns sparse vectors. Normal inference is mocked // and always returns dense vectors. assert!(matches!(response, VectorPersisted::Sparse(..))); let bm25_config = InferenceInput::parse_bm25_config(input.options).unwrap(); // Re-run bm25 and check that response is correct. let bm25 = Bm25::new(bm25_config).doc_embed(input.data.as_str().unwrap()); assert_eq!(response, bm25); } else { let expected_vector = VectorPersisted::Dense(vec![0.0; idx]); assert_eq!(response, expected_vector); } } } async fn run_inference_with_mocked_remote( inference_inputs: Vec<InferenceInput>, ) -> InferenceResponse { // Request a new server from the pool let mut server = mockito::Server::new_async().await; // Create dummy dense vectors for non-bm25 inputs with the length of the index. // The dummy dense vector have the dimension of the position they appeared in `inference_inputs`, // so we can easily check for correct ordering later, although it is a bit hacky. let expected_embeddings: Vec<_> = inference_inputs .iter() .enumerate() .filter(|(_, item)| item.model != BM25_LOCAL_MODEL_NAME) .map(|(index, _)| { let values = vec![0.0; index]; VectorPersisted::Dense(values) }) .collect(); // Create an HTTP mock let mock = server .mock("POST", "/") .with_status(200) .with_header("content-type", "text/json") .with_body( json!(InferenceResponse { embeddings: expected_embeddings, usage: None, }) .to_string(), ) .create_async() .await; let config = InferenceConfig { address: Some(server.url()), // Use mock's URL as address when doing inference. timeout: None, token: Some(String::default()), }; let service = InferenceService::new(Some(config)); let has_remote_inference_items = inference_inputs .iter() .any(|i| i.model != BM25_LOCAL_MODEL_NAME); let res = service .infer( inference_inputs, InferenceType::Update, InferenceParams::new("key", None), ) .await .expect("Failed to do inference"); // We expect exactly 1 request if there is any inference (non-bm25) request // and 0 if all inputs are bm25. if has_remote_inference_items { mock.expect(1).assert_async().await; } else { mock.expect(0).assert_async().await; } res } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/batch_processing_grpc.rs
src/common/inference/batch_processing_grpc.rs
use std::collections::HashSet; use api::grpc::qdrant::vector_input::Variant; use api::grpc::qdrant::{ ContextInput, ContextInputPair, DiscoverInput, PrefetchQuery, Query, RecommendInput, VectorInput, query, }; use api::rest::schema as rest; use tonic::Status; use super::service::{InferenceData, InferenceInput, InferenceRequest}; pub struct BatchAccumGrpc { pub(crate) objects: HashSet<InferenceData>, } impl BatchAccumGrpc { pub fn new() -> Self { Self { objects: HashSet::new(), } } pub fn add(&mut self, data: InferenceData) { self.objects.insert(data); } pub fn extend(&mut self, other: BatchAccumGrpc) { self.objects.extend(other.objects); } pub fn is_empty(&self) -> bool { self.objects.is_empty() } } impl From<&BatchAccumGrpc> for InferenceRequest { fn from(batch: &BatchAccumGrpc) -> Self { Self { inputs: batch .objects .iter() .cloned() .map(InferenceInput::from) .collect(), inference: None, token: None, } } } fn collect_vector_input(vector: &VectorInput, batch: &mut BatchAccumGrpc) -> Result<(), Status> { let Some(variant) = &vector.variant else { return Ok(()); }; match variant { Variant::Id(_) => {} Variant::Dense(_) => {} Variant::Sparse(_) => {} Variant::MultiDense(_) => {} Variant::Document(document) => { let doc = rest::Document::try_from(document.clone()) .map_err(|e| Status::internal(format!("Document conversion error: {e:?}")))?; batch.add(InferenceData::Document(doc)); } Variant::Image(image) => { let img = rest::Image::try_from(image.clone()) .map_err(|e| Status::internal(format!("Image conversion error: {e:?}")))?; batch.add(InferenceData::Image(img)); } Variant::Object(object) => { let obj = rest::InferenceObject::try_from(object.clone()) .map_err(|e| Status::internal(format!("Object conversion error: {e:?}")))?; batch.add(InferenceData::Object(obj)); } } Ok(()) } pub(crate) fn collect_context_input( context: &ContextInput, batch: &mut BatchAccumGrpc, ) -> Result<(), Status> { let ContextInput { pairs } = context; for pair in pairs { collect_context_input_pair(pair, batch)?; } Ok(()) } fn collect_context_input_pair( pair: &ContextInputPair, batch: &mut BatchAccumGrpc, ) -> Result<(), Status> { let ContextInputPair { positive, negative } = pair; if let Some(positive) = positive { collect_vector_input(positive, batch)?; } if let Some(negative) = negative { collect_vector_input(negative, batch)?; } Ok(()) } pub(crate) fn collect_discover_input( discover: &DiscoverInput, batch: &mut BatchAccumGrpc, ) -> Result<(), Status> { let DiscoverInput { target, context } = discover; if let Some(vector) = target { collect_vector_input(vector, batch)?; } if let Some(context) = context { for pair in &context.pairs { collect_context_input_pair(pair, batch)?; } } Ok(()) } pub(crate) fn collect_recommend_input( recommend: &RecommendInput, batch: &mut BatchAccumGrpc, ) -> Result<(), Status> { let RecommendInput { positive, negative, strategy: _, } = recommend; for vector in positive { collect_vector_input(vector, batch)?; } for vector in negative { collect_vector_input(vector, batch)?; } Ok(()) } pub(crate) fn collect_query(query: &Query, batch: &mut BatchAccumGrpc) -> Result<(), Status> { let Some(variant) = &query.variant else { return Ok(()); }; match variant { query::Variant::Nearest(nearest) => collect_vector_input(nearest, batch)?, query::Variant::Recommend(recommend) => collect_recommend_input(recommend, batch)?, query::Variant::Discover(discover) => collect_discover_input(discover, batch)?, query::Variant::Context(context) => collect_context_input(context, batch)?, query::Variant::OrderBy(_) => {} query::Variant::Fusion(_) => {} query::Variant::Rrf(_) => {} query::Variant::Sample(_) => {} query::Variant::Formula(_) => {} query::Variant::NearestWithMmr(nearest_with_mmr) => { nearest_with_mmr .nearest .as_ref() .map(|vector| collect_vector_input(vector, batch)) .transpose()?; } } Ok(()) } pub(crate) fn collect_prefetch( prefetch: &PrefetchQuery, batch: &mut BatchAccumGrpc, ) -> Result<(), Status> { let PrefetchQuery { prefetch, query, using: _, filter: _, params: _, score_threshold: _, limit: _, lookup_from: _, } = prefetch; if let Some(query) = query { collect_query(query, batch)?; } for p in prefetch { collect_prefetch(p, batch)?; } Ok(()) } #[cfg(test)] mod tests { use api::rest::schema::{Document, Image, InferenceObject}; use serde_json::json; use super::*; fn create_test_document(text: &str) -> Document { Document { text: text.to_string(), model: "test-model".to_string(), options: Default::default(), } } fn create_test_image(url: &str) -> Image { Image { image: json!({"data": url.to_string()}), model: "test-model".to_string(), options: Default::default(), } } fn create_test_object(data: &str) -> InferenceObject { InferenceObject { object: json!({"data": data}), model: "test-model".to_string(), options: Default::default(), } } #[test] fn test_batch_accum_basic() { let mut batch = BatchAccumGrpc::new(); assert!(batch.objects.is_empty()); let doc = InferenceData::Document(create_test_document("test")); batch.add(doc.clone()); assert_eq!(batch.objects.len(), 1); batch.add(doc); assert_eq!(batch.objects.len(), 1); } #[test] fn test_batch_accum_extend() { let mut batch1 = BatchAccumGrpc::new(); let mut batch2 = BatchAccumGrpc::new(); let doc1 = InferenceData::Document(create_test_document("test1")); let doc2 = InferenceData::Document(create_test_document("test2")); batch1.add(doc1); batch2.add(doc2); batch1.extend(batch2); assert_eq!(batch1.objects.len(), 2); } #[test] fn test_deduplication() { let mut batch = BatchAccumGrpc::new(); let doc1 = InferenceData::Document(create_test_document("same")); let doc2 = InferenceData::Document(create_test_document("same")); batch.add(doc1); batch.add(doc2); assert_eq!(batch.objects.len(), 1); } #[test] fn test_different_model_same_content() { let mut batch = BatchAccumGrpc::new(); let mut doc1 = create_test_document("same"); let mut doc2 = create_test_document("same"); doc1.model = "model1".to_string(); doc2.model = "model2".to_string(); batch.add(InferenceData::Document(doc1)); batch.add(InferenceData::Document(doc2)); assert_eq!(batch.objects.len(), 2); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/mod.rs
src/common/inference/mod.rs
