peijun1's picture
Deploy AI Studio Proxy API to Hugging Face Spaces
a5784e9
Raw
History Blame Contribute Delete
24 kB
"""
Rendering Logic for Grid Logger
"""
import logging
import re
import sys
import threading
import time
from datetime import datetime
from typing import Any, List, Optional, Tuple
from colorama import Fore, Style
from .constants import SOURCE_MAP, Colors, Columns
from .context import (
request_id_var,
source_var,
)
def normalize_source(source: str) -> str:
"""Normalize source name to fixed 5-letter code."""
key = source.lower().replace(" ", "_").replace("-", "_")
if key in SOURCE_MAP:
return SOURCE_MAP[key]
# Try partial match
for map_key, map_val in SOURCE_MAP.items():
if map_key in key or key in map_key:
return map_val
# Default: first 5 chars, uppercase, padded
return source[:5].upper().ljust(5)
# =============================================================================
# Semantic Highlighter (Enhanced)
# =============================================================================
class SemanticHighlighter:
"""Applies semantic coloring to log message content."""
# Compiled regex patterns for efficiency
TAG_PATTERN = re.compile(r"^\[([A-Z]{2,10})\]\s*")
STRING_PATTERN = re.compile(r"'([^']*)'|\"([^\"]*)\"")
BOOLEAN_PATTERN = re.compile(r"\b(True|False|None)\b")
NUMBER_PATTERN = re.compile(r"\b(\d+\.?\d*(?:ms|s|kb|mb|gb|KB|MB|GB|%|px)?)\b")
HEX_PATTERN = re.compile(r"\b(0x[0-9a-fA-F]+)\b")
URL_PATTERN = re.compile(
r"(https?://[^\s]+|[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}(?::\d+)?(?:/[^\s]*)?)"
)
# Key phrases to bold
PHRASE_ERROR = re.compile(r"\b(Error|ERROR|error)\b")
PHRASE_FAILED = re.compile(r"\b(Failed|FAILED|failed|Failure|failure)\b")
PHRASE_SUCCESS = re.compile(
r"\b(Success|SUCCESS|success|Successful|successful|Complete|complete|Completed|completed)\b"
)
PHRASE_WARNING = re.compile(r"\b(Warning|WARNING|warning)\b")
# Status phrases - Success patterns (Green)
PHRASE_MATCHES = re.compile(r"\((Matches page|Cached|Matches)\)")
# Status phrases - Action patterns (Yellow)
PHRASE_UPDATING = re.compile(r"\((Updating|Toggling|Loading)\.\.\.\)")
# Model IDs (gemini-*, claude-*, etc.)
MODEL_ID_PATTERN = re.compile(
r"\b(gemini-[\w.-]+|claude-[\w.-]+|gpt-[\w.-]+|veo-[\w.-]+)\b"
)
# Request IDs (7-char alphanumeric)
REQ_ID_PATTERN = re.compile(r"\b([a-z0-9]{7})\b")
@classmethod
def highlight(cls, text: str, colorize: bool = True) -> str:
"""Apply semantic highlighting to text."""
