Document toc check and doctest check scripts (huggingface#25319)

* Clean doc toc check and make doctest list better

* Add to Makefile

Makefile

@@ -80,6 +80,7 @@ fix-copies:
 	python utils/check_copies.py --fix_and_overwrite
 	python utils/check_table.py --fix_and_overwrite
 	python utils/check_dummies.py --fix_and_overwrite
+	python utils/check_doctest_list.py --fix_and_overwrite
 	python utils/check_task_guides.py --fix_and_overwrite
 
 # Run tests for the library

utils/check_doc_toc.py

@@ -12,6 +12,25 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+"""
+This script is responsible for cleaning the model section of the table of content by removing duplicates and sorting
+the entries in alphabetical order.
+
+Usage (from the root of the repo):
+
+Check that the table of content is properly sorted (used in `make quality`):
+
+```bash
+python utils/check_doc_toc.py
+```
+
+Auto-sort the table of content if it is not properly sorted (used in `make style`):
+
+```bash
+python utils/check_doc_toc.py --fix_and_overwrite
+```
+"""
+
 
 import argparse
 from collections import defaultdict
@@ -24,7 +43,15 @@
 
 def clean_model_doc_toc(model_doc):
     """
-    Cleans the table of content of the model documentation by removing duplicates and sorting models alphabetically.
+    Cleans a section of the table of content of the model documentation (one specific modality) by removing duplicates
+    and sorting models alphabetically.
+
+    Args:
+        model_doc (`List[dict]`):
+            The list of dictionaries extracted from the `_toctree.yml` file for this specific modality.
+
+    Returns:
+        `List[dict]`: List of dictionaries like the input, but cleaned up and sorted.
     """
     counts = defaultdict(int)
     for doc in model_doc:
@@ -51,6 +78,14 @@ def clean_model_doc_toc(model_doc):
 
 
 def check_model_doc(overwrite=False):
+    """
+    Check that the content of the table of content in `_toctree.yml` is clean (no duplicates and sorted for the model
+    API doc) and potentially auto-cleans it.
+
+    Args:
+        overwrite (`bool`, *optional*, defaults to `False`):
+            Whether to just check if the TOC is clean or to auto-clean it (when `overwrite=True`).
+    """
     with open(PATH_TO_TOC, encoding="utf-8") as f:
         content = yaml.safe_load(f.read())
 
@@ -67,6 +102,7 @@ def check_model_doc(overwrite=False):
 
     model_doc = api_doc[model_idx]["sections"]
 
+    # Extract the modalities and clean them one by one.
     modalities_docs = [(idx, section) for idx, section in enumerate(model_doc) if "sections" in section]
     diff = False
     for idx, modality_doc in modalities_docs:

utils/check_doctest_list.py

@@ -12,28 +12,65 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+"""
+This script is responsible for cleaning the list of doctests by making sure the entries all exist and are in
+alphabetical order.
 
+Usage (from the root of the repo):
+
+Check that the doctest list is properly sorted and all files exist (used in `make repo-consistency`):
+
+```bash
+python utils/check_doctest_list.py
+```
+
+Auto-sort the doctest list if it is not properly sorted (used in `make fix-copies`):
+
+```bash
+python utils/check_doctest_list.py --fix_and_overwrite
+```
+"""
+import argparse
 import os
 
 
 # All paths are set with the intent you should run this script from the root of the repo with the command
 # python utils/check_doctest_list.py
 REPO_PATH = "."
+DOCTEST_FILE_PATHS = ["documentation_tests.txt", "slow_documentation_tests.txt"]
 
 
-if __name__ == "__main__":
-    doctest_file_path = os.path.join(REPO_PATH, "utils/documentation_tests.txt")
+def clean_doctest_list(doctest_file, overwrite=False):
     non_existent_paths = []
     all_paths = []
-    with open(doctest_file_path) as fp:
-        for line in fp:
+    with open(doctest_file, "r", encoding="utf-8") as f:
+        for line in f:
             line = line.strip()
             path = os.path.join(REPO_PATH, line)
             if not (os.path.isfile(path) or os.path.isdir(path)):
                 non_existent_paths.append(line)
-            all_paths.append(path)
+            all_paths.append(line)
+
     if len(non_existent_paths) > 0:
-        non_existent_paths = "\n".join(non_existent_paths)
+        non_existent_paths = "\n".join([f"- {f}" for f in non_existent_paths])
         raise ValueError(f"`utils/documentation_tests.txt` contains non-existent paths:\n{non_existent_paths}")
-    if all_paths != sorted(all_paths):
-        raise ValueError("Files in `utils/documentation_tests.txt` are not in alphabetical order.")
+
+    sorted_paths = sorted(all_paths)
+    if all_paths != sorted_paths:
+        if not overwrite:
+            raise ValueError(
+                f"Files in `{doctest_file}` are not in alphabetical order, run `make fix-copies` to fix "
+                "this automatically."
+            )
+        with open(doctest_file, "w", encoding="utf-8") as f:
+            f.write("\n".join(sorted_paths) + "\n")
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--fix_and_overwrite", action="store_true", help="Whether to fix inconsistencies.")
+    args = parser.parse_args()
+
+    for doctest_file in DOCTEST_FILE_PATHS:
+        doctest_file = os.path.join(REPO_PATH, "utils", doctest_file)
+        clean_doctest_list(doctest_file, args.fix_and_overwrite)