|
| 1 | +# coding=utf-8 |
| 2 | +# Copyright 2024 The HuggingFace Inc. team. |
| 3 | +# |
| 4 | +# Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | +# you may not use this file except in compliance with the License. |
| 6 | +# You may obtain a copy of the License at |
| 7 | +# |
| 8 | +# http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | +# |
| 10 | +""" |
| 11 | +Utility that checks that modules like attention processors are listed in the documentation file. |
| 12 | +
|
| 13 | +```bash |
| 14 | +python utils/check_support_list.py |
| 15 | +``` |
| 16 | +
|
| 17 | +It has no auto-fix mode. |
| 18 | +""" |
| 19 | + |
| 20 | +import os |
| 21 | +import re |
| 22 | + |
| 23 | + |
| 24 | +# All paths are set with the intent that you run this script from the root of the repo |
| 25 | +REPO_PATH = "." |
| 26 | + |
| 27 | + |
| 28 | +def read_documented_classes(doc_path, autodoc_regex=r"\[\[autodoc\]\]\s([^\n]+)"): |
| 29 | + """ |
| 30 | + Reads documented classes from a doc file using a regex to find lines like [[autodoc]] my.module.Class. |
| 31 | + Returns a list of documented class names (just the class name portion). |
| 32 | + """ |
| 33 | + with open(os.path.join(REPO_PATH, doc_path), "r") as f: |
| 34 | + doctext = f.read() |
| 35 | + matches = re.findall(autodoc_regex, doctext) |
| 36 | + return [match.split(".")[-1] for match in matches] |
| 37 | + |
| 38 | + |
| 39 | +def read_source_classes(src_path, class_regex, exclude_conditions=None): |
| 40 | + """ |
| 41 | + Reads class names from a source file using a regex that captures class definitions. |
| 42 | + Optionally exclude classes based on a list of conditions (functions that take class name and return bool). |
| 43 | + """ |
| 44 | + if exclude_conditions is None: |
| 45 | + exclude_conditions = [] |
| 46 | + with open(os.path.join(REPO_PATH, src_path), "r") as f: |
| 47 | + doctext = f.read() |
| 48 | + classes = re.findall(class_regex, doctext) |
| 49 | + # Filter out classes that meet any of the exclude conditions |
| 50 | + filtered_classes = [c for c in classes if not any(cond(c) for cond in exclude_conditions)] |
| 51 | + return filtered_classes |
| 52 | + |
| 53 | + |
| 54 | +def check_documentation(doc_path, src_path, doc_regex, src_regex, exclude_conditions=None): |
| 55 | + """ |
| 56 | + Generic function to check if all classes defined in `src_path` are documented in `doc_path`. |
| 57 | + Returns a set of undocumented class names. |
| 58 | + """ |
| 59 | + documented = set(read_documented_classes(doc_path, doc_regex)) |
| 60 | + source_classes = set(read_source_classes(src_path, src_regex, exclude_conditions=exclude_conditions)) |
| 61 | + |
| 62 | + # Find which classes in source are not documented in a deterministic way. |
| 63 | + undocumented = sorted(source_classes - documented) |
| 64 | + return undocumented |
| 65 | + |
| 66 | + |
| 67 | +if __name__ == "__main__": |
| 68 | + # Define the checks we need to perform |
| 69 | + checks = { |
| 70 | + "Attention Processors": { |
| 71 | + "doc_path": "docs/source/en/api/attnprocessor.md", |
| 72 | + "src_path": "src/diffusers/models/attention_processor.py", |
| 73 | + "doc_regex": r"\[\[autodoc\]\]\s([^\n]+)", |
| 74 | + "src_regex": r"class\s+(\w+Processor(?:\d*_?\d*))[:(]", |
| 75 | + "exclude_conditions": [lambda c: "LoRA" in c, lambda c: c == "Attention"], |
| 76 | + }, |
| 77 | + "Image Processors": { |
| 78 | + "doc_path": "docs/source/en/api/image_processor.md", |
| 79 | + "src_path": "src/diffusers/image_processor.py", |
| 80 | + "doc_regex": r"\[\[autodoc\]\]\s([^\n]+)", |
| 81 | + "src_regex": r"class\s+(\w+Processor(?:\d*_?\d*))[:(]", |
| 82 | + }, |
| 83 | + "Activations": { |
| 84 | + "doc_path": "docs/source/en/api/activations.md", |
| 85 | + "src_path": "src/diffusers/models/activations.py", |
| 86 | + "doc_regex": r"\[\[autodoc\]\]\s([^\n]+)", |
| 87 | + "src_regex": r"class\s+(\w+)\s*\(.*?nn\.Module.*?\):", |
| 88 | + }, |
| 89 | + "Normalizations": { |
| 90 | + "doc_path": "docs/source/en/api/normalization.md", |
| 91 | + "src_path": "src/diffusers/models/normalization.py", |
| 92 | + "doc_regex": r"\[\[autodoc\]\]\s([^\n]+)", |
| 93 | + "src_regex": r"class\s+(\w+)\s*\(.*?nn\.Module.*?\):", |
| 94 | + "exclude_conditions": [ |
| 95 | + # Exclude LayerNorm as it's an intentional exception |
| 96 | + lambda c: c == "LayerNorm" |
| 97 | + ], |
| 98 | + }, |
| 99 | + "LoRA Mixins": { |
| 100 | + "doc_path": "docs/source/en/api/loaders/lora.md", |
| 101 | + "src_path": "src/diffusers/loaders/lora_pipeline.py", |
| 102 | + "doc_regex": r"\[\[autodoc\]\]\s([^\n]+)", |
| 103 | + "src_regex": r"class\s+(\w+)\s*\(.*?nn\.Module.*?\):", |
| 104 | + }, |
| 105 | + } |
| 106 | + |
| 107 | + missing_items = {} |
| 108 | + for category, params in checks.items(): |
| 109 | + undocumented = check_documentation( |
| 110 | + doc_path=params["doc_path"], |
| 111 | + src_path=params["src_path"], |
| 112 | + doc_regex=params["doc_regex"], |
| 113 | + src_regex=params["src_regex"], |
| 114 | + exclude_conditions=params.get("exclude_conditions"), |
| 115 | + ) |
| 116 | + if undocumented: |
| 117 | + missing_items[category] = undocumented |
| 118 | + |
| 119 | + # If we have any missing items, raise a single combined error |
| 120 | + if missing_items: |
| 121 | + error_msg = ["Some classes are not documented properly:\n"] |
| 122 | + for category, classes in missing_items.items(): |
| 123 | + error_msg.append(f"- {category}: {', '.join(sorted(classes))}") |
| 124 | + raise ValueError("\n".join(error_msg)) |
0 commit comments