linter: expand to 21 rules covering all 8 source types (v0.3.0)

Close the defense-in-depth gap flagged after iter 6: chapter, magazine, newspaper, social_media, podcast, and video exemplars had ZERO applicable linter rules, so their 100% linter pass rate was trivially true. Every exemplar now has 3-6 structural rules firing. Bump LINTER_VERSION to v0.3.0. Per harness discipline, this invalidates prior run logs' scores for cross-version comparison (they stay in the log as history). The re-baseline run still scores scalar 1.000, canary 1.000 on all 14 exemplars — formatter was already producing output that matches the new structural conventions, so tightening the linter didn't surface any regressions. New rules (12): - chapter_book_title_italicized - chapter_in_book_marker - magazine_name_italicized - newspaper_name_italicized - periodical_comma_before_date (magazine + newspaper; not journal) - social_media_post_quoted - social_media_platform_comma_date - podcast_series_italicized - podcast_episode_quoted - podcast_format_label - video_title_quoted (in quotes, NOT italicized) - video_format_label Extended: - article_title_quoted now covers journal, magazine, newspaper, and accepts ?/! as terminal punctuation Each rule has a positive + negative unit test with the CMOS/Purdue source cited in the docstring. Tests: 48 → 78 (30 new). Per-exemplar applicable-rule counts after this change: book 4 journal 5-6 web 3 chapter 4 magazine 5 social 4 podcast 5 video 4 newspaper 5
2026-04-10 22:01:56 -04:00
parent ced9c73474
commit fe5e071066
2 changed files with 646 additions and 10 deletions
@@ -15,12 +15,24 @@ from cmos.linter import (
    lint,
    rule_article_title_quoted,
    rule_book_title_italicized,
+    rule_chapter_book_title_italicized,
+    rule_chapter_in_book_marker,
    rule_doi_is_https_url,
    rule_ends_with_period,
    rule_journal_title_italicized,
+    rule_magazine_name_italicized,
+    rule_newspaper_name_italicized,
    rule_no_ibid,
    rule_no_place_of_publication_for_books,
    rule_page_range_uses_en_dash,
+    rule_periodical_comma_before_date,
+    rule_podcast_episode_quoted,
+    rule_podcast_format_label,
+    rule_podcast_series_italicized,
+    rule_social_media_platform_comma_date,
+    rule_social_media_post_quoted,
+    rule_video_format_label,
+    rule_video_title_quoted,
    rule_web_page_title_quoted,
 )
 from harness.score import Exemplar
@@ -81,6 +93,114 @@ WEB = Exemplar(
    expected_bibliography='Google. "Privacy Policy." Privacy & Terms. Effective November 15, 2023. https://policies.google.com/privacy.',
 )

+CHAPTER = Exemplar(
+    name="chapter",
+    source="test",
+    type="chapter",
+    tags=["chapter"],
+    canary=False,
+    messy_input="",
+    canonical={
+        "author": "Doyle, Kathleen",
+        "chapter_title": "The Queen Mary Psalter",
+        "book_title": "The Book by Design: The Remarkable Story of the World's Greatest Invention",
+        "editors": "edited by P. J. M. Marks and Stephen Parkin",
+        "publisher": "University of Chicago Press",
+        "year": 2023,
+    },
+    expected_bibliography='Doyle, Kathleen. "The Queen Mary Psalter." In *The Book by Design: The Remarkable Story of the World\'s Greatest Invention*, edited by P. J. M. Marks and Stephen Parkin. University of Chicago Press, 2023.',
+)
+
+MAGAZINE = Exemplar(
+    name="magazine",
+    source="test",
+    type="magazine",
+    tags=["magazine"],
+    canary=False,
+    messy_input="",
+    canonical={
+        "author": "Mead, Rebecca",
+        "article_title": "Terms of Aggrievement",
+        "magazine": "New Yorker",
+        "date": "December 18, 2023",
+    },
+    expected_bibliography='Mead, Rebecca. "Terms of Aggrievement." *New Yorker*, December 18, 2023.',
+)
+
+NEWSPAPER = Exemplar(
+    name="newspaper",
+    source="test",
+    type="newspaper",
+    tags=["newspaper"],
+    canary=False,
+    messy_input="",
+    canonical={
