andreferraro
6a45368cb0
- Added scripts/extract_article_contents.py for batch scraping with stealth Foxcape and triple extraction (Trafilatura, Newspaper4k, Readability) - Created unit, integration, and E2E test suite in tests/test_extract_article_contents.py (90/90 passing) - Updated specs/003-article-content-extractor and README.md with usage documentation and CLI contracts - Passed ruff linting/formatting and mypy type checking cleanly
1 line
6.8 KiB
JSON
1 line
6.8 KiB
JSON
{"nodes": [{"id": "$graphify-root$_tests_test_language_py", "label": "test_language.py", "file_type": "code", "source_file": "tests/test_language.py", "source_location": "L1"}, {"id": "$graphify-root$_tests_test_language_test_normalize_text", "label": "test_normalize_text()", "file_type": "code", "source_file": "tests/test_language.py", "source_location": "L6", "_callable": true}, {"id": "$graphify-root$_tests_test_language_test_detect_portuguese", "label": "test_detect_portuguese()", "file_type": "code", "source_file": "tests/test_language.py", "source_location": "L14", "_callable": true}, {"id": "$graphify-root$_tests_test_language_test_detect_english", "label": "test_detect_english()", "file_type": "code", "source_file": "tests/test_language.py", "source_location": "L21", "_callable": true}, {"id": "$graphify-root$_tests_test_language_test_detect_spanish", "label": "test_detect_spanish()", "file_type": "code", "source_file": "tests/test_language.py", "source_location": "L28", "_callable": true}, {"id": "$graphify-root$_tests_test_language_test_detect_german", "label": "test_detect_german()", "file_type": "code", "source_file": "tests/test_language.py", "source_location": "L37", "_callable": true}, {"id": "$graphify-root$_tests_test_language_test_detect_italian", "label": "test_detect_italian()", "file_type": "code", "source_file": "tests/test_language.py", "source_location": "L44", "_callable": true}, {"id": "$graphify-root$_tests_test_language_test_detect_french", "label": "test_detect_french()", "file_type": "code", "source_file": "tests/test_language.py", "source_location": "L51", "_callable": true}, {"id": "$graphify-root$_tests_test_language_test_empty_language", "label": "test_empty_language()", "file_type": "code", "source_file": "tests/test_language.py", "source_location": "L58", "_callable": true}, {"id": "$graphify-root$_tests_test_language_rationale_1", "label": "Unit tests for language detection and text normalization.", "file_type": "rationale", "source_file": "tests/test_language.py", "source_location": "L1"}], "edges": [{"source": "$graphify-root$_tests_test_language_py", "target": "src_language", "relation": "imports_from", "context": "import", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L3", "weight": 1.0}, {"source": "$graphify-root$_tests_test_language_py", "target": "$graphify-root$_tests_test_language_test_normalize_text", "relation": "contains", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L6", "weight": 1.0}, {"source": "$graphify-root$_tests_test_language_py", "target": "$graphify-root$_tests_test_language_test_detect_portuguese", "relation": "contains", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L14", "weight": 1.0}, {"source": "$graphify-root$_tests_test_language_py", "target": "$graphify-root$_tests_test_language_test_detect_english", "relation": "contains", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L21", "weight": 1.0}, {"source": "$graphify-root$_tests_test_language_py", "target": "$graphify-root$_tests_test_language_test_detect_spanish", "relation": "contains", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L28", "weight": 1.0}, {"source": "$graphify-root$_tests_test_language_py", "target": "$graphify-root$_tests_test_language_test_detect_german", "relation": "contains", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L37", "weight": 1.0}, {"source": "$graphify-root$_tests_test_language_py", "target": "$graphify-root$_tests_test_language_test_detect_italian", "relation": "contains", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L44", "weight": 1.0}, {"source": "$graphify-root$_tests_test_language_py", "target": "$graphify-root$_tests_test_language_test_detect_french", "relation": "contains", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L51", "weight": 1.0}, {"source": "$graphify-root$_tests_test_language_py", "target": "$graphify-root$_tests_test_language_test_empty_language", "relation": "contains", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L58", "weight": 1.0}, {"source": "$graphify-root$_tests_test_language_rationale_1", "target": "$graphify-root$_tests_test_language_py", "relation": "rationale_for", "confidence": "EXTRACTED", "source_file": "tests/test_language.py", "source_location": "L1", "weight": 1.0}], "raw_calls": [{"caller_nid": "$graphify-root$_tests_test_language_test_normalize_text", "callee": "normalize_text", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L7", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_normalize_text", "callee": "normalize_text", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L8", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_normalize_text", "callee": "normalize_text", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L10", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_normalize_text", "callee": "normalize_text", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L11", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_detect_portuguese", "callee": "detect_language", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L16", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_detect_english", "callee": "detect_language", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L23", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_detect_spanish", "callee": "detect_language", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L32", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_detect_german", "callee": "detect_language", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L39", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_detect_italian", "callee": "detect_language", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L46", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_detect_french", "callee": "detect_language", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L53", "receiver": null}, {"caller_nid": "$graphify-root$_tests_test_language_test_empty_language", "callee": "detect_language", "is_member_call": false, "source_file": "tests/test_language.py", "source_location": "L59", "receiver": null}]} |