Add LXX2012+WEB Bible translation with UI/API selection

.github/workflows/tests.yml

@@ -2,10 +2,14 @@ name: Tests
 
 on:
   push:
-    branches-ignore: [main]
+    branches: [main]
   pull_request:
     branches: [main]
 
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref || github.ref_name }}
+  cancel-in-progress: true
+
 jobs:
   test:
     runs-on: ubuntu-latest

bible/migrations/0003_add_translation_field.py

@@ -0,0 +1,40 @@
+# Generated by Django 6.0.7 on 2026-07-30 20:26
+
+from django.db import migrations, models
+
+
+def backfill_translations(apps, schema_editor):
+    # AddField's default only gives every existing row 'kjv' -- correct for
+    # the English rows, wrong for Romanian/Serbian, which were never KJV.
+    Verse = apps.get_model('bible', 'Verse')
+    Verse.objects.filter(language='ro').update(translation='rccv')
+    Verse.objects.filter(language='sr').update(translation='srp1865')
+
+
+def unbackfill_translations(apps, schema_editor):
+    Verse = apps.get_model('bible', 'Verse')
+    Verse.objects.filter(language__in=('ro', 'sr')).update(translation='kjv')
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('bible', '0002_load_scriptures'),
+    ]
+
+    operations = [
+        migrations.AlterUniqueTogether(
+            name='verse',
+            unique_together=set(),
+        ),
+        migrations.AddField(
+            model_name='verse',
+            name='translation',
+            field=models.CharField(default='kjv', max_length=20),
+        ),
+        migrations.RunPython(backfill_translations, unbackfill_translations),
+        migrations.AlterUniqueTogether(
+            name='verse',
+            unique_together={('book', 'chapter', 'verse', 'language', 'translation')},
+        ),
+    ]

bible/migrations/0004_load_lxx2012_web.py

@@ -0,0 +1,46 @@
+from django.db import migrations
+
+from bible.parse import parse_usfx
+
+TRANSLATION = 'lxx2012-web'
+
+# Only these WEB books are used -- the rest of that file is WEB's own OT and
+# Deuterocanon, which we don't want here since LXX2012 supplies the OT half
+# of this pairing instead.
+NT_BOOKS = {
+    'MAT', 'MRK', 'LUK', 'JHN', 'ACT', 'ROM', '1CO', '2CO', 'GAL', 'EPH',
+    'PHP', 'COL', '1TH', '2TH', '1TI', '2TI', 'TIT', 'PHM', 'HEB', 'JAS',
+    '1PE', '2PE', '1JN', '2JN', '3JN', 'JUD', 'REV',
+}
+
+
+def load_lxx2012_web(apps, schema_editor):
+    Verse = apps.get_model('bible', 'Verse')
+
+    for verse in parse_usfx('data/eng-lxx2012_usfx.xml'):
+        if verse['chapter'] is None or verse['verse'] is None:
+            continue
+        Verse.objects.create(language='en', translation=TRANSLATION, **verse)
+
+    for verse in parse_usfx('data/eng-web_usfx.xml'):
+        if verse['book'] not in NT_BOOKS:
+            continue
+        if verse['chapter'] is None or verse['verse'] is None:
+            continue
+        Verse.objects.create(language='en', translation=TRANSLATION, **verse)
+
+
+def unload_lxx2012_web(apps, schema_editor):
+    Verse = apps.get_model('bible', 'Verse')
+    Verse.objects.filter(translation=TRANSLATION).delete()
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('bible', '0003_add_translation_field'),
+    ]
+
+    operations = [
+        migrations.RunPython(load_lxx2012_web, unload_lxx2012_web),
+    ]

bible/models.py

@@ -18,8 +18,21 @@ class ReferenceParseError(Exception):
     pass
 
 
+# The translation used when a caller doesn't specify one -- keeps every
+# existing call site (which predates the translation field) resolving to
+# exactly the same rows as before, per language.
+DEFAULT_TRANSLATIONS = {
+    'en': 'kjv',
+    'ro': 'rccv',
+    'sr': 'srp1865',
+}
+
+
 class VerseManager(models.Manager):
-    def lookup_reference(self, reference, language='en'):
+    def lookup_reference(self, reference, language='en', translation=None):
+        if translation is None:
+            translation = DEFAULT_TRANSLATIONS[language]
+
         conditionals = []
         book = ''
 