#![allow(dead_code)] mod batch_processing; mod batch_processing_grpc; pub mod bm25; pub(crate) mod config; mod infer_processing; pub mod inference_input; mod local_model; pub mod params; pub mod query_requests_grpc; pub mod query_requests_rest; pub mod service; pub mod token; pub mod update_requests;
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/infer_processing.rs
src/common/inference/infer_processing.rs
use std::collections::{HashMap, HashSet}; use api::rest::models::InferenceUsage; use collection::operations::point_ops::VectorPersisted; use storage::content_manager::errors::StorageError; use super::batch_processing::BatchAccum; use super::service::{ InferenceData, InferenceInput, InferenceResponse, InferenceService, InferenceType, }; use crate::common::inference::params::InferenceParams; pub struct BatchAccumInferred { pub(crate) objects: HashMap<InferenceData, VectorPersisted>, } impl BatchAccumInferred { pub fn new() -> Self { Self { objects: HashMap::new(), } } pub async fn from_objects( objects: HashSet<InferenceData>, inference_type: InferenceType, inference_params: InferenceParams, ) -> Result<(Self, Option<InferenceUsage>), StorageError> { if objects.is_empty() { return Ok((Self::new(), None)); } let Some(service) = InferenceService::get_global() else { return Err(StorageError::service_error( "InferenceService is not initialized. Please check if it was properly configured and initialized during startup.", )); }; service.validate()?; let objects_serialized: Vec<_> = objects.into_iter().collect(); let inference_inputs: Vec<_> = objects_serialized .iter() .cloned() .map(InferenceInput::from) .collect(); let InferenceResponse { embeddings, usage } = service .infer(inference_inputs, inference_type, inference_params) .await?; if embeddings.is_empty() { return Err(StorageError::service_error( "Inference service returned no vectors. Check if models are properly loaded.", )); } let objects = objects_serialized.into_iter().zip(embeddings).collect(); Ok((Self { objects }, usage)) } pub async fn from_batch_accum( batch: BatchAccum, inference_type: InferenceType, inference_params: &InferenceParams, ) -> Result<(Self, Option<InferenceUsage>), StorageError> { let BatchAccum { objects } = batch; Self::from_objects(objects, inference_type, inference_params.clone()).await } pub fn get_vector(&self, data: &InferenceData) -> Option<&VectorPersisted> { self.objects.get(data) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/inference/token.rs
src/common/inference/token.rs
use std::convert::Infallible; use std::future::{Ready, ready}; use actix_web::{FromRequest, HttpMessage}; #[derive(Debug, Clone, PartialEq, Default)] pub struct InferenceToken(pub Option<String>); impl InferenceToken { pub fn new(key: impl Into<String>) -> Self { InferenceToken(Some(key.into())) } pub fn as_str(&self) -> Option<&str> { self.0.as_deref() } } impl From<&str> for InferenceToken { fn from(s: &str) -> Self { InferenceToken::new(s) } } impl FromRequest for InferenceToken { type Error = Infallible; type Future = Ready<Result<Self, Self::Error>>; fn from_request( req: &actix_web::HttpRequest, _payload: &mut actix_web::dev::Payload, ) -> Self::Future { let api_key = req.extensions().get::<InferenceToken>().cloned(); ready(Ok(api_key.unwrap_or_default())) } } pub fn extract_token<R>(req: &tonic::Request<R>) -> InferenceToken { req.extensions() .get::<InferenceToken>() .cloned() .unwrap_or(InferenceToken(None)) }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/telemetry_ops/requests_telemetry.rs
src/common/telemetry_ops/requests_telemetry.rs
use std::collections::HashMap; use std::sync::Arc; use common::types::TelemetryDetail; use parking_lot::Mutex; use schemars::JsonSchema; use segment::common::anonymize::{Anonymize, anonymize_collection_values}; use segment::common::operation_time_statistics::{ OperationDurationStatistics, OperationDurationsAggregator, ScopeDurationMeasurer, }; use serde::Serialize; use storage::rbac::{Access, AccessRequirements}; pub type HttpStatusCode = u16; #[derive(Serialize, Clone, Default, Debug, JsonSchema)] pub struct WebApiTelemetry { pub responses: HashMap<String, HashMap<HttpStatusCode, OperationDurationStatistics>>, } #[derive(Serialize, Clone, Default, Debug, JsonSchema, Anonymize)] pub struct GrpcTelemetry { #[anonymize(with = anonymize_collection_values)] pub responses: HashMap<String, OperationDurationStatistics>, } pub struct ActixTelemetryCollector { pub workers: Vec<Arc<Mutex<ActixWorkerTelemetryCollector>>>, } #[derive(Default)] pub struct ActixWorkerTelemetryCollector { methods: HashMap<String, HashMap<HttpStatusCode, Arc<Mutex<OperationDurationsAggregator>>>>, } pub struct TonicTelemetryCollector { pub workers: Vec<Arc<Mutex<TonicWorkerTelemetryCollector>>>, } #[derive(Default)] pub struct TonicWorkerTelemetryCollector { methods: HashMap<String, Arc<Mutex<OperationDurationsAggregator>>>, } impl ActixTelemetryCollector { pub fn create_web_worker_telemetry(&mut self) -> Arc<Mutex<ActixWorkerTelemetryCollector>> { let worker: Arc<Mutex<_>> = Default::default(); self.workers.push(worker.clone()); worker } pub fn get_telemetry_data(&self, detail: TelemetryDetail) -> WebApiTelemetry { let mut result = WebApiTelemetry::default(); for web_data in &self.workers { let lock = web_data.lock().get_telemetry_data(detail); result.merge(&lock); } result } } impl TonicTelemetryCollector { pub fn create_grpc_telemetry_collector(&mut self) -> Arc<Mutex<TonicWorkerTelemetryCollector>> { let worker: Arc<Mutex<_>> = Default::default(); self.workers.push(worker.clone()); worker } pub fn get_telemetry_data(&self, detail: TelemetryDetail) -> GrpcTelemetry { let mut result = GrpcTelemetry::default(); for grpc_data in &self.workers { let lock = grpc_data.lock().get_telemetry_data(detail); result.merge(&lock); } result } } impl TonicWorkerTelemetryCollector { pub fn add_response(&mut self, method: String, instant: std::time::Instant) { let aggregator = self .methods .entry(method) .or_insert_with(OperationDurationsAggregator::new); ScopeDurationMeasurer::new_with_instant(aggregator, instant); } pub fn get_telemetry_data(&self, detail: TelemetryDetail) -> GrpcTelemetry { let mut responses = HashMap::new(); for (method, aggregator) in self.methods.iter() { responses.insert(method.clone(), aggregator.lock().get_statistics(detail)); } GrpcTelemetry { responses } } } impl ActixWorkerTelemetryCollector { pub fn add_response( &mut self, method: String, status_code: HttpStatusCode, instant: std::time::Instant, ) { let aggregator = self .methods .entry(method) .or_default() .entry(status_code) .or_insert_with(OperationDurationsAggregator::new); ScopeDurationMeasurer::new_with_instant(aggregator, instant); } pub fn get_telemetry_data(&self, detail: TelemetryDetail) -> WebApiTelemetry { let mut responses = HashMap::new(); for (method, status_codes) in &self.methods { let mut status_codes_map = HashMap::new(); for (status_code, aggregator) in status_codes { status_codes_map.insert(*status_code, aggregator.lock().get_statistics(detail)); } responses.insert(method.clone(), status_codes_map); } WebApiTelemetry { responses } } } impl GrpcTelemetry { pub fn merge(&mut self, other: &GrpcTelemetry) { for (method, other_statistics) in &other.responses { let entry = self.responses.entry(method.clone()).or_default(); *entry = entry.clone() + other_statistics.clone(); } } } impl WebApiTelemetry { pub fn merge(&mut self, other: &WebApiTelemetry) { for (method, status_codes) in &other.responses { let status_codes_map = self.responses.entry(method.clone()).or_default(); for (status_code, statistics) in status_codes { let entry = status_codes_map.entry(*status_code).or_default(); *entry = entry.clone() + statistics.clone(); } } } } #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] pub struct RequestsTelemetry { pub rest: WebApiTelemetry, pub grpc: GrpcTelemetry, } impl RequestsTelemetry { pub fn collect( access: &Access, actix_collector: &ActixTelemetryCollector, tonic_collector: &TonicTelemetryCollector, detail: TelemetryDetail, ) -> Option<Self> { let global_access = AccessRequirements::new(); if access.check_global_access(global_access).is_ok() { let rest = actix_collector.get_telemetry_data(detail); let grpc = tonic_collector.get_telemetry_data(detail); Some(Self { rest, grpc }) } else { None } } } impl Anonymize for WebApiTelemetry { fn anonymize(&self) -> Self { let responses = self .responses .iter() .map(|(key, value)| (key.clone(), anonymize_collection_values(value))) .collect(); WebApiTelemetry { responses } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/telemetry_ops/hardware.rs