+        "author": "Blum, Dani",
+        "article_title": "Are Flax Seeds All That?",
+        "newspaper": "New York Times",
+        "date": "December 13, 2023",
+        "url": "https://www.nytimes.com/2023/12/13/well/eat/flax-seeds-benefits.html",
+    },
+    expected_bibliography='Blum, Dani. "Are Flax Seeds All That?" *New York Times*, December 13, 2023. https://www.nytimes.com/2023/12/13/well/eat/flax-seeds-benefits.html.',
+)
+
+SOCIAL = Exemplar(
+    name="social_media",
+    source="test",
+    type="social_media",
+    tags=["social_media"],
+    canary=False,
+    messy_input="",
+    canonical={
+        "author": "Chicago Manual of Style",
+        "post_text": "Is the world ready for singular they? We thought so in 1993.",
+        "platform": "Facebook",
+        "date": "April 17, 2015",
+        "url": "https://www.facebook.com/ChicagoManual/posts/10152906193679151",
+    },
+    expected_bibliography='Chicago Manual of Style. "Is the world ready for singular they? We thought so in 1993." Facebook, April 17, 2015. https://www.facebook.com/ChicagoManual/posts/10152906193679151.',
+)
+
+PODCAST = Exemplar(
+    name="podcast",
+    source="test",
+    type="podcast",
+    tags=["podcast"],
+    canary=False,
+    messy_input="",
+    canonical={
+        "host": "Ober, Lauren, host",
+        "podcast": "The Loudest Girl in the World",
+        "season_episode": "Season 1, episode 2",
+        "episode_title": "Goodbye, Routine; Hello, Meltdown!",
+        "publisher": "Pushkin Industries",
+        "date": "September 13, 2022",
+        "format": "Podcast",
+        "duration": "41 min., 37 sec.",
+        "url": "https://www.pushkin.fm/podcasts/loudest-girl-in-the-world",
+    },
+    expected_bibliography='Ober, Lauren, host. *The Loudest Girl in the World*. Season 1, episode 2, "Goodbye, Routine; Hello, Meltdown!" Pushkin Industries, September 13, 2022. Podcast, 41 min., 37 sec. https://www.pushkin.fm/podcasts/loudest-girl-in-the-world.',
+)
+
+VIDEO = Exemplar(
+    name="video",
+    source="test",
+    type="video",
+    tags=["video"],
+    canary=False,
+    messy_input="",
+    canonical={
+        "author": "Cowan, Vaitea",
+        "title": "How Green Hydrogen Could End the Fossil Fuel Era",
+        "venue": "TED Talk, Vancouver, BC",
+        "date": "April 2022",
+        "format": "Video",
+        "duration": "9 min., 15 sec.",
+        "url": "https://www.ted.com/talks/vaitea_cowan_how_green_hydrogen_could_end_the_fossil_fuel_era",
+    },
+    expected_bibliography='Cowan, Vaitea. "How Green Hydrogen Could End the Fossil Fuel Era." TED Talk, Vancouver, BC, April 2022. Video, 9 min., 15 sec. https://www.ted.com/talks/vaitea_cowan_how_green_hydrogen_could_end_the_fossil_fuel_era.',
+)
+

 def test_linter_version_is_set():
    assert LINTER_VERSION
@@ -254,13 +374,229 @@ def test_web_page_title_quoted_fails_when_unquoted():
    assert result.passed is False


+# --- Article title rule extended to magazines and newspapers -----------------
+
+
+def test_article_title_quoted_passes_on_magazine():
+    # Magazines follow the same article-title convention as journals.
+    result = rule_article_title_quoted(MAGAZINE.expected_bibliography, MAGAZINE)
+    assert result.applicable is True
+    assert result.passed is True
+
+
+def test_article_title_quoted_passes_on_newspaper():
+    result = rule_article_title_quoted(NEWSPAPER.expected_bibliography, NEWSPAPER)
+    assert result.applicable is True
+    assert result.passed is True
+
+
+def test_article_title_quoted_not_applicable_to_book():
+    result = rule_article_title_quoted(BOOK.expected_bibliography, BOOK)
+    assert result.applicable is False
+
+
+# --- Chapter rules ------------------------------------------------------------
+
+
+def test_chapter_book_title_italicized_passes():
+    # Chapters have their enclosing book's title italicized (CMOS 18,
+    # "Article, Chapter, Essay, etc., in a Book" model).
