Christian Krakau-Louis
56 lines
2.5 KiB
Markdown
56 lines
2.5 KiB
Markdown
## Problem Description
|
|
|
|
7 tests in `tests/test_file_upload.py` are failing in the CI pipeline due to Celery task mocking issues. The tests expect the `delay` method to be called once, but it's being called 0 times.
|
|
|
|
### Failing Tests
|
|
- `test_upload_valid_text_file`
|
|
- `test_upload_valid_image_jpeg`
|
|
- `test_upload_valid_png_image`
|
|
- `test_upload_office_document_docx`
|
|
- `test_upload_csv_file`
|
|
- `test_upload_executable_file`
|
|
- `test_image_by_extension`
|
|
|
|
### Error Message
|
|
```
|
|
AssertionError: Expected 'delay' to have been called once. Called 0 times.
|
|
```
|
|
|
|
### Root Cause Analysis
|
|
|
|
The test fixture `mock_celery_tasks` is patching the Celery tasks at:
|
|
- `app.api.files.process_document.delay`
|
|
- `app.api.files.convert_to_pdf.delay`
|
|
|
|
However, looking at the actual implementation in `app/api/files.py` (lines 656, 662, 669, 674), the tasks are invoked using `.delay()` method correctly:
|
|
- Line 656: `task = process_document.delay(target_path, original_filename=safe_filename)`
|
|
- Line 662: `task = convert_to_pdf.delay(target_path, original_filename=safe_filename)`
|
|
- Line 669: `task = convert_to_pdf.delay(target_path, original_filename=safe_filename)`
|
|
- Line 674: `task = convert_to_pdf.delay(target_path, original_filename=safe_filename)`
|
|
|
|
The issue is that the patches are applied correctly but the mocks are not being invoked. This suggests:
|
|
1. The patch path might need adjustment
|
|
2. The tasks might be imported/accessed differently at runtime
|
|
3. There might be an issue with how the tasks are being called in the test environment
|
|
|
|
### Files to Fix
|
|
- `tests/test_file_upload.py` - Fix the mock patching in the `mock_celery_tasks` fixture (lines 20-36)
|
|
|
|
### Expected Behavior
|
|
All tests should pass with 100% success rate. The mocked Celery tasks should be called appropriately when files are uploaded through the `/api/ui-upload` endpoint.
|
|
|
|
### Solution Approach
|
|
1. Verify the correct patch path for the Celery tasks
|
|
2. Ensure the patch is applied before the import of the tasks
|
|
3. Consider patching at the module level where tasks are defined rather than where they're used
|
|
4. Alternative: Patch the task objects directly: `app.tasks.process_document.process_document` and `app.tasks.convert_to_pdf.convert_to_pdf`
|
|
|
|
### Reference
|
|
- Failing job: https://github.com/christianlouis/DocuElevate/actions/runs/21801196852/job/62896908441
|
|
- Commit: 17795620c62c0bcf744eb7e940a3262c364f9bed
|
|
|
|
### Acceptance Criteria
|
|
- All 7 failing tests pass
|
|
- Total test suite passes with 180/180 tests passing (currently 173/180)
|
|
- CI pipeline succeeds
|
|
- No regression in other tests |