aboutsummaryrefslogtreecommitdiff
path: root/tools/extract_missalemeum_oracle.py
diff options
context:
space:
mode:
Diffstat (limited to 'tools/extract_missalemeum_oracle.py')
-rw-r--r--tools/extract_missalemeum_oracle.py19
1 files changed, 14 insertions, 5 deletions
diff --git a/tools/extract_missalemeum_oracle.py b/tools/extract_missalemeum_oracle.py
index 4b010e2..1abc574 100644
--- a/tools/extract_missalemeum_oracle.py
+++ b/tools/extract_missalemeum_oracle.py
@@ -57,11 +57,20 @@
# under 40 characters that contains a chapter:verse-shaped digit pair
# (`\d+\s*[:,.]\s*\d+`, matching all three separators actually used in
# this source -- "Titus 2:11-15", "4 Kings, 5:1-15", "John 20. 19-31").
-# Verified exhaustively over the whole fixture (all 1 458 Lectio/
-# Evangelium sections across the 728 days that carry one): exactly one
-# candidate per section, zero ambiguous, zero false positives from a
-# longer rubric sentence. TWO days (Good Friday, both years -- the "Missa
-# Praesanctificatorum" liturgy) have no "Lectio"/"Evangelium" section at
+# Verified exhaustively over the whole fixture (all 1 456 Lectio/
+# Evangelium sections, 728 days x 2) -- CORRECTED, fix round 1
+# (coordinator review): this used to claim
+# "exactly one candidate per section, zero ambiguous". WRONG: 2 of the
+# 1 456 sections (Holy Saturday's own "Lectio", both years) yield TWO
+# candidates, ['Col 3:1-4', 'Ps 117, 1'] -- the second is the Tractus
+# verse that follows the real Epistle citation in the same paragraph,
+# also matching the verse-shaped regex. The fixture is still correct on
+# both rows (Col 3:1-4 IS the real citation) because [extract_citation]
+# below returns candidates[0] and the real citation happens to come
+# FIRST in the text -- correctness here rests on ORDERING, not
+# uniqueness, and this header previously overclaimed the latter. TWO
+# days (Good Friday, both years -- the "Missa Praesanctificatorum"
+# liturgy) have no "Lectio"/"Evangelium" section at
# all (a multi-lesson structure instead, "Lectiones"/"Passio", with no
# single reading occupying the Epistle/Gospel slot this schema assumes,
# the SAME shape colitur's own test_lectionary.ml records for its own