# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import glob import json import os import sys from pathlib import Path from typing import List, Union import h5py import numpy as np import yaml # from isaaclab.utils.datasets import episode_data def find_usd_files( path: Union[str, Path], extension: str = None, recursive: bool = True, exclude_underscore_dirs: bool = True, exclude_materials: bool = True ) -> List[Path]: """ Find USD files in a directory. Args: path: File or directory path to search extension: Specific extension to find (e.g., '.usd', '.usda'). If None, finds all USD types (.usd, .usda, .usdc, .usdz) recursive: Search subdirectories recursively exclude_underscore_dirs: Skip directories starting with '_' exclude_materials: Skip 'materials' directories Returns: Sorted list of Path objects to USD files """ path = Path(path) if path.is_file(): if extension is None or path.suffix.lower() == extension.lower(): return [path] return [] if not path.is_dir(): return [] # Determine extensions to search if extension: extensions = [extension] else: extensions = ['.usd', '.usda', '.usdc', '.usdz'] usd_files = [] for ext in extensions: pattern = f"**/*{ext}" if recursive else f"*{ext}" for file_path in path.glob(pattern): if not file_path.is_file(): continue # Get relative path parts (excluding filename) try: rel_parts = file_path.relative_to(path).parts[:-1] except ValueError: rel_parts = () # Apply exclusion filters if exclude_underscore_dirs and any(part.startswith('_') for part in rel_parts): continue if exclude_materials and any(part.lower() == 'materials' for part in rel_parts): continue usd_files.append(file_path) return sorted(set(usd_files)) # Helper function to convert file paths based on path_type def convert_file_path(filepath: str, path_type: str, output_dir: str = None) -> str: """Convert file paths based on the specified path_type.""" if not isinstance(filepath, str): return str(filepath) # Check if this looks like a file path (absolute or with file extension) is_file_path = (os.path.isabs(filepath) or any(filepath.endswith(ext) for ext in ['.py', '.usd', '.usda', '.usdc', '.usdz', '.json', '.csv', '.md', '.txt', '.yaml', '.yml'])) if not is_file_path: return filepath if path_type == "absolute": # Keep as-is return filepath elif path_type == "relative": # Convert absolute paths to relative paths if os.path.isabs(filepath): try: relative_path = os.path.relpath(filepath, output_dir) # Normalize path separators for markdown (use forward slashes) return relative_path.replace(os.sep, '/') except ValueError: # If conversion fails (e.g., different drives on Windows), return original return filepath return filepath elif path_type == "filename_only": # Extract just the filename return os.path.basename(filepath) else: # Invalid path_type, return as-is return filepath def get_relative_path(input_path: str, relative_path: str) -> str: """Get the relative path of a file from a search directory to a relative directory.""" # Get the absolute path of the search directory search_path = os.path.abspath(input_path) # Get the absolute path of the relative directory relative_path = os.path.abspath(relative_path) # Get the relative path of the file from the search directory to the relative directory return os.path.relpath(search_path, relative_path) def load_hdf5_episode(filepath: str, episode: int) -> dict: """Load episode from an HDF5 file.""" # Check if file exists if not os.path.exists(filepath): raise FileNotFoundError(f"File not found: {filepath}") with h5py.File(filepath, 'r') as f: data = f.get('data') if data is not None: ep = data.get(f'demo_{episode}') # Copy all sub-datasets into a dict of numpy arrays episode_data = {} for name, ds in ep.items(): # ds[()] reads the full dataset into memory episode_data[name] = ds[()] else: raise ValueError(f"Episode {episode} not found in {filepath}") return episode_data def load_hdf5_episode_data(filepath: str, episode: int, key: str) -> np.ndarray: """Load actions from an HDF5 file.""" with h5py.File(filepath, 'r') as f: data = f.get('data') ep = data.get(f'demo_{episode}') if ep is not None: actions = ep.get(key) # Only copy "actions" into an np array actions = ep.get(key)[()] else: raise ValueError(f"Episode {episode} not found in {filepath}") return actions def load_param_file(file_path: str, parent_dir: str = None) -> dict: file_path = validate_file_path(file_path, parent_dir) if file_path.endswith(".yaml"): with open(file_path, "r") as file: data = yaml.safe_load(file) if data is None: raise ValueError(f"File '{file_path}' is invalid") elif file_path.endswith(".json"): with open(file_path, "r") as f: data = json.load(f) else: raise ValueError( f"File '{file_path}' is an unsupported type; supported