@@ -77,7 +90,7 @@ def lookup_reference(self, reference, language='en'):
 
         # Run the query
         expression = functools.reduce(operator.or_, conditionals)
-        return self.filter(language=language).filter(expression)
+        return self.filter(language=language, translation=translation).filter(expression)
 
 
 class Verse(models.Model):
@@ -87,11 +100,12 @@ class Verse(models.Model):
     content = models.TextField()
     paragraph_start = models.BooleanField(default=False)
     language = models.CharField(max_length=10)
+    translation = models.CharField(max_length=20, default='kjv')
 
     objects = VerseManager()
 
     class Meta:
-        unique_together = 'book', 'chapter', 'verse', 'language'
+        unique_together = 'book', 'chapter', 'verse', 'language', 'translation'
 
     def __str__(self):
         blurb = textwrap.shorten(self.content, width=20, placeholder='...')

bible/parse.py

@@ -8,6 +8,7 @@ def parse_usfx(filename):
     book, chapter, verse = None, None, None
     paragraph_start = False
     is_valid_content = False
+    was_valid_content = False
     strings = []
 
     def make_verse():
@@ -59,6 +60,12 @@ def make_verse():
                     yield make_verse()
 
                 verse = node.getAttribute('id')
+                if verse and '-' in verse:
+                    # A verse bridge (e.g. "1-2"), where the source merges
+                    # two verses into one printed unit -- store it under the
+                    # first number rather than failing to parse an int; the
+                    # combined text is already all in this one entry.
+                    verse = verse.split('-')[0]
                 is_valid_content = True
             case [pulldom.START_ELEMENT, 've']:
                 yield make_verse()
@@ -67,11 +74,20 @@ def make_verse():
             case [pulldom.START_ELEMENT, 'p']:
                 paragraph_start = True
 
-            # Footnote Element
-            case [pulldom.START_ELEMENT, 'f']:
+            # Footnote and cross-reference elements -- <x> (cross-reference,
+            # e.g. WEB's "11:33 Daniel 6:22-23" pointing back to an OT
+            # parallel) is structurally the same kind of aside as <f>
+            # (footnote), so it needs the same is_valid_content suppression;
+            # without it, the reference text gets appended straight into the
+            # verse content.
+            case [pulldom.START_ELEMENT, 'f' | 'x']:
+                was_valid_content = is_valid_content
                 is_valid_content = False
-            case [pulldom.END_ELEMENT, 'f']:
-                is_valid_content = True
+            case [pulldom.END_ELEMENT, 'f' | 'x']:
+                # Restore whatever was in effect before the aside rather than
+                # assuming True -- these can appear in content (like Psalm
+                # title <d> blocks) that isn't part of a verse.
+                is_valid_content = was_valid_content
 
             # Character content
             case [pulldom.CHARACTERS, _]:

bible/tests/test_parse.py

@@ -6,5 +6,15 @@
 class ParseTest(TestCase):
     def test_gen_9_23(self):
         expected = 'And Shem and Japheth took a garment, and laid it upon both their shoulders, and went backward, and covered the nakedness of their father; and their faces were backward, and they saw not their father’s nakedness.'
-        verse = Verse.objects.get(book='GEN', chapter=9, verse=23, language='en')
+        verse = Verse.objects.get(book='GEN', chapter=9, verse=23, language='en', translation='kjv')
         self.assertEqual(expected, verse.content)
+
+    def test_web_cross_references_stripped(self):
+        """WEB annotates verses with <x> cross-reference elements (e.g.
+        pointing Heb 11:33 back to Daniel 6) that aren't part of the verse
+        text itself. These must not leak into the stored content the way
+        <f> footnotes were already excluded."""
+        expected = 'who through faith subdued kingdoms, worked out righteousness, obtained promises, stopped the mouths of lions,'
+        verse = Verse.objects.get(book='HEB', chapter=11, verse=33, language='en', translation='lxx2012-web')
+        self.assertEqual(expected, verse.content)
+        self.assertNotIn('Daniel', verse.content)

