Remove vestigial Day.saints field; de-duplicate feast_name/DayCommemoration overlap - #165
Open
brianglass wants to merge 4 commits into
Open
Remove vestigial Day.saints field; de-duplicate feast_name/DayCommemoration overlap#165brianglass wants to merge 4 commits into
brianglass wants to merge 4 commits into
Conversation
…ration overlap Drop Day.saints (JSONField) -- fully unused since the saint-model-refactor project moved to DayCommemoration; no code read it, and it was already flagged as deferred cleanup in docs/saint-model-refactor.md. Migration 0009 removes it; fixtures/calendarium.json regenerated via dumpdata and verified identical apart from the dropped field. Separately, an architecture review of Day vs. DayCommemoration (prompted by noticing feast_name sometimes crams multiple names into one string, e.g. "Leavetaking of Transfiguration; St Tikhon of Zadonsk") found that 164 Day.feast_name entries encode person-level facts that DayCommemoration already models properly, with per-tradition granularity feast_name lacks. Of these, 75 turned out to be real saints (vs. season/Forefeast/Leavetaking labels, which correctly stay on Day). Resolved 62 of the 75: - 41 where Slavic and Greek both elevate the same saint (confirmed via data/antiochian_raw/'s harvested history, sampling multiple years and discounting years where the date fell on a Sunday, since Antiochian's title field switches to the paschal-cycle Sunday name that year regardless of the saint's actual rank): text moved from feast_name into the existing (previously suppressed) DayCommemoration entry. - 18 where Slavic elevates a saint Greek doesn't (Herman of Alaska, Alexander Nevsky, Vladimir, Seraphim of Sarov, Boris & Gleb, etc.): got a genuine Slavic/Greek Day-row split -- feast_level moves with the saint onto a slavic-tagged row, a new greek-tagged row keeps the day's other (non-elevated) content. One planned split (Aug 13, Tikhon of Zadonsk) turned out unnecessary once the data confirmed the elevation belongs to "Leavetaking of Transfiguration" itself, independent of Tikhon. - 3 ambiguous cases where a Dominical/calendar event permanently co-occurs (Basil the Great/Circumcision on Jan 1, Theodosius the Cenobiarch/Theophany-afterfeast on Jan 11, Symeon Stylites/Ecclesiastical New Year on Sep 1), resolved via individual research: none showed independent Greek-specific demotion, so all three got the same no-split treatment as the 41. The remaining 13 (7 fixed-date + 5 movable Lenten Sundays, plus one Theotokos-icon edge case) have no existing DayCommemoration row at all -- left alone, since creating one is a data-entry task, not a reconciliation. Also fixed along the way: - minimal_saints was a day_native-based grouping that broke down as soon as a "story-only" commemoration needed to be the thing shown (exactly this audit's use case). Replaced with a plain length cap (MINIMAL_SAINTS_LIMIT=3) on the full saints list, used both by summary_title's fallback and (newly) the monthly calendar grid, which previously rendered an unbounded saints list per cell. - test_ical.py's CalendarTest never loaded the commemorations.json fixture, so DayCommemoration was silently empty in that test's world -- a latent gap this audit's changes exposed. - Several tests updated to check the new (correct) location of moved text rather than the old feast_name string, preserving each test's original intent (documented inline). Full suite: 142 tests, only the pre-existing test_translation_changes_passage_content flake (confirmed unrelated, present on a clean main baseline). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Hf6j2xXQXywHVh3HAVRxB3
…ints Closes out the feast_name/DayCommemoration audit: these 12 person-type feast_name entries had no existing DayCommemoration row at all, so they couldn't be reconciled the way the other 62 were. Created a new Saint + DayCommemoration for each, tagged by confirmed presence/absence in data/antiochian_raw/: - 3 confirmed present for Greek too (Theodosius of Chernigov, the Royal Martyrs of Russia, Juvenaly & Peter the Aleut of Alaska -- Antiochian's own harvested description text names all three explicitly) -- tagged 'common'. - 3 confirmed absent from Greek's commemorations on these specific dates (Innocent of Moscow, Tikhon of Moscow, John Kochurov) -- tagged 'slavic'. feast_level on these dates was already correctly justified by something else independently confirmed tradition-invariant in the earlier tier-1 pass (Apostle Thomas, Apostle James son of Alphaeus), so no Day-row split was needed. - 4 