Coverage for src/local_deep_research/research_library/extraction_catalog.py: 100%

1"""Catalog identifiers for arXiv HTML-extracted text provenance. 

2 

3``LibraryDocument.extraction_method`` and ``.extraction_source`` 

4(``database/models/library.py``) are plain unconstrained strings, not 

5SQLAlchemy ``Enum`` columns, so adding a new catalog value never changes 

6the schema Alembic ships -- see the docstring on ``ExtractionMethod`` / 

7``ExtractionSource`` in that file. 

8 

9The ``"html_extraction"`` / ``"arxiv_html"`` values ``DownloadService`` 

10persists for an HTML-extracted arXiv download (see 

11``_arxiv_text_provenance`` in 

12``research_library/services/download_service.py``) are defined here, 

13outside ``database/models/``, rather than as new members on the 

14pre-existing enums in ``library.py``. That keeps every file under 

15``database/models/`` byte-identical to the merge base, which two tests 

16enforce: 

17 

18* ``test_migration_chain_integrity.py::test_every_file_behind_the_alembic_metadata_is_unchanged_since_the_fork`` 

19* ``test_upgrade_from_pre_migration_install.py::test_pr_changes_no_orm_model`` 

20 

21Both treat any edit to a file backing ``Base.metadata`` (or anything 

22under ``database/models/``) as presumed schema drift, regardless of 

23whether the specific edit actually changes a column -- so a value that 

24does not need to live there should not. 

25""" 

26 

27import enum 

28 

29 

30class ExtractionMethodExtra(str, enum.Enum): 

31 """Extraction-method catalog values added after the merge base.""" 

32 

33 HTML_EXTRACTION = "html_extraction" 

34 

35 

36class ExtractionSourceExtra(str, enum.Enum): 

37 """Extraction-source catalog values added after the merge base.""" 

38 

39 ARXIV_HTML = "arxiv_html"