calendarium/api.py

@@ -18,7 +18,7 @@
 from pydantic import AnyUrl, AnyHttpUrl, conint, constr, validator
 
 from . import datetools, liturgics, views
-from .datetools import Calendar, Tradition
+from .datetools import Calendar, Tradition, Translation
 from orthocal.decorators import etag, etag_date, instrument_endpoint
 
 logger = logging.getLogger(__name__)
@@ -162,9 +162,9 @@ class OembedSchema(Schema):
 def not_implemented_handler(request, exc):
     return api.create_response(request, {'message': 'Not Implemented'}, status=501)
 
-async def _get_calendar_day(request, cal, tradition, year, month, day):
+async def _get_calendar_day(request, cal, tradition, year, month, day, translation=None):
     try:
-        day = liturgics.Day(year, month, day, calendar=cal, tradition=tradition, language=request.LANGUAGE_CODE)
+        day = liturgics.Day(year, month, day, calendar=cal, tradition=tradition, language=request.LANGUAGE_CODE, translation=translation)
     except ValueError:
         # The date is out of range or invalid
         raise Http404
@@ -178,26 +178,32 @@ async def _get_calendar_day(request, cal, tradition, year, month, day):
 @api.get('{cal:cal}/{year:year}/{month:month}/{day:day}/', response=DaySchema)
 @instrument_endpoint
 @decorate_view(etag)
-async def get_calendar_day(request, cal: Calendar, year: year, month: month, day: day):
+async def get_calendar_day(request, cal: Calendar, year: year, month: month, day: day, translation: Translation = None):
     """Get information about the liturgical day for the given calendar and date.
     The *cal* path parameter should be `gregorian` or `julian`. The legacy `oca` or `rocor`
     will still work, but should be avoided for new code. This serves the Slavic/OCA
     tradition; see the `{tradition}/{cal}/...` routes below for the Greek tradition.
+    The optional *translation* query parameter selects the Bible translation
+    (`kjv` or `lxx2012-web`); it only affects English content and defaults to
+    `kjv` when omitted.
     """
 
-    return await _get_calendar_day(request, cal, Tradition.Slavic, year, month, day)
+    return await _get_calendar_day(request, cal, Tradition.Slavic, year, month, day, translation)
 
 @api.get('{tradition:tradition}/{cal:cal}/{year:year}/{month:month}/{day:day}/', response=DaySchema)
 @instrument_endpoint
 @decorate_view(etag)
-async def get_calendar_day_tradition(request, tradition: Tradition, cal: Calendar, year: year, month: month, day: day):
+async def get_calendar_day_tradition(request, tradition: Tradition, cal: Calendar, year: year, month: month, day: day, translation: Translation = None):
     """Get information about the liturgical day for the given tradition, calendar, and date.
     The *tradition* path parameter should be `slavic` or `greek`. The legacy `oca`,
     `antiochian`, and `goa` will still work, but should be avoided for new code.
     The *cal* path parameter should be `gregorian` or `julian`.
+    The optional *translation* query parameter selects the Bible translation
+    (`kjv` or `lxx2012-web`); it only affects English content and defaults to
+    `kjv` when omitted.
     """
 
