More utils doc (#25457)

* Document and clean more utils.

* More documentation and fixes

* Switch to Lysandre's token

* Address review comments

* Actually put else
This commit is contained in:
Sylvain Gugger
2023-08-17 07:58:35 +02:00
committed by GitHub
parent 36f183ebab
commit 2defb6b048
9 changed files with 411 additions and 84 deletions

View File

@@ -12,12 +12,30 @@
# 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.
"""
Utility that sorts the names in the auto mappings defines in the auto modules in alphabetical order.
Use from the root of the repo with:
```bash
python utils/sort_auto_mappings.py
```
to auto-fix all the auto mappings (used in `make style`).
To only check if the mappings are properly sorted (as used in `make quality`), do:
```bash
python utils/sort_auto_mappings.py --check_only
```
"""
import argparse
import os
import re
from typing import Optional
# Path are set with the intent you should run this script from the root of the repo.
PATH_TO_AUTO_MODULE = "src/transformers/models/auto"
@@ -28,7 +46,18 @@ _re_intro_mapping = re.compile(r"[A-Z_]+_MAPPING(\s+|_[A-Z_]+\s+)=\s+OrderedDict
_re_identifier = re.compile(r'\s*\(\s*"(\S[^"]+)"')
def sort_auto_mapping(fname, overwrite: bool = False):
def sort_auto_mapping(fname: str, overwrite: bool = False) -> Optional[bool]:
"""
Sort all auto mappings in a file.
Args:
fname (`str`): The name of the file where we want to sort auto-mappings.
overwrite (`bool`, *optional*, defaults to `False`): Whether or not to fix and overwrite the file.
Returns:
`Optional[bool]`: Returns `None` if `overwrite=True`. Otherwise returns `True` if the file has an auto-mapping
improperly sorted, `False` if the file is okay.
"""
with open(fname, "r", encoding="utf-8") as f:
content = f.read()
@@ -37,8 +66,8 @@ def sort_auto_mapping(fname, overwrite: bool = False):
line_idx = 0
while line_idx < len(lines):
if _re_intro_mapping.search(lines[line_idx]) is not None:
indent = len(re.search(r"^(\s*)\S", lines[line_idx]).groups()[0]) + 8
# Start of a new mapping!
indent = len(re.search(r"^(\s*)\S", lines[line_idx]).groups()[0]) + 8
while not lines[line_idx].startswith(" " * indent + "("):
new_lines.append(lines[line_idx])
line_idx += 1
@@ -65,11 +94,17 @@ def sort_auto_mapping(fname, overwrite: bool = False):
if overwrite:
with open(fname, "w", encoding="utf-8") as f:
f.write("\n".join(new_lines))
elif "\n".join(new_lines) != content:
return True
else:
return "\n".join(new_lines) != content
def sort_all_auto_mappings(overwrite: bool = False):
"""
Sort all auto mappings in the library.
Args:
overwrite (`bool`, *optional*, defaults to `False`): Whether or not to fix and overwrite the file.
"""
fnames = [os.path.join(PATH_TO_AUTO_MODULE, f) for f in os.listdir(PATH_TO_AUTO_MODULE) if f.endswith(".py")]
diffs = [sort_auto_mapping(fname, overwrite=overwrite) for fname in fnames]