pre-commit hook run only on staged/committed files on current branch?

[natarajanc-prodigygame]

vulture seems like a very useful tool, yet I can't configure it quite right. Perhaps it's a bug - or a configuration issue that could be explained slightly better.

i want to use the vulture pre-commit hook to only run on the staged/committed files in my branch individually and not run it based on directories specified in paths under the pyproject.toml

If i use the pyproject.toml config and specify the directories - the unused imported lib in 1 file can be used in another file in the same directory and hence the detection of unused imports does not happen

is it possible to only run on staged/committed files individually?

[RJ722]

If Vulture is run on staged/committed files only, it will report false "unused import" results.

I'm not sure how pre-commit works under the hood, but I'm wondering if it can be configured to do the following: Run Vulture on all your files, and then grep the results to show only the ones for staged/committed files.

[natarajanc-prodigygame]

@RJ722 taking a simple example - assuming 2 files in same directory

file a.py
import lib_a # unused in this file

file b.py
import lib_a # actually used in the file

if i run vulture on all files , the unused import for file a.py wont be surfaced (even with grep solution you indicated)

the unused import for file a.py only happens when i run vulture a.py which is what i want to happen - but unfortunately vulture is running on pre-commit hook on all files together

see below - running on single file gives a unused import , while running vulture on files in the parent folder and grepping for file gives nothing

$ vulture --min-confidence=90 dags/locations_nces_to_snowflake.py
dags/locations_nces_to_snowflake.py:10: unused import 'time' (90% confidence)

$ vulture --min-confidence=90 dags/ | grep locations_nces_to_snowflake.py
$  

[anita-clement]

Hi, is there an update on this issue?

I'm running into the same issue. vulture is running on all files when I'm running pre-commit while the expected behaviour from pre-commit is for it to run on staged files only. Thanks

[jendrikseipp]

I agree with @RJ722 : it's important to run Vulture on all files. Otherwise, if you edit a file containing name definitions but these names are not used in the file, you'll get many false positives.

As @RJ722 says: if it's possible to only report the unused code from changed files, that would be a useful change, though. PRs welcome :-)