if not colorize:
return text
result = text
# Handle tags at the start (e.g., [UI], [NET], [SYS])
tag_match = cls.TAG_PATTERN.match(result)
if tag_match:
full_tag = tag_match.group(0)
tag_name = tag_match.group(1)
# Use source color if available, otherwise use TAG color
normalized = normalize_source(tag_name)
tag_color = Colors.SOURCES.get(normalized, Colors.TAG)
colored_tag = f"{tag_color}[{tag_name}]{Colors.RESET} "
result = colored_tag + result[len(full_tag) :]
# Highlight URLs first (before strings, to avoid conflict)
def replace_url(match: re.Match[str]) -> str:
return f"{Colors.URL}{Style.DIM}{match.group(1)}{Colors.RESET}"
result = cls.URL_PATTERN.sub(replace_url, result)
# Highlight model IDs
def replace_model_id(match: re.Match[str]) -> str:
return f"{Colors.STRING}{match.group(1)}{Colors.RESET}"
result = cls.MODEL_ID_PATTERN.sub(replace_model_id, result)
# Highlight quoted strings
def replace_string(match: re.Match[str]) -> str:
# Match group 1 is single-quote, group 2 is double-quote
content = match.group(1) if match.group(1) else match.group(2)
quote = "'" if match.group(1) else '"'
return f"{Colors.STRING}{quote}{content}{quote}{Colors.RESET}"
result = cls.STRING_PATTERN.sub(replace_string, result)
# Highlight booleans with distinct colors
def replace_boolean(match: re.Match[str]) -> str:
val = match.group(1)
if val == "True":
return f"{Colors.BOOLEAN_TRUE}{val}{Colors.RESET}"
elif val == "False":
return f"{Colors.BOOLEAN_FALSE}{val}{Colors.RESET}"
else: # None
return f"{Colors.BOOLEAN_NONE}{val}{Colors.RESET}"
result = cls.BOOLEAN_PATTERN.sub(replace_boolean, result)
# Highlight numbers (avoid matching inside ANSI codes)
def replace_number(match: re.Match[str]) -> str:
return f"{Colors.NUMBER}{match.group(1)}{Colors.RESET}"
result = cls.NUMBER_PATTERN.sub(replace_number, result)
# Highlight hex numbers
def replace_hex(match: re.Match[str]) -> str:
return f"{Colors.NUMBER}{match.group(1)}{Colors.RESET}"
result = cls.HEX_PATTERN.sub(replace_hex, result)
# Bold key phrases
def replace_error(match: re.Match[str]) -> str:
return f"{Colors.PHRASE_ERROR}{match.group(1)}{Colors.RESET}"
result = cls.PHRASE_ERROR.sub(replace_error, result)
def replace_failed(match: re.Match[str]) -> str:
return f"{Colors.PHRASE_FAILED}{match.group(1)}{Colors.RESET}"
result = cls.PHRASE_FAILED.sub(replace_failed, result)
def replace_success(match: re.Match[str]) -> str:
return f"{Colors.PHRASE_SUCCESS}{match.group(1)}{Colors.RESET}"
result = cls.PHRASE_SUCCESS.sub(replace_success, result)
def replace_warning(match: re.Match[str]) -> str:
return f"{Colors.PHRASE_WARNING}{match.group(1)}{Colors.RESET}"
result = cls.PHRASE_WARNING.sub(replace_warning, result)
# Highlight status phrases - Success (Green)
def replace_matches(match: re.Match[str]) -> str:
return f"{Colors.PHRASE_SUCCESS}{match.group(0)}{Colors.RESET}"
result = cls.PHRASE_MATCHES.sub(replace_matches, result)
# Highlight status phrases - Action (Yellow)
def replace_updating(match: re.Match[str]) -> str:
return f"{Colors.PHRASE_WARNING}{match.group(0)}{Colors.RESET}"
result = cls.PHRASE_UPDATING.sub(replace_updating, result)
return result
# =============================================================================
# Burst Suppression State (Thread-safe)
# =============================================================================
class BurstBuffer:
"""
Thread-safe buffer for burst suppression.
Tracks duplicate log messages and suppresses them, appending
a count like (x3) when a different message arrives.
"""
def __init__(self):
self._lock = threading.Lock()
self._last_key: Optional[str] = None
self._last_formatted: Optional[str] = None
self._count: int = 0
def process(
self, key: str, formatted_line: str
) -> Tuple[Optional[str], Optional[str]]:
"""
Process a log entry for burst suppression.
Args:
key: Unique key for the log (source + message)
formatted_line: The fully formatted log line
Returns:
Tuple of (line_to_print_now, deferred_line_if_any)
- If same as last: returns (None, None) - suppressed
- If different: returns (last_with_count_if_needed, current_line)
"""
with self._lock:
if self._last_key is None:
# First message
self._last_key = key
self._last_formatted = formatted_line
self._count = 1
return (formatted_line, None)
if key == self._last_key:
# Duplicate - suppress and increment counter
self._count += 1
return (None, None)
else:
# Different message - flush previous with count
prev_line = self._last_formatted
prev_count = self._count
# Update to new message
self._last_key = key
self._last_formatted = formatted_line
self._count = 1
if prev_count > 1 and prev_line:
# Create a dim footnote line instead of appending count
# Extract timestamp/level/source/id columns from prev_line to align properly
# Format: " └─ (Repeated N times)"
indent = (
" " * 31
) # Align with message column (timestamp + level + source + id + spaces)
count_line = f"{indent}{Colors.TREE}\u2514\u2500 {Colors.BURST_COUNT}(Repeated {prev_count} times){Colors.RESET}"
# Return previous line (without count), count line, and current line
return (f"{prev_line}\n{count_line}", formatted_line)
else:
# Previous was single, just print current
return (formatted_line, None)
def flush(self) -> Optional[str]:
"""Flush any remaining buffered message."""
