Commit 4f21ba72 authored by Jan Reimes's avatar Jan Reimes
Browse files

feat(02-checkout-graph-deprecation-config-01): enhance classification error messages 

- Include expected file types in warning
- Add re-download suggestion with CLI command
- Log folder contents for debugging
- Add test for improved error messages
parent e70e1df2

packages/3gpp-ai/threegpp_ai/operations/classify.py

+17 −1
Original line number Diff line number Diff line
@@ -229,7 +229,23 @@ def classify_document_files( 


    files = _scan_document_files(folder_path)

    if not files:

        logger.warning("No document files found in %s for document %s", folder_path, document_id)

        # Log folder contents for debugging

        try:

            folder_contents = list(folder_path.iterdir())

            contents_summary = [f.name for f in folder_contents]

        except (OSError, PermissionError):

            contents_summary = ["<cannot read folder>"]

        

        logger.warning(

            "No document files found in %s for document %s. "

            "Expected: .pdf/.docx/.xlsx/.pptx files. "

            "Folder contents: %s. "

            "Run 'tdoc-crawler checkout %s' to re-download if folder is empty.",

            folder_path,

            document_id,

            contents_summary,

            document_id,

        )

        return []



    if len(files) == 1:

tests/test_checkout.py

+28 −0
Original line number Diff line number Diff line
@@ -338,3 +338,31 @@ class TestCheckoutTdocIfNeeded: 
            # Should trigger checkout because folder doesn't exist

            mock_checkout.assert_called_once()

            assert result == new_path





class TestClassifyDocumentFiles:

    """Tests for classify_document_files error messages."""



    def test_empty_folder_warning_includes_redownload_suggestion(

        self,

        tmp_path: Path,

        caplog: pytest.LogCaptureFixture,

    ) -> None:

        """Test that empty folder warning includes re-download suggestion."""

        from threegpp_ai.operations.classify import classify_document_files

        

        folder_path = tmp_path / "empty_folder"

        folder_path.mkdir(parents=True)

        # Add only .ai subfolder

        ai_folder = folder_path / ".ai"

        ai_folder.mkdir(parents=True)

        (ai_folder / "processed.md").write_text("content")

        

        with caplog.at_level("WARNING"):

            result = classify_document_files("S4-251234", folder_path)

        

        assert result == []

        assert "No document files found" in caplog.text

        assert ".pdf/.docx/.xlsx/.pptx" in caplog.text

        assert "tdoc-crawler checkout S4-251234" in caplog.text

        assert "Folder contents:" in caplog.text