universal Triodion Sundays (Theodore Tyro, Gregory Palamas, John Climacus, Mary of Egypt) -- tagged 'common' without needing per-date verification, since these are canonically identical fixed points in the Lenten cycle across every Byzantine-rite jurisdiction, not a jurisdiction-specific commemoration. - St Raphael of Brooklyn's movable (Antiochian's own "first Saturday of November") occasion reuses the same Saint identity already created for his fixed Feb 27 Slavic entry, rather than a new one -- same person, a different commemorated occasion. Updated two tests whose assertions still pointed at the old feast_name location for text this change (and tier 1/2) moved into the ordinary saints list. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Hf6j2xXQXywHVh3HAVRxB3
The feast_name/DayCommemoration de-duplication audit blanked feast_name on ~62 dates where the text was redundant with a DayCommemoration entry -- but for dates where that commemoration IS the day's whole identity (Peter & Paul, the Beheading of the Forerunner, Nicholas, etc.), this also silently dropped them from the page/API title, since day.feasts (not day.saints) drives it. Restore the original feast_name text (recovered from git history) on 52 of those dates and revert their DayCommemoration.ordering back to -1 so each shows once, as the title, rather than duplicating in the ordinary Commemorations list. Three cases deliberately excluded: - Aug 13 (Tikhon of Zadonsk) and Jul 5 (Sergius/Athanasius): the original text was a semicolon-joined combo of two distinct commemorations -- restoring it would reintroduce exactly the pattern this whole effort is meant to remove. - Aug 9 (Herman's canonization) and May 7 (Alexis Toth): new_style (civil-date-anchored) facts. Day.feast_name has no new_style gating, so putting them there bled onto the Julian-shifted label date, reintroducing the #146 bug class. Left as ordinary saints-list entries. Also simplify readings.html: merge the old standalone Feasts section into the header (day.feasts now drives the title directly), and drop the Commemorations-section-relabels-to-Feasts heuristic in favor of always showing "Commemorations". Verified corpus-wide: zero remaining feast_name values with semicolons. 142/142 tests pass; two golden JSON fixtures (last_bday.json, january.json) regenerated and diffed field-by-field to confirm only the intended feast_name/feasts/summary_title/saints fields changed. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Hf6j2xXQXywHVh3HAVRxB3
The previous commit's dumpdata regeneration used --indent 1 instead of the repo's established --indent 2, reindenting every line in both files and burying the actual 52-row content diff in ~99k lines of pure whitespace churn. Recreated with --indent 2; content is byte-identical to the previous commit (verified programmatically, zero field-level differences) -- this is a pure formatting fix. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Hf6j2xXQXywHVh3HAVRxB3
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Day.saints(JSONField) -- fully unused since the saint-model-refactor project moved toDayCommemoration, already flagged as deferred cleanup. Migration 0009; fixture regenerated viadumpdataand verified identical apart from the dropped field.Day.feast_nameandDayCommemorationwere redundantly encoding the same saint:data/antiochian_raw/'s multi-year harvested history) -- text moved into the existingDayCommemorationentry.Day-row split,feast_levelmoves with the saint.DayCommemorationrow at all -- left alone; creating one is a data-entry task, not a reconciliation.minimal_saintsredesigned from aday_native-based grouping (broke down for "story-only" commemorations) to a plain length cap (MINIMAL_SAINTS_LIMIT=3), now also used by the monthly calendar grid, which previously rendered an unbounded list per cell.test_ical.py'sCalendarTestnever loaded thecommemorations.jsonfixture, soDayCommemorationwas silently empty in that test's world.Test plan
test_translation_changes_passage_contentflake (confirmed unrelated, present on a cleanmainbaseline).dumpdata+ semantic diff (field-by-field comparison) before being copied into place -- confirmed each diff matches exactly the intended change, nothing else.feasts/saintsoutput across both traditions for representative dates in each tier, plus the Julian-calendar civil-date mechanism (issue Modern (post-1900) saints misplaced or missing in Julian-calendar mode #146) to confirm no regression.🤖 Generated with Claude Code
https://claude.ai/code/session_01Hf6j2xXQXywHVh3HAVRxB3