src/common/telemetry_ops/hardware.rs
use std::collections::HashMap; use api::rest::models::HardwareUsage; use schemars::JsonSchema; use segment::common::anonymize::Anonymize; use serde::Serialize; use storage::dispatcher::Dispatcher; use storage::rbac::{Access, AccessRequirements}; #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] pub struct HardwareTelemetry { pub(crate) collection_data: HashMap<String, HardwareUsage>, } impl HardwareTelemetry { pub(crate) fn new(dispatcher: &Dispatcher, access: &Access) -> Self { let mut all_hw_metrics = dispatcher.all_hw_metrics(); let collection_data = match access { Access::Global(_) => all_hw_metrics, Access::Collection(collection_access_list) => { let required_access = AccessRequirements::new(); let allowed_collections = collection_access_list.meeting_requirements(required_access); let mut resolved_collection_data = HashMap::new(); for collection in allowed_collections { if let Some(hw_metrics) = all_hw_metrics.remove(collection) { resolved_collection_data.insert(collection.clone(), hw_metrics); } } resolved_collection_data } }; Self { collection_data } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/telemetry_ops/collections_telemetry.rs
src/common/telemetry_ops/collections_telemetry.rs
use std::sync::atomic::AtomicBool; use std::time::Duration; use collection::operations::types::CollectionResult; use collection::telemetry::{ CollectionSnapshotTelemetry, CollectionTelemetry, CollectionsAggregatedTelemetry, }; use common::types::{DetailsLevel, TelemetryDetail}; use schemars::JsonSchema; use segment::common::anonymize::Anonymize; use serde::Serialize; use storage::content_manager::toc::TableOfContent; use storage::rbac::Access; #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] #[serde(untagged)] pub enum CollectionTelemetryEnum { Full(Box<CollectionTelemetry>), Aggregated(CollectionsAggregatedTelemetry), } #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize, Default)] pub struct CollectionsTelemetry { #[anonymize(false)] pub number_of_collections: usize, #[anonymize(false)] #[serde(skip_serializing_if = "Option::is_none")] pub max_collections: Option<usize>, #[serde(skip_serializing_if = "Option::is_none")] pub collections: Option<Vec<CollectionTelemetryEnum>>, #[serde(skip_serializing_if = "Option::is_none")] pub snapshots: Option<Vec<CollectionSnapshotTelemetry>>, } impl CollectionsTelemetry { pub async fn collect( detail: TelemetryDetail, access: &Access, toc: &TableOfContent, timeout: Duration, is_stopped: &AtomicBool, ) -> CollectionResult<Self> { let number_of_collections = toc.all_collections(access).await.len(); let (collections, snapshots) = if detail.level >= DetailsLevel::Level1 { let telemetry_data = if detail.level >= DetailsLevel::Level2 { let toc_telemetry = toc .get_telemetry_data(detail, access, timeout, is_stopped) .await?; let collections: Vec<_> = toc_telemetry .collection_telemetry .into_iter() .map(|t| CollectionTelemetryEnum::Full(Box::new(t))) .collect(); (collections, toc_telemetry.snapshot_telemetry) } else { let collections = toc .get_aggregated_telemetry_data(access, timeout, is_stopped) .await? .into_iter() .map(CollectionTelemetryEnum::Aggregated) .collect(); (collections, vec![]) }; (Some(telemetry_data.0), Some(telemetry_data.1)) } else { (None, None) }; let max_collections = toc.max_collections(); Ok(CollectionsTelemetry { number_of_collections, max_collections, collections, snapshots, }) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/telemetry_ops/memory_telemetry.rs
src/common/telemetry_ops/memory_telemetry.rs
use schemars::JsonSchema; use segment::common::anonymize::Anonymize; use serde::Serialize; use storage::rbac::Access; #[cfg(all( not(target_env = "msvc"), any(target_arch = "x86_64", target_arch = "aarch64") ))] use storage::rbac::AccessRequirements; #[cfg(all( not(target_env = "msvc"), any(target_arch = "x86_64", target_arch = "aarch64") ))] use tikv_jemalloc_ctl::{epoch, stats}; #[derive(Debug, Clone, Default, JsonSchema, Serialize, Anonymize)] #[anonymize(false)] pub struct MemoryTelemetry { /// Total number of bytes in active pages allocated by the application pub active_bytes: usize, /// Total number of bytes allocated by the application pub allocated_bytes: usize, /// Total number of bytes dedicated to metadata pub metadata_bytes: usize, /// Maximum number of bytes in physically resident data pages mapped pub resident_bytes: usize, /// Total number of bytes in virtual memory mappings pub retained_bytes: usize, } impl MemoryTelemetry { #[cfg(all( not(target_env = "msvc"), any(target_arch = "x86_64", target_arch = "aarch64") ))] pub fn collect(access: &Access) -> Option<MemoryTelemetry> { let required_access = AccessRequirements::new(); if epoch::advance().is_ok() && access.check_global_access(required_access).is_ok() { Some(MemoryTelemetry { active_bytes: stats::active::read().unwrap_or_default(), allocated_bytes: stats::allocated::read().unwrap_or_default(), metadata_bytes: stats::metadata::read().unwrap_or_default(), resident_bytes: stats::resident::read().unwrap_or_default(), retained_bytes: stats::retained::read().unwrap_or_default(), }) } else { log::info!("Failed to advance Jemalloc stats epoch"); None } } #[cfg(target_env = "msvc")] pub fn collect(_access: &Access) -> Option<MemoryTelemetry> { None } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/telemetry_ops/app_telemetry.rs
src/common/telemetry_ops/app_telemetry.rs
use std::path::Path; use chrono::{DateTime, SubsecRound, Utc}; use common::flags::FeatureFlags; use common::types::{DetailsLevel, TelemetryDetail}; use schemars::JsonSchema; use segment::common::anonymize::Anonymize; use segment::types::HnswGlobalConfig; use serde::Serialize; use crate::settings::Settings; pub struct AppBuildTelemetryCollector { pub startup: DateTime<Utc>, } impl AppBuildTelemetryCollector { pub fn new() -> Self { AppBuildTelemetryCollector { startup: Utc::now().round_subsecs(2), } } } #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] pub struct AppFeaturesTelemetry { pub debug: bool, pub service_debug_feature: bool, pub recovery_mode: bool, pub gpu: bool, pub rocksdb: bool, } #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] pub struct RunningEnvironmentTelemetry { #[anonymize(false)] distribution: Option<String>, #[anonymize(false)] distribution_version: Option<String>, is_docker: bool, #[anonymize(false)] cores: Option<usize>, ram_size: Option<usize>, disk_size: Option<usize>, #[anonymize(false)] cpu_flags: String, #[serde(skip_serializing_if = "Option::is_none")] cpu_endian: Option<CpuEndian>, #[serde(skip_serializing_if = "Option::is_none")] gpu_devices: Option<Vec<GpuDeviceTelemetry>>, } #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] pub struct AppBuildTelemetry { #[anonymize(false)] pub name: String, #[anonymize(false)] pub version: String, #[serde(skip_serializing_if = "Option::is_none")] pub features: Option<AppFeaturesTelemetry>, #[anonymize(value = None)] #[serde(skip_serializing_if = "Option::is_none")] pub runtime_features: Option<FeatureFlags>, #[serde(skip_serializing_if = "Option::is_none")] pub hnsw_global_config: Option<HnswGlobalConfig>, #[serde(skip_serializing_if = "Option::is_none")] pub system: Option<RunningEnvironmentTelemetry>, #[serde(skip_serializing_if = "Option::is_none")] pub jwt_rbac: Option<bool>, #[serde(skip_serializing_if = "Option::is_none")] pub hide_jwt_dashboard: Option<bool>, pub startup: DateTime<Utc>, } impl AppBuildTelemetry { pub fn collect( detail: TelemetryDetail, collector: &AppBuildTelemetryCollector, settings: &Settings, ) -> Self { AppBuildTelemetry { name: env!("CARGO_PKG_NAME").to_string(), version: env!("CARGO_PKG_VERSION").to_string(), features: (detail.level >= DetailsLevel::Level1).then(|| AppFeaturesTelemetry { debug: cfg!(debug_assertions), service_debug_feature: cfg!(feature = "service_debug"), recovery_mode: settings.storage.recovery_mode.is_some(), gpu: cfg!(feature = "gpu"), rocksdb: cfg!