types: [json/yaml]" ) return data def load_file(file_path: str): """Load a file from a path.""" file_path = validate_file_path(file_path) if file_path.endswith(".json"): try: with open(file_path, "r") as f: return json.load(f) except json.JSONDecodeError: raise ValueError(f"File '{file_path}' is not a valid JSON file") elif file_path.endswith(".yaml"): try: with open(file_path, "r") as f: return yaml.safe_load(f) except yaml.YAMLError: raise ValueError(f"File '{file_path}' is not a valid YAML file") else: raise ValueError(f"File '{file_path}' is an unsupported type; supported types: [json/yaml]") def validate_file_extension(file_path: str, ext: str) -> str: """ Ensures that the given filepath has the correct extension. Args: filepath (str): The file path or filename. ext (str): The required extension, with or without a leading dot. Returns: str: The filepath with the correct extension if it was missing or incorrect. """ # Ensure extension starts with a dot if not ext.startswith("."): ext = "." + ext # Extract current file extension base, current_ext = os.path.splitext(file_path) # Check if the extension matches if current_ext.lower() != ext.lower(): return base + ext return file_path def validate_file_path(file_path: str, parent_dir: str = None) -> str: """ Checks if the file exists. If the path is not absolute, prepends the parent directory. Args: file_path (str): The file path to check. parent_dir (str): The parent directory to use if the path is not absolute. Returns: str: The absolute file path if the file exists, otherwise raises an error. """ if not os.path.isabs(file_path): if parent_dir is not None: file_path = os.path.join(parent_dir, file_path) else: raise ValueError( f"File is not an absolute filepath: {file_path} and no parent is specified" ) if os.path.exists(file_path): return os.path.abspath(file_path) else: raise FileNotFoundError(f"File does not exist: {file_path}") def get_filename_without_extension(filepath: str) -> str: """Returns the filename only, without the extension or the full filepath.""" return os.path.basename(filepath).split(".")[0] def find_ext(dir: str, ext: Union[str, List[str]], full_filepath=True, recursive=True): """Given a directory and an extension, return all valid files matching that extension. Args: dir: valid directory path ext (str or list of str): an extension, starting with the dot. full_filepath (bool): If the list of paths should specify the full filepath. recursive (bool): if the directory should be crawled recursively """ # Check that directory is a valid directory if not os.path.isdir(dir): raise FileNotFoundError( f"Directory {dir} does not exist. Did you provide the correct path?" ) # Check if a single extension or a list of extensions are provided if isinstance(ext, str): single_ext = True elif all(isinstance(e, str) for e in ext): single_ext = False else: raise ValueError("Argument 'ext' needs to be of type str or list of str.") # Iterate through directories ext_files = [] iterator = os.walk(dir) if recursive else ((next(os.walk(dir))),) for root, dirs, files in iterator: for file in files: # Check for all the possible extensions. for e in ext if not single_ext else [ext]: if file.lower().endswith(e): filepath = os.path.join(root, file) if full_filepath else file ext_files.append(filepath) return ext_files def check_ext(file: str, ext: Union[str, List[str]]) -> bool: """checks a full filepath is indeed a file and has the correct extension. Returns True if the file is of the type of extension specified. Args: file (str): a full filepath to the file to be examined. ext (str or list of str): acceptable extensions, wit or without the dot. """ path = Path(file) # First check if file is valid if not path.is_file(): raise FileNotFoundError(f"Specified file '{file}' is not a valid file.") # Check if a single extension or a list of extensions are provided if isinstance(ext, str): single_ext = True elif all(isinstance(e, str) for e in ext): single_ext = False else: raise ValueError("Argument 'ext' needs to be of type str or list of str.") # Check ext against a set of possible extensions. for e in ext if not single_ext else [ext]: if e.startswith("."): if path.suffix == e: return True else: if path.suffix == "." + e: return True return False def h5tree(filepath): """Prints a tree of the data in the file.""" with h5py.File(filepath, "r") as hf: print(f"{os.path.basename(filepath)}") h5tree_recurse(hf) def h5tree_recurse(val, pre=""): """The recursive function for printing a data tree. You can call this function directly with the h5py file.""" items = len(val) for key, val in val.items(): items -= 1 if items == 0: # the last item if isinstance(val, h5py._hl.group.Group): print(f"{pre}└── {key}") # print(pre + '└── ' + key) h5tree_recurse(val, pre + " ") else: print(f"{pre}└── {key} {val.shape}") # print(pre + '└── ' + key + ' (%d)' % len(val)) else: if isinstance(val, h5py._hl.group.Group): print(f"{pre}├── {key}") # print(pre + '├── ' + key) h5tree_recurse(val, pre + "│ ") else: print(f"{pre}├── {key} {val.shape}") # print(pre + '├── ' + key + ' (%d)' % len(val)) def save_json(data: dict | list, filepath: str | Path, indent: int = 2) -> None: """Write data to a JSON file. Args: data: Serializable data (dict or list). filepath: Destination path (str or Path). indent: JSON indentation level. """ with open(filepath, "w") as f: json.dump(data, f, indent=indent) def write_dict_to_json(key: str, data_dict: dict, filepath: str): """Write a dictionary to file with a specific key. Note, the heading will be replaced if it already exists. The json file must be a single dictionary with subheadings: { "key1": { "subkey1": value1, "subkey2": value2 }, "key2": [ ... ] } """ data = dict() # Check if file exists. If so, load the existing data from it. if os.path.isfile(filepath): with open(filepath, "r") as json_file: data = json.load(json_file) if not isinstance(data, dict): raise ValueError( "The json file provided must contain a single dictionary." ) # Replace data with key with the current dict. data[key] = data_dict # Write to file. with open(filepath, "w") as json_file: json.dump(data, json_file, indent=2) print(f"'{key}' written to {filepath}") return def get_latest_subdirectory(parent_dir) -> str: """Returns the latest modified subdirectory""" if not os.path.isdir(parent_dir): raise ValueError(f"Parent directory does not exist: {parent_dir}") all_subdirs = [ os.path.join(parent_dir, d) for d in os.listdir(parent_dir) if os.path.isdir(os.path.join(parent_dir, d)) ] latest_subdir = max(all_subdirs, key=os.path.getmtime) return latest_subdir def get_incremented_filepath(filepath): """ Get an incremented filepath if the filepath already exists in that directory. Args: filepath (str): The base filepath (including extension). Returns: str: A unique filepath with an incremented suffix. """ if not os.path.exists(filepath): return filepath base, ext = os.path.splitext(filepath) i = 1 new_filepath = f"{base}_{i}{ext}" while os.path.exists(new_filepath): i += 1 new_filepath = f"{base}_{i}{ext}" return new_filepath def get_list_of_files_with_extension( dir: str, ext: Union[str, List], recursive: bool = False ) -> List[str]: if not os.path.isdir(dir): raise ValueError(f"Directory '{dir}' is not a valid directory") if isinstance(ext, str): ext = [ext] # Make sure all extensions start with a dot ext = [ extension if extension.startswith(".") else "." + extension for extension in ext ] file_paths = [] if recursive: # Walk through the directory and its subdirectories for root, dirs, files in os.walk(dir): for file in files: # Check if the file ends with any of the specified extensions if any(file.endswith(extension) for extension in ext): # Construct the full path and add to the list file_paths.append(os.path.join(root, file)) else: # Only walk through the current level directory folder_items = os.listdir(dir) for file in folder_items: if any(file.endswith(extension) for extension in ext): file_paths.append(os.path.join(dir, file)) return file_paths def expand_folder_patterns( patterns: list[str], base_dir: str | None = None, ) -> tuple[list[str], bool]: """Expand glob patterns to matching directory paths. Each pattern can be a literal path or a glob (e.g., 'pi0_*'). Resolution order for a relative pattern: 1. As-given (CWD-relative) — standard CLI behavior. 2. base_dir-prepended fallback (only if (1) had no matches and base_dir is set) — preserves shorthand like `pi0_*` resolving to `/pi0_*`. Args: patterns: Folder names or glob patterns. base_dir: Optional fallback base directory used only when CWD-relative resolution finds no matches. Returns: (deduplicated list of directory paths, whether any pattern was expanded) """ folders: list[str] = [] pattern_expanded = False for pattern in patterns: candidates: list[str] = [] if os.path.isabs(pattern): candidates.append(pattern) else: candidates.append(pattern) if base_dir is not None: candidates.append(os.path.join(base_dir, pattern)) matches: list[str] = [] full_pattern = candidates[0] for cand in candidates: cand_matches = sorted(glob.glob(cand)) cand_matches = [m for m in cand_matches if os.path.isdir(m)] if cand_matches: full_pattern = cand matches = cand_matches break if matches: if len(matches) > 1 or matches[0] != full_pattern: pattern_expanded = True folders.extend(os.path.abspath(m) for m in matches) else: folders.append(os.path.abspath(full_pattern)) seen: set[str] = set() unique: list[str] = [] for f in