with self._lock:
if self._last_formatted and self._count > 1:
# Create footnote line for repeated message
indent = " " * 31
count_line = f"{indent}{Colors.TREE}\u2514\u2500 {Colors.BURST_COUNT}(Repeated {self._count} times){Colors.RESET}"
result = f"{self._last_formatted}\n{count_line}"
self._last_key = None
self._last_formatted = None
self._count = 0
return result
# Don't return anything if count <= 1 (no repetition to report)
self._last_key = None
self._last_formatted = None
self._count = 0
return None
# Global burst buffer instance
_burst_buffer = BurstBuffer()
# =============================================================================
# Grid Formatter
# =============================================================================
class GridFormatter(logging.Formatter):
"""
Fixed-width grid formatter with semantic highlighting and burst suppression.
Format: TIME | LVL | SOURCE | ID | MESSAGE
Example output:
22:47:51.690 INF API y74ebn9 Received /v1/chat/completions request
22:47:51.692 INF WORKR y74ebn9 Processing request logic...
22:47:51.695 INF PROXY Sniff HTTPS requests (x5)
"""
def __init__(
self,
colorize: bool = True,
burst_suppression: bool = True,
show_tree: bool = True, # Deprecated, kept for compatibility
):
super().__init__()
self.colorize = colorize
self.burst_suppression = burst_suppression
def format(self, record: logging.LogRecord) -> str:
"""Format log record into grid layout."""
# Skip during Python shutdown to avoid ImportError
if sys.meta_path is None:
return record.getMessage()
# Extract context variables with defaults
try:
req_id = request_id_var.get()
except LookupError:
req_id = " "
try:
source = source_var.get()
except LookupError:
source = "SYS"
# Normalize source to 5-letter code
source_normalized = normalize_source(source)
# Column 1: Time (HH:MM:SS.mmm) - no date
now = datetime.now()
timestamp = now.strftime("%H:%M:%S.") + f"{int(now.microsecond / 1000):03d}"
if self.colorize:
time_col = f"{Colors.TIME}{timestamp}{Colors.RESET}"
else:
time_col = timestamp
# Column 2: Level (3 chars)
level_abbrev = Colors.LEVEL_ABBREV.get(
record.levelname, record.levelname[:3].upper()
)
if self.colorize:
level_color = Colors.LEVELS.get(record.levelname, Fore.WHITE)
level_col = f"{level_color}{level_abbrev}{Colors.RESET}"
else:
level_col = level_abbrev
# Column 3: Source (fixed 5-char width)
if self.colorize:
source_color = Colors.SOURCES.get(source_normalized, Fore.WHITE)
source_col = f"{source_color}{source_normalized}{Colors.RESET}"
else:
source_col = source_normalized
# Column 4: Request ID (fixed 7-char width)
id_display = req_id[: Columns.ID].ljust(Columns.ID)
if self.colorize:
id_col = f"{Colors.REQUEST_ID}{id_display}{Colors.RESET}"
else:
id_col = id_display
# Column 6: Message with semantic highlighting
message = record.getMessage()
# Skip separator lines (dashed lines)
if message.strip().startswith("---") or message.strip().startswith("==="):
# Skip separator lines entirely
return ""
if self.colorize:
message = SemanticHighlighter.highlight(message)
# Combine all columns
line = f"{time_col} {level_col} {source_col} {id_col} {message}"
# Apply burst suppression
if self.burst_suppression:
# Create key from source + original message (pre-highlight)
burst_key = f"{source_normalized}:{record.getMessage()}"
prev_line, current_line = _burst_buffer.process(burst_key, line)
if prev_line is None and current_line is None:
# Suppressed - return empty string (won't be printed)
return ""
elif current_line is None:
# Just the previous (with count if needed)
return prev_line or ""
elif prev_line:
# Both lines - print previous then current
return f"{prev_line}\n{current_line}"
else:
# Just current
return current_line
return line
# =============================================================================
# Plain Formatter (for file/WebSocket - no ANSI codes)
# =============================================================================
class PlainGridFormatter(logging.Formatter):
"""Plain-text grid formatter for file/WebSocket logging (no ANSI codes)."""