(feature = "rocksdb"), }), runtime_features: (detail.level >= DetailsLevel::Level1) .then(common::flags::feature_flags), hnsw_global_config: (detail.level >= DetailsLevel::Level1) .then(|| settings.storage.hnsw_global_config.clone()), system: (detail.level >= DetailsLevel::Level1).then(get_system_data), jwt_rbac: settings.service.jwt_rbac, hide_jwt_dashboard: settings.service.hide_jwt_dashboard, startup: collector.startup, } } } fn get_system_data() -> RunningEnvironmentTelemetry { let distribution = if let Ok(release) = sys_info::linux_os_release() { release.id } else { sys_info::os_type().ok() }; let distribution_version = if let Ok(release) = sys_info::linux_os_release() { release.version_id } else { sys_info::os_release().ok() }; let mut cpu_flags = vec![]; #[cfg(any(target_arch = "x86", target_arch = "x86_64"))] { if std::arch::is_x86_feature_detected!("sse") { cpu_flags.push("sse"); } if std::arch::is_x86_feature_detected!("sse2") { cpu_flags.push("sse2"); } if std::arch::is_x86_feature_detected!("avx") { cpu_flags.push("avx"); } if std::arch::is_x86_feature_detected!("avx2") { cpu_flags.push("avx2"); } if std::arch::is_x86_feature_detected!("fma") { cpu_flags.push("fma"); } if std::arch::is_x86_feature_detected!("f16c") { cpu_flags.push("f16c"); } if std::arch::is_x86_feature_detected!("avx512f") { cpu_flags.push("avx512f"); } if std::arch::is_x86_feature_detected!("avx512vl") { cpu_flags.push("avx512vl"); } if std::arch::is_x86_feature_detected!("avx512vpopcntdq") { cpu_flags.push("avx512vpopcntdq"); } } #[cfg(all(target_arch = "aarch64", target_feature = "neon"))] { if std::arch::is_aarch64_feature_detected!("neon") { cpu_flags.push("neon"); } if std::arch::is_aarch64_feature_detected!("fp16") { cpu_flags.push("fp16"); } } #[cfg(feature = "gpu")] let gpu_devices = segment::index::hnsw_index::gpu::GPU_DEVICES_MANAGER .read() .as_ref() .map(|gpu_devices_manager| { gpu_devices_manager .all_found_device_names() .iter() .map(|name| GpuDeviceTelemetry { name: name.clone() }) .collect::<Vec<_>>() }); #[cfg(not(feature = "gpu"))] let gpu_devices = None; RunningEnvironmentTelemetry { distribution, distribution_version, is_docker: cfg!(unix) && Path::new("/.dockerenv").exists(), cores: sys_info::cpu_num().ok().map(|x| x as usize), ram_size: sys_info::mem_info().ok().map(|x| x.total as usize), disk_size: sys_info::disk_info().ok().map(|x| x.total as usize), cpu_flags: cpu_flags.join(","), cpu_endian: Some(CpuEndian::current()), gpu_devices, } } #[derive(Serialize, Clone, Copy, Debug, JsonSchema, Anonymize)] #[serde(rename_all = "snake_case")] pub enum CpuEndian { Little, Big, Other, } impl CpuEndian { /// Get the current used byte order pub const fn current() -> Self { if cfg!(target_endian = "little") { CpuEndian::Little } else if cfg!(target_endian = "big") { CpuEndian::Big } else { CpuEndian::Other } } } #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] pub struct GpuDeviceTelemetry { #[anonymize(false)] pub name: String, }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/telemetry_ops/mod.rs
src/common/telemetry_ops/mod.rs
pub mod app_telemetry; pub mod cluster_telemetry; pub mod collections_telemetry; pub mod hardware; pub mod memory_telemetry; pub mod requests_telemetry;
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/common/telemetry_ops/cluster_telemetry.rs
src/common/telemetry_ops/cluster_telemetry.rs
use std::collections::HashMap; use collection::operations::types::PeerMetadata; use collection::shards::shard::PeerId; use common::types::{DetailsLevel, TelemetryDetail}; use schemars::JsonSchema; use segment::common::anonymize::{Anonymize, anonymize_collection_values_opt}; use serde::Serialize; use storage::dispatcher::Dispatcher; use storage::rbac::{Access, AccessRequirements}; use storage::types::{ClusterStatus, ConsensusThreadStatus, PeerInfo, StateRole}; use crate::settings::Settings; #[derive(Serialize, Clone, Debug, JsonSchema)] pub struct P2pConfigTelemetry { connection_pool_size: usize, } #[derive(Serialize, Clone, Debug, JsonSchema)] pub struct ConsensusConfigTelemetry { max_message_queue_size: usize, tick_period_ms: u64, bootstrap_timeout_sec: u64, } #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] #[anonymize(false)] pub struct ClusterConfigTelemetry { grpc_timeout_ms: u64, p2p: P2pConfigTelemetry, consensus: ConsensusConfigTelemetry, } impl From<&Settings> for ClusterConfigTelemetry { fn from(settings: &Settings) -> Self { ClusterConfigTelemetry { grpc_timeout_ms: settings.cluster.grpc_timeout_ms, p2p: P2pConfigTelemetry { connection_pool_size: settings.cluster.p2p.connection_pool_size, }, consensus: ConsensusConfigTelemetry { max_message_queue_size: settings.cluster.consensus.max_message_queue_size, tick_period_ms: settings.cluster.consensus.tick_period_ms, bootstrap_timeout_sec: settings.cluster.consensus.bootstrap_timeout_sec, }, } } } #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] pub struct ClusterStatusTelemetry { #[anonymize(false)] pub number_of_peers: usize, #[anonymize(false)] pub term: u64, #[anonymize(false)] pub commit: u64, #[anonymize(false)] pub pending_operations: usize, pub role: Option<StateRole>, pub is_voter: bool, #[anonymize(false)] pub peer_id: Option<PeerId>, pub consensus_thread_status: ConsensusThreadStatus, } #[derive(Serialize, Clone, Debug, JsonSchema, Anonymize)] pub struct ClusterTelemetry { pub enabled: bool, #[serde(skip_serializing_if = "Option::is_none")] pub status: Option<ClusterStatusTelemetry>, #[serde(skip_serializing_if = "Option::is_none")] pub config: Option<ClusterConfigTelemetry>, #[serde(skip_serializing_if = "Option::is_none")] #[anonymize(with = anonymize_collection_values_opt)] pub peers: Option<HashMap<PeerId, PeerInfo>>, #[serde(skip_serializing_if = "Option::is_none")] #[anonymize(false)] pub peer_metadata: Option<HashMap<PeerId, PeerMetadata>>, #[serde(skip_serializing_if = "Option::is_none")] pub metadata: Option<HashMap<String, serde_json::Value>>, } impl ClusterTelemetry { pub fn collect( access: &Access, detail: TelemetryDetail, dispatcher: &Dispatcher, settings: &Settings, ) -> Option<ClusterTelemetry> { let global_access = AccessRequirements::new(); if access.check_global_access(global_access).is_err() { return None; } Some(ClusterTelemetry { enabled: settings.cluster.enabled, status: (detail.level >= DetailsLevel::Level1) .then(|| match dispatcher.cluster_status() { ClusterStatus::Disabled => None, ClusterStatus::Enabled(cluster_info) => Some(ClusterStatusTelemetry { number_of_peers: cluster_info.peers.len(), term: cluster_info.raft_info.term, commit: cluster_info.raft_info.commit, pending_operations: cluster_info.raft_info.pending_operations, role: cluster_info.raft_info.role, is_voter: cluster_info.raft_info.is_voter, peer_id: Some(cluster_info.peer_id), consensus_thread_status: cluster_info.consensus_thread_status, }), }) .flatten(), config: (detail.level >= DetailsLevel::Level2) .then(|| ClusterConfigTelemetry::from(settings)), peers: (detail.level >= DetailsLevel::Level2) .then(|| match dispatcher.cluster_status() { ClusterStatus::Disabled => None, ClusterStatus::Enabled(cluster_info) => Some(cluster_info.peers), }) .flatten(), peer_metadata: (detail.level >= DetailsLevel::Level3) .then(|| { dispatcher .consensus_state() .map(|state| state.persistent.read().peer_metadata_by_id()) }) .flatten(), metadata: (detail.level >= DetailsLevel::Level1) .then(|| { dispatcher .consensus_state() .map(|state| state.persistent.read().cluster_metadata.clone()) .filter(|metadata| !metadata.is_empty()) }) .flatten(), }) } pub fn this_peer_id(&self) -> Option<PeerId> { self.enabled .then(|| self.status.as_ref().and_then(|j| j.peer_id)) .flatten() } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/migrations/single_to_cluster.rs
src/migrations/single_to_cluster.rs