folders: real = os.path.realpath(f) if real not in seen: seen.add(real) unique.append(f) return unique, pattern_expanded def confirm_folders( folders: list[str], default_yes: bool = True, ) -> list[str]: """Prompt user to confirm a folder selection. Args: folders: Candidate folder paths. default_yes: When True the default action (Enter) includes all folders; when False the default action aborts. Returns: Confirmed folders, or an empty list if the user aborts. """ if default_yes: prompt = "\nInclude all? [Y/n/s] (s=select individually): " else: prompt = "Proceed with these folders? [y/N/s] (s=select individually): " if not sys.stdin.isatty(): if default_yes: print(f"{prompt}[non-interactive stdin: auto-including all {len(folders)} folder(s)]") return folders else: print(f"{prompt}[non-interactive stdin: aborting]") return [] while True: response = input(prompt).strip().lower() if response == "": if default_yes: return folders else: print("Aborted.") return [] elif response == "y": return folders elif response == "n": print("Aborted.") return [] elif response == "s": confirmed: list[str] = [] for folder in folders: name = os.path.basename(folder) while True: r = input(f" Include '{name}'? [Y/n]: ").strip().lower() if r in ("", "y"): confirmed.append(folder) break elif r == "n": break else: print(" Please enter 'y' or 'n'") if not confirmed: print("No folders selected.") return [] print(f"\nSelected {len(confirmed)} folder(s):") for f in confirmed: print(f" - {os.path.basename(f)}") return confirmed else: print("Please enter 'y', 'n', or 's'") def get_folders_in_dir(dir: str) -> List[str]: if not os.path.isdir(dir): raise ValueError(f"Directory '{dir}' is not a valid directory") folder_items = os.listdir(dir) folders = [ instance for instance in folder_items if os.path.isdir(os.path.join(dir, instance)) ] return folders def get_filename(filepath, without_extension: bool = True) -> str: filename = os.path.basename(filepath) if without_extension: return os.path.splitext(filename)[0] else: return filename def get_filepath_with_extensions(dir: str, filename: str, ext: Union[str, List]): """ Gets a file with names with a set of acceptable extensions. If both are available, it returns the first ext instead. """ if isinstance(ext, str): ext = [ext] # Make sure all extensions start with a dot ext = [ extension if extension.startswith(".") else "." + extension for extension in ext ] # Check if the current file exists, and if so, return the full file path directly if any(filename.endswith(extension) for extension in ext) and os.path.isfile( os.path.join(dir, filename) ): return os.path.join(dir, filename) # strip the extension and search for other files with the same name and same extension filename_wo_ext = get_filename(filename, without_extension=True) for extension in ext: # Check if each file exists path = os.path.join(dir, filename_wo_ext + extension) if os.path.isfile(path): print(f"Found alternative file '{filename_wo_ext + extension}'.") return path err = f"Provided file name {filename} cannot be found in directory {dir} with an valid extension. Valid extensions: {ext}" raise FileNotFoundError(err) def write_class_params_to_file(obj, filepath): import json """ Saves only the values of parameters defined in the __init__ method of the given object's class to a JSON file. Parameters: - obj: The object whose `__init__` parameters will be saved. - filepath: The JSON file to write the parameters to. """ init_values = get_class_params_as_dict(obj) # Write the result to a JSON file with open(filepath, "w") as f: json.dump(init_values, f, indent=4) def get_class_params_as_dict(obj) -> dict: """ Saves only the values of parameters defined in the __init__ method of the given object's class to a JSON file. Parameters: - obj: The object whose `__init__` parameters will be saved. - filepath: The JSON file to write the parameters to. """ import inspect # Get the `__init__` method of the class init_method = getattr(obj.__class__, "__init__", None) if not init_method: raise ValueError( f"The class {obj.__class__.__name__} does not have an __init__ method." ) # Get the parameter names from the `__init__` method, excluding 'self' init_params = inspect.signature(init_method).parameters init_param_names = [param for param in init_params if param != "self"] # Filter instance attributes based on the `__init__` parameters # init_values = {key: getattr(obj, key) for key in init_param_names if hasattr(obj, key)} init_values = {} for key in init_param_names: if hasattr(obj, key): value = getattr(obj, key) try: # Test if the value is serializable json.dumps(value) init_values[key] = value except (TypeError, ValueError): # Skip unserializable values continue return init_values