chendl/compositional_test
1
1# coding=utf-82# Copyright 2023 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 at7#8# http://www.apache.org/licenses/LICENSE-2.09#10# Unless required by applicable law or agreed to in writing, software11# distributed under the License is distributed on an "AS IS" BASIS,12# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.13# See the License for the specific language governing permissions and14# limitations under the License.15 16import argparse17import os18 19from transformers.utils import direct_transformers_import20 21 22# All paths are set with the intent you should run this script from the root of the repo with the command23# python utils/check_task_guides.py24TRANSFORMERS_PATH = "src/transformers"25PATH_TO_TASK_GUIDES = "docs/source/en/tasks"26 27 28def _find_text_in_file(filename, start_prompt, end_prompt):29 """30 Find the text in `filename` between a line beginning with `start_prompt` and before `end_prompt`, removing empty31 lines.32 """33 with open(filename, "r", encoding="utf-8", newline="\n") as f:34 lines = f.readlines()35 # Find the start prompt.36 start_index = 037 while not lines[start_index].startswith(start_prompt):38 start_index += 139 start_index += 140 41 end_index = start_index42 while not lines[end_index].startswith(end_prompt):43 end_index += 144 end_index -= 145 46 while len(lines[start_index]) <= 1:47 start_index += 148 while len(lines[end_index]) <= 1:49 end_index -= 150 end_index += 151 return "".join(lines[start_index:end_index]), start_index, end_index, lines52 53 54# This is to make sure the transformers module imported is the one in the repo.55transformers_module = direct_transformers_import(TRANSFORMERS_PATH)56 57TASK_GUIDE_TO_MODELS = {58 "asr.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_CTC_MAPPING_NAMES,59 "audio_classification.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES,60 "language_modeling.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_CAUSAL_LM_MAPPING_NAMES,61 "image_classification.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES,62 "masked_language_modeling.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_MASKED_LM_MAPPING_NAMES,63 "multiple_choice.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES,64 "object_detection.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES,65 "question_answering.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES,66 "semantic_segmentation.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_SEMANTIC_SEGMENTATION_MAPPING_NAMES,67 "sequence_classification.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES,68 "summarization.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES,69 "token_classification.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES,70 "translation.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES,71 "video_classification.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES,72 "document_question_answering.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES,73 "monocular_depth_estimation.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_DEPTH_ESTIMATION_MAPPING_NAMES,74}75 76# This list contains model types used in some task guides that are not in `CONFIG_MAPPING_NAMES` (therefore not in any77# `MODEL_MAPPING_NAMES` or any `MODEL_FOR_XXX_MAPPING_NAMES`).78SPECIAL_TASK_GUIDE_TO_MODEL_TYPES = {79 "summarization.mdx": ("nllb",),80 "translation.mdx": ("nllb",),81}82 83 84def get_model_list_for_task(task_guide):85 """86 Return the list of models supporting given task.87 """88 model_maping_names = TASK_GUIDE_TO_MODELS[task_guide]89 special_model_types = SPECIAL_TASK_GUIDE_TO_MODEL_TYPES.get(task_guide, set())90 model_names = {91 code: name92 for code, name in transformers_module.MODEL_NAMES_MAPPING.items()93 if (code in model_maping_names or code in special_model_types)94 }95 return ", ".join([f"[{name}](../model_doc/{code})" for code, name in model_names.items()]) + "\n"96 97 98def check_model_list_for_task(task_guide, overwrite=False):99 """For a given task guide, checks the model list in the generated tip for consistency with the state of the lib and overwrites if needed."""100 101 current_list, start_index, end_index, lines = _find_text_in_file(102 filename=os.path.join(PATH_TO_TASK_GUIDES, task_guide),103 start_prompt="<!--This tip is automatically generated by `make fix-copies`, do not fill manually!-->",104 end_prompt="<!--End of the generated tip-->",105 )106 107 new_list = get_model_list_for_task(task_guide)108 109 if current_list != new_list:110 if overwrite:111 with open(os.path.join(PATH_TO_TASK_GUIDES, task_guide), "w", encoding="utf-8", newline="\n") as f:112 f.writelines(lines[:start_index] + [new_list] + lines[end_index:])113 else:114 raise ValueError(115 f"The list of models that can be used in the {task_guide} guide needs an update. Run `make fix-copies`"116 " to fix this."117 )118 119 120if __name__ == "__main__":121 parser = argparse.ArgumentParser()122 parser.add_argument("--fix_and_overwrite", action="store_true", help="Whether to fix inconsistencies.")123 args = parser.parse_args()124 125 for task_guide in TASK_GUIDE_TO_MODELS.keys():126 check_model_list_for_task(task_guide, args.fix_and_overwrite)127 