use std::sync::Arc; use collection::collection_state::State; use collection::config::{CollectionConfigInternal, ShardingMethod}; use collection::shards::replica_set::replica_set_state::ReplicaState; use collection::shards::shard::PeerId; use storage::content_manager::collection_meta_ops::{ CollectionMetaOperations, CreateCollection, CreateCollectionOperation, CreateShardKey, SetShardReplicaState, }; use storage::content_manager::consensus_manager::ConsensusStateRef; use storage::content_manager::shard_distribution::ShardDistributionProposal; use storage::content_manager::toc::TableOfContent; use storage::dispatcher::Dispatcher; use storage::rbac::{Access, AccessRequirements}; /// Processes the existing collections, which were created outside the consensus: /// - during the migration from single to cluster /// - during restoring from a backup pub async fn handle_existing_collections( toc_arc: Arc<TableOfContent>, consensus_state: ConsensusStateRef, dispatcher_arc: Arc<Dispatcher>, this_peer_id: PeerId, collections: Vec<String>, ) { let full_access = Access::full("Migration from single to cluster"); let multipass = full_access .check_global_access(AccessRequirements::new().manage()) .expect("Full access should have manage rights"); consensus_state.is_leader_established.await_ready(); for collection_name in collections { let Ok(collection_obj) = toc_arc .get_collection(&multipass.issue_pass(&collection_name)) .await else { break; }; let State { config, shards, resharding: _, // resharding can't exist outside of consensus transfers: _, // transfers can't exist outside of consensus shards_key_mapping, payload_index_schema: _, // payload index schema doesn't require special handling in this case } = collection_obj.state().await; let CollectionConfigInternal { params, hnsw_config, optimizer_config, wal_config, quantization_config, strict_mode_config, uuid, metadata, } = config; let shards_number = params.shard_number.get(); let sharding_method = params.sharding_method; let mut collection_create_operation = CreateCollectionOperation::new( collection_name.clone(), CreateCollection { vectors: params.vectors, sparse_vectors: params.sparse_vectors, shard_number: Some(shards_number), sharding_method, replication_factor: Some(params.replication_factor.get()), write_consistency_factor: Some(params.write_consistency_factor.get()), on_disk_payload: Some(params.on_disk_payload), hnsw_config: Some(hnsw_config.into()), wal_config: Some(wal_config.into()), optimizers_config: Some(optimizer_config.into()), quantization_config, strict_mode_config, uuid, metadata, }, ) .expect("Failed to create collection operation"); let mut consensus_operations = Vec::new(); match sharding_method.unwrap_or_default() { ShardingMethod::Auto => { collection_create_operation.set_distribution(ShardDistributionProposal { distribution: shards .iter() .filter_map(|(shard_id, shard_info)| { if shard_info.replicas.contains_key(&this_peer_id) { Some((*shard_id, vec![this_peer_id])) } else { None } }) .collect(), }); consensus_operations.push(CollectionMetaOperations::CreateCollection( collection_create_operation, )); } ShardingMethod::Custom => { // We should create additional consensus operations here to set the shard distribution collection_create_operation.set_distribution(ShardDistributionProposal::empty()); consensus_operations.push(CollectionMetaOperations::CreateCollection( collection_create_operation, )); for (shard_key, shard_ids) in shards_key_mapping.iter() { let mut placement = Vec::new(); for shard_id in shard_ids { let shard_info = shards.get(shard_id).unwrap(); placement.push(shard_info.replicas.keys().copied().collect()); } consensus_operations.push(CollectionMetaOperations::CreateShardKey( CreateShardKey { collection_name: collection_name.clone(), shard_key: shard_key.clone(), placement, initial_state: None, // Initial state can't be set during migration }, )) } } } for operation in consensus_operations { let _res = dispatcher_arc .submit_collection_meta_op(operation, full_access.clone(), None) .await; } for (shard_id, shard_info) in shards { if shard_info.replicas.contains_key(&this_peer_id) { let _res = dispatcher_arc .submit_collection_meta_op( CollectionMetaOperations::SetShardReplicaState(SetShardReplicaState { collection_name: collection_name.clone(), shard_id, peer_id: this_peer_id, state: ReplicaState::Active, from_state: None, }), full_access.clone(), None, ) .await; } } } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/migrations/mod.rs
src/migrations/mod.rs
pub mod single_to_cluster;
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/actix/certificate_helpers.rs
src/actix/certificate_helpers.rs
use std::fmt::Debug; use std::io::{self, BufRead, BufReader}; use std::sync::Arc; use std::time::{Duration, Instant}; use fs_err::File; use parking_lot::RwLock; use rustls::client::VerifierBuilderError; use rustls::pki_types::CertificateDer; use rustls::server::{ClientHello, ResolvesServerCert, WebPkiClientVerifier}; use rustls::sign::CertifiedKey; use rustls::{RootCertStore, ServerConfig, crypto}; use rustls_pemfile::Item; use crate::settings::{Settings, TlsConfig}; type Result<T> = std::result::Result<T, Error>; /// A TTL based rotating server certificate resolver #[derive(Debug)] struct RotatingCertificateResolver { /// TLS configuration used for loading/refreshing certified key tls_config: TlsConfig, /// TTL for each rotation ttl: Option<Duration>, /// Current certified key key: RwLock<CertifiedKeyWithAge>, } impl RotatingCertificateResolver { pub fn new(tls_config: TlsConfig, ttl: Option<Duration>) -> Result<Self> { let certified_key = load_certified_key(&tls_config)?; Ok(Self { tls_config, ttl, key: RwLock::new(CertifiedKeyWithAge::from(certified_key)), }) } /// Get certificate key or refresh /// /// The key is automatically refreshed when the TTL is reached. /// If refreshing fails, an error is logged and the old key is persisted. fn get_key_or_refresh(&self) -> Arc<CertifiedKey> { // Get read-only lock to the key. If TTL is not configured or is not expired, return key. let key = self.key.read(); let ttl = match self.ttl { Some(ttl) if key.is_expired(ttl) => ttl, _ => return key.key.clone(), }; drop(key); // If TTL is expired: // - get read-write lock to the key // - *re-check that TTL is expired* (to avoid refreshing the key multiple times from concurrent threads) // - refresh and return the key let mut key = self.key.write(); if key.is_expired(ttl) && let Err(err) = key.refresh(&self.tls_config) { log::error!("Failed to refresh server TLS certificate, keeping current: {err}"); } key.key.clone() } } impl ResolvesServerCert for RotatingCertificateResolver { fn resolve(&self, _client_hello: ClientHello<'_>) -> Option<Arc<CertifiedKey>> { Some(self.get_key_or_refresh()) } } #[derive(Debug)] struct CertifiedKeyWithAge { /// Last time the certificate was updated/replaced last_update: Instant, /// Current certified key key: Arc<CertifiedKey>, } impl CertifiedKeyWithAge { pub fn from(key: Arc<CertifiedKey>) -> Self { Self { last_update: Instant::now(), key, } } pub fn refresh(&mut self, tls_config: &TlsConfig) -> Result<()> { *self = Self::from(load_certified_key(tls_config)?); Ok(()) } pub fn age(&self) -> Duration { self.last_update.elapsed() } pub fn is_expired(&self, ttl: Duration) -> bool { self.age() >= ttl } } /// Load TLS configuration and construct certified key. fn load_certified_key(tls_config: &TlsConfig) -> Result<Arc<CertifiedKey>> { // Load certificates let certs: Vec<CertificateDer> = with_buf_read(&tls_config.cert, |rd| { rustls_pemfile::read_all(rd).collect::<io::Result<Vec<_>>>() })? .into_iter() .filter_map(|item| match item { Item::X509Certificate(data) => Some(data), _ => None, }) .collect(); if certs.is_empty() { return Err(Error::NoServerCert); } // Load private key let private_key_item = with_buf_read(&tls_config.key, rustls_pemfile::read_one)?.ok_or(Error::NoPrivateKey)?; let private_key = match private_key_item { Item::Pkcs1Key(pkey) => rustls_pki_types::PrivateKeyDer::from(pkey), Item::Pkcs8Key(pkey) => rustls_pki_types::PrivateKeyDer::from(pkey), Item::Sec1Key(pkey) => rustls_pki_types::PrivateKeyDer::from(pkey), _ => return Err(Error::InvalidPrivateKey), }; let signing_key = crypto::ring::sign::any_supported_type(&private_key).map_err(Error::Sign)?; // Construct certified key let certified_key = CertifiedKey::new(certs, signing_key); Ok(Arc::new(certified_key)) } /// Generate an actix server configuration with TLS /// /// Uses TLS settings as configured in configuration by user. pub fn actix_tls_server_config(settings: &Settings) -> Result<ServerConfig> { let config = ServerConfig::builder(); let tls_config = settings .tls .clone() .ok_or_else(Settings::tls_config_is_undefined_error) .map_err(Error::Io)?; // Verify client CA or not let config = if settings.service.verify_https_client_certificate { let mut root_cert_store = RootCertStore::empty(); let ca_cert_path = tls_config.ca_cert.as_ref().ok_or(Error::NoCaCert)?; let ca_certs: Vec<CertificateDer> = with_buf_read(ca_cert_path, |rd| rustls_pemfile::certs(rd).collect())?; root_cert_store.add_parsable_certificates(ca_certs); let client_cert_verifier = WebPkiClientVerifier::builder(root_cert_store.into()) .build() .map_err(Error::ClientCertVerifier)?; config.with_client_cert_verifier(client_cert_verifier) } else { config.with_no_client_auth() }; // Configure