def format(self, record: logging.LogRecord) -> str:
"""Format log record into plain text grid layout."""
try:
req_id = request_id_var.get()
except LookupError:
req_id = " "
try:
source = source_var.get()
except LookupError:
source = "SYS"
source_normalized = normalize_source(source)
now = datetime.now()
timestamp = now.strftime("%H:%M:%S.") + f"{int(now.microsecond / 1000):03d}"
level_abbrev = Colors.LEVEL_ABBREV.get(
record.levelname, record.levelname[:3].upper()
)
id_display = req_id[: Columns.ID].ljust(Columns.ID)
message = record.getMessage()
# Skip separator lines
if message.strip().startswith("---") or message.strip().startswith("==="):
return ""
return f"{timestamp} {level_abbrev} {source_normalized} {id_display} {message}"
# =============================================================================
# Object Dumper (YAML-style)
# =============================================================================
def format_object(obj: Any, indent: int = 0, colorize: bool = True) -> str:
"""
Format an object in YAML-style for clean logging.
Args:
obj: The object to format (dict, list, or primitive)
indent: Current indentation level
colorize: Whether to apply colors
Returns:
Formatted string representation
"""
lines: List[str] = []
prefix = " " * indent
if isinstance(obj, dict):
for key, value in obj.items(): # pyright: ignore[reportUnknownVariableType]
key_str = (
f"{Colors.KEY}{key}{Colors.RESET}" if colorize else str(key) # pyright: ignore[reportUnknownArgumentType]
)
if isinstance(value, (dict, list)) and value:
lines.append(f"{prefix}{key_str}:")
lines.append(format_object(value, indent + 1, colorize))
else:
formatted_value = _format_value(value, colorize)
lines.append(f"{prefix}{key_str}: {formatted_value}")
elif isinstance(obj, list):
for item in obj: # pyright: ignore[reportUnknownVariableType]
if isinstance(item, (dict, list)) and item:
lines.append(f"{prefix}-")
lines.append(format_object(item, indent + 1, colorize))
else:
formatted_value = _format_value(item, colorize)
lines.append(f"{prefix}- {formatted_value}")
else:
lines.append(f"{prefix}{_format_value(obj, colorize)}")
return "\n".join(lines)
def _format_value(value: Any, colorize: bool = True) -> str:
"""Format a single value with appropriate coloring."""
if not colorize:
if isinstance(value, str):
return f"'{value}'"
return str(value)
if isinstance(value, bool):
if value:
return f"{Colors.BOOLEAN_TRUE}{value}{Colors.RESET}"
else:
return f"{Colors.BOOLEAN_FALSE}{value}{Colors.RESET}"
elif isinstance(value, (int, float)):
return f"{Colors.NUMBER}{value}{Colors.RESET}"
elif isinstance(value, str):
# Truncate long strings
if len(value) > 50:
truncated = value[:47] + "..."
return f"{Colors.STRING}'{truncated}'{Colors.RESET}"
return f"{Colors.STRING}'{value}'{Colors.RESET}"
elif value is None:
return f"{Colors.BOOLEAN_NONE}None{Colors.RESET}"
else:
return f"{Colors.STRING}{repr(value)}{Colors.RESET}"
# =============================================================================
# Progress Indicator
# =============================================================================
class ProgressLine:
"""
Progress indicator that updates on the same line.
Uses carriage return (\\r) to overwrite the previous output
instead of creating new log lines.
Usage:
progress = ProgressLine("Waiting for stream")
for i in range(100):
progress.update(i + 1, 100, f"chunks: {i + 1}")
time.sleep(0.05)
progress.finish("Complete")
"""
def __init__(self, message: str, source: Optional[str] = None):
"""
Initialize progress indicator.