+    result = rule_chapter_book_title_italicized(CHAPTER.expected_bibliography, CHAPTER)
+    assert result.applicable is True
+    assert result.passed is True
+
+
+def test_chapter_book_title_italicized_fails_when_plain():
+    bad = CHAPTER.expected_bibliography.replace(
+        "*The Book by Design: The Remarkable Story of the World's Greatest Invention*",
+        "The Book by Design: The Remarkable Story of the World's Greatest Invention",
+    )
+    result = rule_chapter_book_title_italicized(bad, CHAPTER)
+    assert result.passed is False
+
+
+def test_chapter_book_title_italicized_not_applicable_to_book():
+    result = rule_chapter_book_title_italicized(BOOK.expected_bibliography, BOOK)
+    assert result.applicable is False
+
+
+def test_chapter_in_book_marker_passes():
+    # Chapter entries include the keyword "In" before the italicized book
+    # title to distinguish them from standalone books.
+    result = rule_chapter_in_book_marker(CHAPTER.expected_bibliography, CHAPTER)
+    assert result.passed is True
+
+
+def test_chapter_in_book_marker_fails_without_in():
+    bad = CHAPTER.expected_bibliography.replace(" In *", " *")
+    result = rule_chapter_in_book_marker(bad, CHAPTER)
+    assert result.passed is False
+
+
+# --- Magazine and newspaper name italicized ----------------------------------
+
+
+def test_magazine_name_italicized_passes():
+    # CMOS 18: magazine titles are italicized.
+    result = rule_magazine_name_italicized(MAGAZINE.expected_bibliography, MAGAZINE)
+    assert result.passed is True
+
+
+def test_magazine_name_italicized_fails_when_plain():
+    bad = MAGAZINE.expected_bibliography.replace("*New Yorker*", "New Yorker")
+    result = rule_magazine_name_italicized(bad, MAGAZINE)
+    assert result.passed is False
+
+
+def test_newspaper_name_italicized_passes():
+    result = rule_newspaper_name_italicized(NEWSPAPER.expected_bibliography, NEWSPAPER)
+    assert result.passed is True
+
+
+def test_newspaper_name_italicized_fails_when_plain():
+    bad = NEWSPAPER.expected_bibliography.replace("*New York Times*", "New York Times")
+    result = rule_newspaper_name_italicized(bad, NEWSPAPER)
+    assert result.passed is False
+
+
+# --- Periodical comma before date (magazine/newspaper only) -------------------
+
+
+def test_periodical_comma_before_date_passes_on_magazine():
+    # Magazines and newspapers use a COMMA between the italicized name and
+    # the date, not a period: *New Yorker*, December 18, 2023.
+    result = rule_periodical_comma_before_date(MAGAZINE.expected_bibliography, MAGAZINE)
+    assert result.passed is True
+
+
+def test_periodical_comma_before_date_fails_when_period_used():
+    bad = MAGAZINE.expected_bibliography.replace("*New Yorker*,", "*New Yorker*.")
+    result = rule_periodical_comma_before_date(bad, MAGAZINE)
+    assert result.passed is False
+
+
+def test_periodical_comma_before_date_passes_on_newspaper():
+    result = rule_periodical_comma_before_date(NEWSPAPER.expected_bibliography, NEWSPAPER)
+    assert result.passed is True
+
+
+def test_periodical_comma_before_date_not_applicable_to_journal():
+    # Journals use volume/issue/year in parens, no comma-after-italic pattern.
+    result = rule_periodical_comma_before_date(JOURNAL.expected_bibliography, JOURNAL)
+    assert result.applicable is False
+
+
+# --- Social media rules -------------------------------------------------------
+
+
+def test_social_media_post_quoted_passes():
+    # The post content is wrapped in straight double quotes.
+    result = rule_social_media_post_quoted(SOCIAL.expected_bibliography, SOCIAL)
+    assert result.passed is True
+
+
+def test_social_media_post_quoted_fails_when_title_cased():
+    # Title-casing is a common LLM error on social media content.
+    bad = SOCIAL.expected_bibliography.replace(
+        "Is the world ready for singular they? We thought so in 1993.",
+        "Is the World Ready for Singular They? We Thought So in 1993.",
+    )
+    result = rule_social_media_post_quoted(bad, SOCIAL)
+    assert result.passed is False
+
+
+def test_social_media_platform_comma_date_passes():
+    # The platform name is followed by a comma (not a period) before the date.
