Summary

• Certify Python 3.13 support for the nlp, nlp-advanced, and ocr optional profiles based on local install/profile validation.
• Extend CI to include Python 3.13 nlp and nlp-advanced test matrix rows.
• Add Python 3.13 profile-smoke coverage for nlp, nlp-advanced, and ocr.
• Install system Tesseract for OCR profile smoke jobs and assert the binary is available when requested.
• Document the 3.13 optional-profile support boundary: nlp, nlp-advanced, and ocr supported; Donut requires a local model; distributed and all are not newly certified in 4.5.

Local Python 3.13 validation

• Python: 3.13.13
• nlp: installed .[test,cli,nlp]; DATAFOG_INSTALL_PROFILE=nlp ... pytest tests/test_install_profiles.py -q passed; spaCy 3.8.14 imported.
• nlp-advanced: installed .[test,cli,nlp,nlp-advanced]; profile smoke passed; GLiNER 0.2.26, torch 2.6.0, transformers 5.1.0 imported.
• ocr: installed .[test,ocr]; profile smoke passed; Pillow 12.2.0, pytesseract 0.3.13, numpy 2.4.6 imported; system Tesseract 5.5.2 was available.
• nlp-advanced,ocr: Donut dependencies imported on Python 3.13; runtime correctly reported that the configured Donut model must be available locally.

Verification

• actionlint .github/workflows/ci.yml
• .venv312/bin/pre-commit run --files .github/workflows/ci.yml README.md docs/contributing.rst docs/optional-surfaces.rst tests/test_install_profiles.py --show-diff-on-failure
• DATAFOG_NO_TELEMETRY=1 DO_NOT_TRACK=1 .venv312/bin/python -m sphinx -b html docs docs/_build/html
• git diff --check
• DATAFOG_NO_TELEMETRY=1 DO_NOT_TRACK=1 DATAFOG_INSTALL_PROFILE=ocr DATAFOG_REQUIRE_TESSERACT=1 .venv313-ocr/bin/python -m pytest tests/test_install_profiles.py -q

Refs DFPY-68.