Args:
message: Base message to display
source: Optional source override (uses context var if None)
"""
self.message = message
self.source = source
self.last_update = 0.0
self._started = False
self._min_interval = 0.05 # Minimum time between updates (50ms)
def update(self, current: int, total: int, extra: str = "") -> None:
"""
Update progress on the same line.
Args:
current: Current progress value
total: Total value for 100%
extra: Additional info to display after progress bar
"""
now = time.time()
# Rate limit updates to avoid flickering
if now - self.last_update < self._min_interval and current < total:
return
self.last_update = now
self._started = True
# Get context
try:
req_id = request_id_var.get()
except LookupError:
req_id = " "
source = self.source
if source is None:
try:
source = source_var.get()
except LookupError:
source = "SYS"
source_normalized = normalize_source(source)
# Build timestamp
timestamp = (
datetime.now().strftime("%H:%M:%S.") + f"{int(now * 1000) % 1000:03d}"
)
# Calculate progress
percentage = (current / total * 100) if total > 0 else 0
bar_width = 20
filled = int(bar_width * current / total) if total > 0 else 0
bar = "#" * filled + "-" * (bar_width - filled)
extra_text = f" {extra}" if extra else ""
# Build the line with colors
source_color = Colors.SOURCES.get(source_normalized, Fore.WHITE)
line = (
f"\r{Colors.TIME}{timestamp}{Colors.RESET} "
f"{Colors.LEVELS['INFO']}INF{Colors.RESET} "
f"{source_color}{source_normalized}{Colors.RESET} "
f"{Colors.REQUEST_ID}{req_id:<{Columns.ID}}{Colors.RESET} "
f"{Colors.MESSAGE}{self.message} [{bar}] "
f"{Colors.NUMBER}{current}{Colors.RESET}/{Colors.NUMBER}{total}{Colors.RESET} "
f"({Colors.NUMBER}{percentage:.0f}%{Colors.RESET}){extra_text}"
)
# Clear to end of line and print
sys.stdout.write(f"{line}\033[K")
sys.stdout.flush()
def finish(self, message: Optional[str] = None) -> None:
"""
Complete the progress and move to new line.
Args:
message: Optional completion message
"""
if self._started:
if message:
sys.stdout.write(f" - {Colors.STRING}{message}{Colors.RESET}")
sys.stdout.write("\n")
sys.stdout.flush()
# =============================================================================
# JSON Formatter (Structured Logging)
# =============================================================================
class JSONFormatter(logging.Formatter):
"""
JSON formatter for structured logging.
Outputs logs as JSON lines for easy parsing by log aggregation tools
like ELK, Datadog, CloudWatch, etc.
Enable via JSON_LOGS=true environment variable.
Format:
{"timestamp": "2024-12-15T15:30:00.123Z", "level": "ERROR", "source": "API",
"request_id": "abc1234", "message": "...", "exception": "..."}
"""
def format(self, record: logging.LogRecord) -> str:
"""Format log record as JSON."""
import json
from datetime import timezone
# Get context variables
try:
req_id = request_id_var.get()
except LookupError:
req_id = ""
try:
source = source_var.get()
except LookupError:
source = "SYS"
source_normalized = normalize_source(source)
# Build timestamp in ISO 8601 format with milliseconds
now = datetime.now(timezone.utc)
timestamp = (
now.strftime("%Y-%m-%dT%H:%M:%S.") + f"{int(now.microsecond / 1000):03d}Z"
)
# Base log entry
log_entry: dict[str, Any] = {
"timestamp": timestamp,
"level": record.levelname,
"source": source_normalized.strip(),
"message": record.getMessage(),
}
# Add request ID if present
if req_id and req_id.strip():
log_entry["request_id"] = req_id.strip()
# Add logger name
if record.name:
log_entry["logger"] = record.name
# Add exception info if present
if record.exc_info and record.exc_info[1] is not None:
import traceback as tb
log_entry["exception"] = {
"type": type(record.exc_info[1]).__name__,
"message": str(record.exc_info[1]),
"traceback": "".join(tb.format_exception(*record.exc_info)),
}
# Add extra fields from record
# Common extra fields that might be useful
for key in ("funcName", "lineno", "pathname"):
value = getattr(record, key, None)
if value:
log_entry[key] = value
return json.dumps(log_entry, ensure_ascii=False, default=str)