rotating certificate resolver let ttl = match tls_config.cert_ttl { None | Some(0) => None, Some(seconds) => Some(Duration::from_secs(seconds)), }; let cert_resolver = RotatingCertificateResolver::new(tls_config, ttl)?; let config = config.with_cert_resolver(Arc::new(cert_resolver)); Ok(config) } fn with_buf_read<T>(path: &str, f: impl FnOnce(&mut dyn BufRead) -> io::Result<T>) -> Result<T> { let file = File::open(path).map_err(|err| Error::OpenFile(err, path.into()))?; let mut reader = BufReader::new(file); let dyn_reader: &mut dyn BufRead = &mut reader; f(dyn_reader).map_err(|err| Error::ReadFile(err, path.into())) } /// Actix TLS errors. #[derive(thiserror::Error, Debug)] pub enum Error { #[error("TLS file could not be opened: {1}")] OpenFile(#[source] io::Error, String), #[error("TLS file could not be read: {1}")] ReadFile(#[source] io::Error, String), #[error("general TLS IO error")] Io(#[source] io::Error), #[error("no server certificate found")] NoServerCert, #[error("no private key found")] NoPrivateKey, #[error("invalid private key")] InvalidPrivateKey, #[error("TLS signing error")] Sign(#[source] rustls::Error), #[error("client certificate verification")] ClientCertVerifier(#[source] VerifierBuilderError), #[error("No ca_cert provided")] NoCaCert, }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/actix/helpers.rs
src/actix/helpers.rs
use std::fmt::Debug; use std::future::Future; use actix_web::http::header; use actix_web::http::header::HeaderMap; use actix_web::rt::time::Instant; use actix_web::{HttpResponse, ResponseError, http}; use api::rest::models::{ApiResponse, ApiStatus, HardwareUsage, InferenceUsage, Usage}; use collection::operations::types::CollectionError; use common::counter::hardware_accumulator::HwMeasurementAcc; use serde::Serialize; use storage::content_manager::errors::{StorageError, StorageResult}; use storage::content_manager::toc::request_hw_counter::RequestHwCounter; use storage::dispatcher::Dispatcher; pub fn get_request_hardware_counter( dispatcher: &Dispatcher, collection_name: String, report_to_api: bool, wait: Option<bool>, ) -> RequestHwCounter { let report_to_api = report_to_api && wait != Some(false); RequestHwCounter::new( HwMeasurementAcc::new_with_metrics_drain( dispatcher.get_collection_hw_metrics(collection_name), ), report_to_api, ) } pub fn accepted_response( timing: Instant, hardware_usage: Option<HardwareUsage>, inference_usage: Option<InferenceUsage>, ) -> HttpResponse { let usage = { let u = Usage { hardware: hardware_usage, inference: inference_usage, }; if u.is_empty() { None } else { Some(u) } }; HttpResponse::Accepted().json(ApiResponse::<()> { result: None, status: ApiStatus::Accepted, time: timing.elapsed().as_secs_f64(), usage, }) } pub fn process_response_with_inference_usage<T>( response: Result<T, StorageError>, timing: Instant, hardware_usage: Option<HardwareUsage>, inference_usage: Option<InferenceUsage>, ) -> HttpResponse where T: Serialize, { match response { Ok(res) => HttpResponse::Ok().json(ApiResponse { result: Some(res), status: ApiStatus::Ok, time: timing.elapsed().as_secs_f64(), usage: Some(Usage { hardware: hardware_usage, inference: inference_usage, }), }), Err(err) => process_response_error_with_inference_usage( err, timing, hardware_usage, inference_usage, ), } } pub fn process_response<T>( response: Result<T, StorageError>, timing: Instant, hardware_usage: Option<HardwareUsage>, ) -> HttpResponse where T: Serialize, { process_response_with_inference_usage(response, timing, hardware_usage, None) } pub fn process_response_error_with_inference_usage( err: StorageError, timing: Instant, hardware_usage: Option<HardwareUsage>, inference_usage: Option<InferenceUsage>, ) -> HttpResponse { log_service_error(&err); let error = HttpError::from(err); let http_code = error.status_code(); let headers = error.headers(); let json_body = ApiResponse::<()> { result: None, status: ApiStatus::Error(error.to_string()), time: timing.elapsed().as_secs_f64(), usage: Some(Usage { hardware: hardware_usage, inference: inference_usage, }), }; let mut response_builder = HttpResponse::build(http_code); for header_pair in headers { response_builder.insert_header(header_pair); } response_builder.json(json_body) } pub fn process_response_error( err: StorageError, timing: Instant, hardware_usage: Option<HardwareUsage>, ) -> HttpResponse { process_response_error_with_inference_usage(err, timing, hardware_usage, None) } pub fn already_in_progress_response() -> HttpResponse { HttpResponse::build(http::StatusCode::SERVICE_UNAVAILABLE).json(ApiResponse::<()> { result: None, status: ApiStatus::AlreadyInProgress, time: 0.0, usage: None, }) } /// Response wrapper for a `Future` returning `Result`. /// /// # Cancel safety /// /// Future must be cancel safe. pub async fn time<T, Fut>(future: Fut) -> HttpResponse where Fut: Future<Output = StorageResult<T>>, T: serde::Serialize, { time_impl(async { future.await.map(Some) }).await } /// Response wrapper for a `Future` returning `Result`. /// If `wait` is false, returns `202 Accepted` immediately. pub async fn time_or_accept<T, Fut>(future: Fut, wait: bool) -> HttpResponse where Fut: Future<Output = StorageResult<T>> + Send + 'static, T: serde::Serialize + Send + 'static, { let future = async move { let handle = tokio::task::spawn(async move { let result = future.await; if !wait && let Err(err) = &result { log_service_error(err); } result }); if wait { handle.await?.map(Some) } else { Ok(None) } }; time_impl(future).await } /// # Cancel safety /// /// Future must be cancel safe. async fn time_impl<T, Fut>(future: Fut) -> HttpResponse where Fut: Future<Output = Result<Option<T>, StorageError>>, T: serde::Serialize, { let instant = Instant::now(); match future.await.transpose() { Some(res) => process_response(res, instant, None), None => accepted_response(instant, None, None), } } fn log_service_error(err: &StorageError) { if let StorageError::ServiceError { backtrace, .. } = err { log::error!("Error processing request: {err}"); if let Some(backtrace) = backtrace { log::trace!("Backtrace: {backtrace}"); } } } #[derive(Clone, Debug, thiserror::Error)] #[error("{0}")] pub struct HttpError(StorageError); impl HttpError { fn headers(&self) -> HeaderMap { let mut headers = HeaderMap::new(); match &self.0 { StorageError::RateLimitExceeded { description: _, retry_after, } => { if let Some(retry_after) = retry_after { // Retry-After is expressed in seconds `https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Retry-After` // Ceil the value to the nearest second so clients don't retry too early let retry_after_sec = retry_after.as_secs_f32().ceil() as u32; headers.insert( header::RETRY_AFTER, header::HeaderValue::from(retry_after_sec), ); } } StorageError::BadInput { .. } => {} StorageError::AlreadyExists { .. } => {} StorageError::NotFound { .. } => {} StorageError::ServiceError { .. } => {} StorageError::BadRequest { .. } => {} StorageError::Locked { .. } => {} StorageError::Timeout { .. } => {} StorageError::ChecksumMismatch { .. } => {} StorageError::Forbidden { .. } => {} StorageError::PreconditionFailed { .. } => {} StorageError::InferenceError { .. } => {} StorageError::ShardUnavailable { .. } => {} StorageError::EmptyPartialSnapshot { .. } => {} } headers } } impl ResponseError for HttpError { fn status_code(&self) -> http::StatusCode { match &self.0 { StorageError::BadInput { .. } => http::StatusCode::BAD_REQUEST, StorageError::NotFound { .. } => http::StatusCode::NOT_FOUND, StorageError::ServiceError { .. } => http::StatusCode::INTERNAL_SERVER_ERROR, StorageError::BadRequest { .. } => http::StatusCode::BAD_REQUEST, StorageError::Locked { .. } => http::StatusCode::FORBIDDEN, StorageError::Timeout { .. } => http::StatusCode::REQUEST_TIMEOUT, StorageError::AlreadyExists { .. } => http::StatusCode::CONFLICT, StorageError::ChecksumMismatch { .. } => http::StatusCode::BAD_REQUEST, StorageError::Forbidden { .. } => http::StatusCode::FORBIDDEN, StorageError::PreconditionFailed { .. } => http::StatusCode::INTERNAL_SERVER_ERROR, StorageError::InferenceError { .. } => http::StatusCode::BAD_REQUEST, StorageError::RateLimitExceeded { .. } => http::StatusCode::TOO_MANY_REQUESTS, StorageError::ShardUnavailable { .. } => http::StatusCode::SERVICE_UNAVAILABLE, StorageError::EmptyPartialSnapshot { .. } => http::StatusCode::NOT_MODIFIED, } } } impl From<StorageError> for HttpError { fn from(err: StorageError) -> Self { HttpError(err) } } impl From<CollectionError> for HttpError { fn from(err: CollectionError) -> Self { HttpError(err.into()) } } impl From<std::io::Error> for HttpError { fn from(err: std::io::Error) -> Self { HttpError(err.into()) // TODO: Is this good enough?.. 🤔 } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/actix/web_ui.rs