+    result = rule_social_media_platform_comma_date(SOCIAL.expected_bibliography, SOCIAL)
+    assert result.passed is True
+
+
+def test_social_media_platform_comma_date_fails_when_period_used():
+    bad = SOCIAL.expected_bibliography.replace("Facebook, April", "Facebook. April")
+    result = rule_social_media_platform_comma_date(bad, SOCIAL)
+    assert result.passed is False
+
+
+# --- Podcast rules ------------------------------------------------------------
+
+
+def test_podcast_series_italicized_passes():
+    # Podcast series titles are italicized like book titles.
+    result = rule_podcast_series_italicized(PODCAST.expected_bibliography, PODCAST)
+    assert result.passed is True
+
+
+def test_podcast_series_italicized_fails_when_plain():
+    bad = PODCAST.expected_bibliography.replace(
+        "*The Loudest Girl in the World*", "The Loudest Girl in the World"
+    )
+    result = rule_podcast_series_italicized(bad, PODCAST)
+    assert result.passed is False
+
+
+def test_podcast_episode_quoted_passes():
+    # Episode titles are in quotes (including terminal punctuation like !).
+    result = rule_podcast_episode_quoted(PODCAST.expected_bibliography, PODCAST)
+    assert result.passed is True
+
+
+def test_podcast_episode_quoted_fails_when_unquoted():
+    bad = PODCAST.expected_bibliography.replace(
+        '"Goodbye, Routine; Hello, Meltdown!"',
+        "Goodbye, Routine; Hello, Meltdown!",
+    )
+    result = rule_podcast_episode_quoted(bad, PODCAST)
+    assert result.passed is False
+
+
+def test_podcast_format_label_passes():
+    # CMOS includes a "Podcast," format label before duration.
+    result = rule_podcast_format_label(PODCAST.expected_bibliography, PODCAST)
+    assert result.passed is True
+
+
+def test_podcast_format_label_fails_when_missing():
+    bad = PODCAST.expected_bibliography.replace("Podcast, ", "")
+    result = rule_podcast_format_label(bad, PODCAST)
+    assert result.passed is False
+
+
+# --- Video rules --------------------------------------------------------------
+
+
+def test_video_title_quoted_passes():
+    # Video/TED Talk titles are in quotes (not italicized).
+    result = rule_video_title_quoted(VIDEO.expected_bibliography, VIDEO)
+    assert result.passed is True
+
+
+def test_video_title_quoted_fails_when_italicized():
+    bad = VIDEO.expected_bibliography.replace(
+        '"How Green Hydrogen Could End the Fossil Fuel Era."',
+        "*How Green Hydrogen Could End the Fossil Fuel Era.*",
+    )
+    result = rule_video_title_quoted(bad, VIDEO)
+    assert result.passed is False
+
+
+def test_video_format_label_passes():
+    # CMOS includes a "Video," format label before duration.
+    result = rule_video_format_label(VIDEO.expected_bibliography, VIDEO)
+    assert result.passed is True
+
+
+def test_video_format_label_fails_when_missing():
+    bad = VIDEO.expected_bibliography.replace("Video, ", "")
+    result = rule_video_format_label(bad, VIDEO)
+    assert result.passed is False
+
+
 # --- Aggregate check_passes ---------------------------------------------------


 def test_check_passes_on_canonical_outputs():
-    # All three seed exemplars' expected_bibliography strings must lint clean.
-    # If any of them fail, either the linter is too strict or the exemplar is
-    # wrong — either way, we want to know immediately.
-    assert check_passes(BOOK.expected_bibliography, BOOK)
-    assert check_passes(JOURNAL.expected_bibliography, JOURNAL)
-    assert check_passes(WEB.expected_bibliography, WEB)
+    # Every in-test exemplar's expected_bibliography must lint clean.
+    # If any fail, either the linter is too strict or the exemplar is wrong —
+    # either way, we want to know immediately.
+    for ex in (BOOK, JOURNAL, WEB, CHAPTER, MAGAZINE, NEWSPAPER, SOCIAL, PODCAST, VIDEO):
+        assert check_passes(ex.expected_bibliography, ex), f"linter rejected {ex.name}"