Files
catena/parity/golden/bibclean/records.ndjson
T

34 lines
9.2 KiB
Plaintext
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
{"clean_underscores": "The Human Condition", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "arendt hannah", "reasons": [], "severity": 0, "title_case": "The Human Condition", "title_in_author": false, "valid_isbn": true}
{"clean_underscores": "THE HUMAN CONDITION", "is_shouting": true, "isbn_from_filename": null, "looks_like_person": true, "normalise": "arendt hannah", "reasons": ["allcaps_title", "missing_isbn"], "severity": 2, "title_case": "The Human Condition", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "Untitled", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": false, "normalise": "arguing about war", "reasons": ["placeholder_title", "title_in_author", "missing_isbn"], "severity": 7, "title_case": "Untitled", "title_in_author": true, "valid_isbn": false}
{"clean_underscores": "Untitled", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "default", "reasons": ["placeholder_title", "placeholder_author", "missing_isbn"], "severity": 8, "title_case": "Untitled", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "chapter1.", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "unknown", "reasons": ["unknown_author", "filename_title", "lowercase_title", "missing_isbn"], "severity": 8, "title_case": "Chapter1.pdf", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "A:\\pwecent2005.wpd", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": false, "normalise": "", "reasons": ["unknown_author", "path_title", "filename_title", "missing_isbn"], "severity": 10, "title_case": "A:\\pwecent2005.wpd", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "/home/matt/book", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "1993", "reasons": ["path_title", "numeric_author", "lowercase_title", "missing_isbn"], "severity": 10, "title_case": "/Home/matt/book", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "Microsoft Word - ch3.doc", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "oxford university press", "reasons": ["filename_title", "junk_title", "publisher_as_author", "missing_isbn"], "severity": 9, "title_case": "Microsoft Word - Ch3.doc", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "Board Advisors etc", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "c s lewis madeleine l engle", "reasons": ["underscore_title", "missing_isbn"], "severity": 1, "title_case": "Board_advisors_etc", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "The Fall: And the Rise", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "plato", "reasons": ["underscore_title", "missing_isbn"], "severity": 1, "title_case": "The Fall_ and the Rise_txt", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "heidegger areopagite", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "smith jr", "reasons": ["lowercase_title", "missing_isbn"], "severity": 2, "title_case": "Heidegger Areopagite", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "Lossky s Argument", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": false, "normalise": "kierkegaard s conception of god", "reasons": ["missing_apostrophe", "missing_isbn"], "severity": 2, "title_case": "Lossky S Argument", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "Rama", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "arthur c clarke gentry lee", "reasons": ["missing_isbn"], "severity": 0, "title_case": "Rama", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "MIT", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": false, "normalise": "an ethnohistorical dictionary china", "reasons": ["missing_isbn"], "severity": 0, "title_case": "Mit", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "ÊTRE ET TEMPS", "is_shouting": true, "isbn_from_filename": null, "looks_like_person": true, "normalise": "heidegger martin", "reasons": ["allcaps_title", "missing_isbn"], "severity": 2, "title_case": "Être Et Temps", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "ÉTUDES SUR LE MOYEN ÂGE", "is_shouting": true, "isbn_from_filename": null, "looks_like_person": true, "normalise": "gilson tienne", "reasons": ["allcaps_title", "missing_isbn"], "severity": 2, "title_case": "Études Sur Le Moyen Âge", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "DON'T LOOK NOW", "is_shouting": true, "isbn_from_filename": null, "looks_like_person": true, "normalise": "du maurier daphne", "reasons": ["allcaps_title", "missing_isbn"], "severity": 2, "title_case": "Don't Look Now", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "THE WELL-BUILT WHEEL", "is_shouting": true, "isbn_from_filename": null, "looks_like_person": false, "normalise": "", "reasons": ["unknown_author", "allcaps_title", "missing_isbn"], "severity": 5, "title_case": "The Well-Built Wheel", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "The (Second) Estate", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "bloch marc", "reasons": ["missing_isbn"], "severity": 0, "title_case": "The (Second) Estate", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": null, "is_shouting": false, "isbn_from_filename": null, "looks_like_person": false, "normalise": "", "reasons": ["unknown_author", "missing_isbn"], "severity": 3, "title_case": null, "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": false, "normalise": "", "reasons": ["unknown_author", "missing_isbn"], "severity": 3, "title_case": "", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "spaced", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "arendt hannah", "reasons": [], "severity": 0, "title_case": "Spaced", "title_in_author": false, "valid_isbn": true}
{"clean_underscores": "A Book", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "anonymous", "reasons": ["unknown_author"], "severity": 3, "title_case": "A Book", "title_in_author": false, "valid_isbn": true}
{"clean_underscores": "Scan 001", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "administrator", "reasons": ["placeholder_author"], "severity": 4, "title_case": "Scan 001", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "xxx", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "xxxx", "reasons": ["placeholder_title", "placeholder_author", "missing_isbn"], "severity": 8, "title_case": "Xxx", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "The Document", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "oxford michael", "reasons": ["missing_isbn"], "severity": 0, "title_case": "The Document", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "Full Page Fax Print", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "user", "reasons": ["placeholder_author", "junk_title", "missing_isbn"], "severity": 7, "title_case": "Full Page Fax Print", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "Tam Sayfa Faks Yazdırma", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "routledge", "reasons": ["junk_title", "publisher_as_author", "missing_isbn"], "severity": 6, "title_case": "Tam Sayfa Faks Yazdırma", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "A Book", "is_shouting": false, "isbn_from_filename": "9780198237907", "looks_like_person": true, "normalise": "arendt", "reasons": ["missing_isbn"], "severity": 0, "title_case": "A Book", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "A Book", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "arendt", "reasons": ["missing_isbn"], "severity": 0, "title_case": "A Book", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "A Book", "is_shouting": false, "isbn_from_filename": "080213825X", "looks_like_person": true, "normalise": "arendt", "reasons": ["missing_isbn"], "severity": 0, "title_case": "A Book", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "Ἀθῆναι καὶ Μόσχα", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "", "reasons": ["missing_isbn"], "severity": 0, "title_case": "Ἀθῆναι Καὶ Μόσχα", "title_in_author": false, "valid_isbn": false}
{"clean_underscores": "東京の手紙", "is_shouting": false, "isbn_from_filename": null, "looks_like_person": true, "normalise": "", "reasons": ["missing_isbn"], "severity": 0, "title_case": "東京の手紙", "title_in_author": false, "valid_isbn": false}