src/actix/web_ui.rs
use std::path::Path; use actix_web::dev::HttpServiceFactory; use actix_web::http::header::HeaderValue; use actix_web::middleware::DefaultHeaders; use actix_web::web; use crate::settings::Settings; const DEFAULT_STATIC_DIR: &str = "./static"; pub const WEB_UI_PATH: &str = "/dashboard"; pub fn web_ui_folder(settings: &Settings) -> Option<String> { let web_ui_enabled = settings.service.enable_static_content.unwrap_or(true); if web_ui_enabled { let static_folder = settings .service .static_content_dir .clone() .unwrap_or_else(|| DEFAULT_STATIC_DIR.to_string()); let static_folder_path = Path::new(&static_folder); if !static_folder_path.exists() || !static_folder_path.is_dir() { // enabled BUT folder does not exist log::warn!( "Static content folder for Web UI '{}' does not exist", static_folder_path.display(), ); None } else { // enabled AND folder exists Some(static_folder) } } else { // not enabled None } } pub fn web_ui_factory(static_folder: &str) -> impl HttpServiceFactory + use<> { web::scope(WEB_UI_PATH) .wrap(DefaultHeaders::new().add(("X-Frame-Options", HeaderValue::from_static("DENY")))) .service(actix_files::Files::new("/", static_folder).index_file("index.html")) } #[cfg(test)] mod tests { use actix_web::App; use actix_web::http::StatusCode; use actix_web::http::header::{self, HeaderMap}; use actix_web::test::{self, TestRequest}; use super::*; fn assert_html_custom_headers(headers: &HeaderMap) { let content_type = header::HeaderValue::from_static("text/html; charset=utf-8"); assert_eq!(headers.get(header::CONTENT_TYPE), Some(&content_type)); let x_frame_options = header::HeaderValue::from_static("DENY"); assert_eq!(headers.get(header::X_FRAME_OPTIONS), Some(&x_frame_options),); } #[actix_web::test] async fn test_web_ui() { let static_dir = String::from("static"); let mut settings = Settings::new(None).unwrap(); settings.service.static_content_dir = Some(static_dir.clone()); let maybe_static_folder = web_ui_folder(&settings); if maybe_static_folder.is_none() { println!("Skipping test because the static folder was not found."); return; } let static_folder = maybe_static_folder.unwrap(); let srv = test::init_service(App::new().service(web_ui_factory(&static_folder))).await; // Index path (no trailing slash) let req = TestRequest::with_uri(WEB_UI_PATH).to_request(); let res = test::call_service(&srv, req).await; assert_eq!(res.status(), StatusCode::OK); let headers = res.headers(); assert_html_custom_headers(headers); // Index path (trailing slash) let req = TestRequest::with_uri(format!("{WEB_UI_PATH}/").as_str()).to_request(); let res = test::call_service(&srv, req).await; assert_eq!(res.status(), StatusCode::OK); let headers = res.headers(); assert_html_custom_headers(headers); // Index path (index.html file) let req = TestRequest::with_uri(format!("{WEB_UI_PATH}/index.html").as_str()).to_request(); let res = test::call_service(&srv, req).await; assert_eq!(res.status(), StatusCode::OK); let headers = res.headers(); assert_html_custom_headers(headers); // Static asset (favicon.ico) let req = TestRequest::with_uri(format!("{WEB_UI_PATH}/favicon.ico").as_str()).to_request(); let res = test::call_service(&srv, req).await; assert_eq!(res.status(), StatusCode::OK); let headers = res.headers(); assert_eq!( headers.get(header::CONTENT_TYPE), Some(&header::HeaderValue::from_static("image/x-icon")), ); // Non-existing path (404 Not Found) let fake_path = uuid::Uuid::new_v4().to_string(); let srv = test::init_service(App::new().service(web_ui_factory(&fake_path))).await; let req = TestRequest::with_uri(WEB_UI_PATH).to_request(); let res = test::call_service(&srv, req).await; assert_eq!(res.status(), StatusCode::NOT_FOUND); let headers = res.headers(); assert_eq!(headers.get(header::CONTENT_TYPE), None); assert_eq!(headers.get(header::CONTENT_LENGTH), None); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/actix/auth.rs
src/actix/auth.rs
use std::convert::Infallible; use std::future::{Ready, ready}; use std::sync::Arc; use actix_web::body::{BoxBody, EitherBody}; use actix_web::dev::{Service, ServiceRequest, ServiceResponse, Transform, forward_ready}; use actix_web::{Error, FromRequest, HttpMessage, HttpResponse, ResponseError}; use futures_util::future::LocalBoxFuture; use storage::rbac::Access; use super::helpers::HttpError; use crate::common::auth::{AuthError, AuthKeys}; pub struct Auth { auth_keys: AuthKeys, whitelist: Vec<WhitelistItem>, } impl Auth { pub fn new(auth_keys: AuthKeys, whitelist: Vec<WhitelistItem>) -> Self { Self { auth_keys, whitelist, } } } impl<S, B> Transform<S, ServiceRequest> for Auth where S: Service<ServiceRequest, Response = ServiceResponse<EitherBody<B, BoxBody>>, Error = Error> + 'static, S::Future: 'static, B: 'static, { type Response = ServiceResponse<EitherBody<B, BoxBody>>; type Error = Error; type InitError = (); type Transform = AuthMiddleware<S>; type Future = Ready<Result<Self::Transform, Self::InitError>>; fn new_transform(&self, service: S) -> Self::Future { ready(Ok(AuthMiddleware { auth_keys: Arc::new(self.auth_keys.clone()), whitelist: self.whitelist.clone(), service: Arc::new(service), })) } } #[derive(Clone, Eq, PartialEq, Hash)] pub struct WhitelistItem(pub String, pub PathMode); impl WhitelistItem { pub fn exact<S: Into<String>>(path: S) -> Self { Self(path.into(), PathMode::Exact) } pub fn prefix<S: Into<String>>(path: S) -> Self { Self(path.into(), PathMode::Prefix) } pub fn matches(&self, other: &str) -> bool { self.1.check(&self.0, other) } } #[derive(Copy, Clone, Eq, PartialEq, Hash)] pub enum PathMode { /// Path must match exactly Exact, /// Path must have given prefix Prefix, } impl PathMode { fn check(&self, key: &str, other: &str) -> bool { match self { Self::Exact => key == other, Self::Prefix => other.starts_with(key), } } } pub struct AuthMiddleware<S> { auth_keys: Arc<AuthKeys>, /// List of items whitelisted from authentication. whitelist: Vec<WhitelistItem>, service: Arc<S>, } impl<S> AuthMiddleware<S> { pub fn is_path_whitelisted(&self, path: &str) -> bool { self.whitelist.iter().any(|item| item.matches(path)) } } impl<S, B> Service<ServiceRequest> for AuthMiddleware<S> where S: Service<ServiceRequest, Response = ServiceResponse<EitherBody<B, BoxBody>>, Error = Error> + 'static, S::Future: 'static, B: 'static, { type Response = ServiceResponse<EitherBody<B, BoxBody>>; type Error = Error; type Future = LocalBoxFuture<'static, Result<Self::Response, Self::Error>>; forward_ready!(service); fn call(&self, req: ServiceRequest) -> Self::Future { let path = req.path(); if self.is_path_whitelisted(path) { return Box::pin(self.service.call(req)); } let auth_keys = self.auth_keys.clone(); let service = self.service.clone(); Box::pin(async move { match auth_keys .validate_request(|key| req.headers().get(key).and_then(|val| val.to_str().ok())) .await { Ok((access, inference_token)) => { let previous = req.extensions_mut().insert::<Access>(access); req.extensions_mut().insert(inference_token); debug_assert!( previous.is_none(), "Previous access object should not exist in the request" ); service.call(req).await } Err(e) => { let resp = match e { AuthError::Unauthorized(e) => HttpResponse::Unauthorized().body(e), AuthError::Forbidden(e) => HttpResponse::Forbidden().body(e), AuthError::StorageError(e) => HttpError::from(e).error_response(), }; Ok(req.into_response(resp).map_into_right_body()) } } }) } } pub struct ActixAccess(pub Access); impl FromRequest for ActixAccess { type Error = Infallible; type Future = Ready<Result<Self, Self::Error>>; fn from_request( req: &actix_web::HttpRequest, _payload: &mut actix_web::dev::Payload, ) -> Self::Future { let access = req.extensions_mut().remove::<Access>().unwrap_or_else(|| { Access::full("All requests have full by default access when API key is not configured") }); ready(Ok(ActixAccess(access))) } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false
qdrant/qdrant
https://github.com/qdrant/qdrant/blob/f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd/src/actix/mod.rs
src/actix/mod.rs
