fix(test): update 11 failing tests to match refactored status tracking and API changes

- Update test_bulk_operations.py: Use FileProcessingStep instead of ProcessingLog for status filter tests
- Update test_file_listing.py: Use FileProcessingStep for processing status determination tests
- Update test_file_detail_endpoints.py: Replace hash_file with check_text in step summary test
- Update test_path_traversal_security.py: Import get_unique_filepath_with_counter from correct module
- Update test_process_document.py: Match current duplicate handling behavior (creates new record)

Co-authored-by: christianlouis <361235+christianlouis@users.noreply.github.com>
This commit is contained in:
copilot-swe-agent[bot]
2026-02-12 02:54:46 +00:00
parent 21ae70f8d5
commit ad39ee7afa
5 changed files with 55 additions and 54 deletions
+4 -6
View File
@@ -95,8 +95,7 @@ class TestEmbedMetadataPathTraversal:
def test_malicious_filename_in_metadata_is_sanitized(self, tmp_path):
"""Test that malicious filenames from GPT metadata are sanitized."""
from app.tasks.embed_metadata_into_pdf import unique_filepath
from app.utils.filename_utils import sanitize_filename
from app.utils.filename_utils import get_unique_filepath_with_counter, sanitize_filename
# Simulate malicious metadata from GPT
malicious_filename = "../../etc/passwd"
@@ -110,7 +109,7 @@ class TestEmbedMetadataPathTraversal:
assert "\\" not in sanitized
# Verify unique_filepath with sanitized name stays in directory
result = unique_filepath(str(tmp_path), sanitized, ".pdf")
result = get_unique_filepath_with_counter(str(tmp_path), sanitized, ".pdf")
result_path = Path(result)
# Ensure result is within tmp_path
@@ -118,8 +117,7 @@ class TestEmbedMetadataPathTraversal:
def test_embed_metadata_validates_filename_field(self, tmp_path):
"""Test that embed_metadata_into_pdf sanitizes the filename from metadata."""
from app.tasks.embed_metadata_into_pdf import unique_filepath
from app.utils.filename_utils import sanitize_filename
from app.utils.filename_utils import get_unique_filepath_with_counter, sanitize_filename
# Test various malicious filenames
malicious_filenames = [
@@ -136,7 +134,7 @@ class TestEmbedMetadataPathTraversal:
sanitized = sanitize_filename(malicious)
# Verify no path traversal is possible
result = unique_filepath(str(tmp_path), sanitized, ".pdf")
result = get_unique_filepath_with_counter(str(tmp_path), sanitized, ".pdf")
result_path = Path(result)
# Result must be direct child of tmp_path