pub mod actix_telemetry; pub mod api; mod auth; mod certificate_helpers; pub mod helpers; pub mod web_ui; use std::io; use std::sync::Arc; use ::api::rest::models::{ApiResponse, ApiStatus, VersionInfo}; use actix_cors::Cors; use actix_multipart::form::MultipartFormConfig; use actix_multipart::form::tempfile::TempFileConfig; use actix_web::middleware::{Compress, Condition, Logger, NormalizePath}; use actix_web::{App, HttpRequest, HttpResponse, HttpServer, Responder, error, get, web}; use actix_web_extras::middleware::Condition as ConditionEx; use api::facet_api::config_facet_api; use collection::operations::validation; use collection::operations::verification::new_unchecked_verification_pass; use storage::dispatcher::Dispatcher; use storage::rbac::Access; use crate::actix::api::cluster_api::config_cluster_api; use crate::actix::api::collections_api::config_collections_api; use crate::actix::api::count_api::count_points; use crate::actix::api::debug_api::config_debugger_api; use crate::actix::api::discovery_api::config_discovery_api; use crate::actix::api::issues_api::config_issues_api; use crate::actix::api::local_shard_api::config_local_shard_api; use crate::actix::api::profiler_api::config_profiler_api; use crate::actix::api::query_api::config_query_api; use crate::actix::api::recommend_api::config_recommend_api; use crate::actix::api::retrieve_api::{get_point, get_points, scroll_points}; use crate::actix::api::search_api::config_search_api; use crate::actix::api::service_api::config_service_api; use crate::actix::api::shards_api::config_shards_api; use crate::actix::api::snapshot_api::config_snapshots_api; use crate::actix::api::update_api::config_update_api; use crate::actix::auth::{Auth, WhitelistItem}; use crate::actix::web_ui::{WEB_UI_PATH, web_ui_factory, web_ui_folder}; use crate::common::auth::AuthKeys; use crate::common::debugger::DebuggerState; use crate::common::health; use crate::common::http_client::HttpClient; use crate::common::telemetry::TelemetryCollector; use crate::settings::{Settings, max_web_workers}; use crate::tracing::LoggerHandle; #[get("/")] pub async fn index() -> impl Responder { HttpResponse::Ok().json(VersionInfo::default()) } pub fn init( dispatcher: Arc<Dispatcher>, telemetry_collector: Arc<tokio::sync::Mutex<TelemetryCollector>>, health_checker: Option<Arc<health::HealthChecker>>, settings: Settings, logger_handle: LoggerHandle, ) -> io::Result<()> { actix_web::rt::System::new().block_on(async { // Nothing to verify here. let pass = new_unchecked_verification_pass(); let auth_keys = AuthKeys::try_create( &settings.service, dispatcher .toc(&Access::full("For JWT validation"), &pass) .clone(), ); let upload_dir = dispatcher .toc(&Access::full("For upload dir"), &pass) .upload_dir() .unwrap(); let dispatcher_data = web::Data::from(dispatcher); let actix_telemetry_collector = telemetry_collector .lock() .await .actix_telemetry_collector .clone(); let debugger_state = web::Data::new(DebuggerState::from_settings(&settings)); let telemetry_collector_data = web::Data::from(telemetry_collector); let logger_handle_data = web::Data::new(logger_handle); let http_client = web::Data::new(HttpClient::from_settings(&settings)?); let health_checker = web::Data::new(health_checker); let web_ui_available = web_ui_folder(&settings); let service_config = web::Data::new(settings.service.clone()); let mut api_key_whitelist = vec![ WhitelistItem::exact("/"), WhitelistItem::exact("/healthz"), WhitelistItem::prefix("/readyz"), WhitelistItem::prefix("/livez"), ]; if web_ui_available.is_some() { api_key_whitelist.push(WhitelistItem::prefix(WEB_UI_PATH)); } let mut server = HttpServer::new(move || { let cors = Cors::default() .allow_any_origin() .allow_any_method() .allow_any_header(); let validate_path_config = actix_web_validator::PathConfig::default() .error_handler(|err, rec| validation_error_handler("path parameters", err, rec)); let validate_query_config = actix_web_validator::QueryConfig::default() .error_handler(|err, rec| validation_error_handler("query parameters", err, rec)); let validate_json_config = actix_web_validator::JsonConfig::default() .limit(settings.service.max_request_size_mb * 1024 * 1024) .error_handler(|err, rec| validation_error_handler("JSON body", err, rec)); let mut app = App::new() .wrap(Compress::default()) // Reads the `Accept-Encoding` header to negotiate which compression codec to use. // api_key middleware // note: the last call to `wrap()` or `wrap_fn()` is executed first .wrap(ConditionEx::from_option(auth_keys.as_ref().map( |auth_keys| Auth::new(auth_keys.clone(), api_key_whitelist.clone()), ))) // Normalize path .wrap(NormalizePath::trim()) .wrap(Condition::new(settings.service.enable_cors, cors)) .wrap( // Set up logger, but avoid logging hot status endpoints Logger::default() .exclude("/") .exclude("/metrics") .exclude("/telemetry") .exclude("/healthz") .exclude("/readyz") .exclude("/livez"), ) .wrap(actix_telemetry::ActixTelemetryTransform::new( actix_telemetry_collector.clone(), )) .app_data(dispatcher_data.clone()) .app_data(telemetry_collector_data.clone()) .app_data(logger_handle_data.clone()) .app_data(http_client.clone()) .app_data(debugger_state.clone()) .app_data(health_checker.clone()) .app_data(validate_path_config) .app_data(validate_query_config) .app_data(validate_json_config) .app_data(TempFileConfig::default().directory(&upload_dir)) .app_data(MultipartFormConfig::default().total_limit(usize::MAX)) .app_data(service_config.clone()) .service(index) .configure(config_collections_api) .configure(config_snapshots_api) .configure(config_update_api) .configure(config_cluster_api) .configure(config_service_api) .configure(config_search_api) .configure(config_recommend_api) .configure(config_discovery_api) .configure(config_query_api) .configure(config_facet_api) .configure(config_shards_api) .configure(config_issues_api) .configure(config_debugger_api) .configure(config_profiler_api) .configure(config_local_shard_api) // Ordering of services is important for correct path pattern matching // See: <https://github.com/qdrant/qdrant/issues/3543> .service(scroll_points) .service(count_points) .service(get_point) .service(get_points); if let Some(static_folder) = web_ui_available.as_deref() { app = app.service(web_ui_factory(static_folder)); } app }) .workers(max_web_workers(&settings)); let port = settings.service.http_port; let bind_addr = format!("{}:{}", settings.service.host, port); // With TLS enabled, bind with certificate helper and Rustls, or bind regularly server = if settings.service.enable_tls { log::info!( "TLS enabled for REST API (TTL: {})", settings .tls .as_ref() .and_then(|tls| tls.cert_ttl) .map(|ttl| ttl.to_string()) .unwrap_or_else(|| "none".into()), ); let config = certificate_helpers::actix_tls_server_config(&settings) .map_err(io::Error::other)?; server.bind_rustls_0_23(bind_addr, config)? } else { log::info!("TLS disabled for REST API"); server.bind(bind_addr)? }; log::info!("Qdrant HTTP listening on {port}"); server.run().await }) } fn validation_error_handler( name: &str, err: actix_web_validator::Error, _req: &HttpRequest, ) -> error::Error { use actix_web_validator::error::DeserializeErrors; // Nicely describe deserialization and validation errors let msg = match &err { actix_web_validator::Error::Validate(errs) => { validation::label_errors(format!("Validation error in {name}"), errs) } actix_web_validator::Error::Deserialize(err) => { format!( "Deserialize error in {name}: {}", match err { DeserializeErrors::DeserializeQuery(err) => err.to_string(), DeserializeErrors::DeserializeJson(err) => err.to_string(), DeserializeErrors::DeserializePath(err) => err.to_string(), } ) } actix_web_validator::Error::JsonPayloadError( actix_web::error::JsonPayloadError::Deserialize(err), ) => { format!("Format error in {name}: {err}",) } err => err.to_string(), }; // Build fitting response let response = match &err { actix_web_validator::Error::Validate(_) => HttpResponse::UnprocessableEntity(), _ => HttpResponse::BadRequest(), } .json(ApiResponse::<()> { result: None, status: ApiStatus::Error(msg), time: 0.0, usage: None, }); error::InternalError::from_response(err, response).into() } #[cfg(test)] mod tests { use ::api::grpc::api_crate_version; #[test] fn test_version() { assert_eq!( api_crate_version(), env!("CARGO_PKG_VERSION"), "Qdrant and lib/api crate versions are not same" ); } }
rust
Apache-2.0
f937d0260ffd7705c6f34d9c25da1e27ebf5ddfd
2026-01-04T15:34:51.524868Z
false