-    return await _get_calendar_day(request, cal, tradition, year, month, day)
+    return await _get_calendar_day(request, cal, tradition, year, month, day, translation)
 
 async def _get_calendar_month(request, cal, tradition, year, month) -> List[DaySchemaLite]:
     days = [d async for d in liturgics.amonth_of_days(year, month, calendar=cal, tradition=tradition)]
@@ -239,26 +245,32 @@ async def get_calendar_month_tradition(request, tradition: Tradition, cal: Calen
 @api.get('{cal:cal}/', response=DaySchema, summary='Get Today')
 @instrument_endpoint
 @decorate_view(etag_date)
-async def get_calendar_default(request, cal: Calendar):
+async def get_calendar_default(request, cal: Calendar, translation: Translation = None):
     """Get information about the current liturgical day for the given calendar.
     The timezone is Pacific Time. The *cal* path parameter should be
     `gregorian` or `julian`. The legacy `oca` or `rocor` will still work, but
     should be avoided for new code. This serves the Slavic/OCA tradition; see
     the `{tradition}/{cal}/` route below for the Greek tradition.
+    The optional *translation* query parameter selects the Bible translation
+    (`kjv` or `lxx2012-web`); it only affects English content and defaults to
+    `kjv` when omitted.
     """
     dt = timezone.localtime()
-    return await _get_calendar_day(request, cal, Tradition.Slavic, dt.year, dt.month, dt.day)
+    return await _get_calendar_day(request, cal, Tradition.Slavic, dt.year, dt.month, dt.day, translation)
 
 @api.get('{tradition:tradition}/{cal:cal}/', response=DaySchema, summary='Get Today (by tradition)')
 @instrument_endpoint
 @decorate_view(etag_date)
-async def get_calendar_default_tradition(request, tradition: Tradition, cal: Calendar):
+async def get_calendar_default_tradition(request, tradition: Tradition, cal: Calendar, translation: Translation = None):
     """Get information about the current liturgical day for the given tradition and calendar.
     The timezone is Pacific Time. The *tradition* path parameter should be
     `slavic` or `greek`. The *cal* path parameter should be `gregorian` or `julian`.
+    The optional *translation* query parameter selects the Bible translation
+    (`kjv` or `lxx2012-web`); it only affects English content and defaults to
+    `kjv` when omitted.
     """
     dt = timezone.localtime()
-    return await _get_calendar_day(request, cal, tradition, dt.year, dt.month, dt.day)
+    return await _get_calendar_day(request, cal, tradition, dt.year, dt.month, dt.day, translation)
 
 @api.get('oembed/calendar/', response=OembedSchema, exclude_none=True)
 @instrument_endpoint

calendarium/datetools.py

@@ -30,6 +30,26 @@ def cal_session_key(tradition):
     always Gregorian and shouldn't clobber a Slavic Julian/Gregorian choice."""
     return f'cal_{tradition}'
 
+class Translation(StrEnum):
+    KJV = 'kjv'
+    LXX2012WEB = 'lxx2012-web'
+
+def translation_session_key(language):
+    """Mirrors cal_session_key(tradition): translation choice is remembered
+    per-language, since it's only meaningful for English today."""
+    return f'translation_{language}'
+
+# Display labels for every translation code that can appear in Verse rows,
+# not just the ones selectable via the dropdown -- rccv/srp1865 are included
+# so the "Scripture Readings (...)" heading is accurate for Romanian/Serbian
+# too, instead of always showing "(KJV)" regardless of language.
+TRANSLATION_LABELS = {
+    'kjv': 'King James Version',
+    'lxx2012-web': 'LXX2012 & WEB',
+    'rccv': 'Romanian Corrected Cornilescu Version',
+    'srp1865': 'Serbian (Karadžić/Daničić, 1865)',
+}
+
 class FastLevels(IntEnum):
     NoFast         = 0
     Fast           = 1

calendarium/liturgics/day.py

@@ -96,12 +96,12 @@ class Day:
     shared and lives here in the base class.
     """
 
-    def __new__(cls, year, month, day, calendar=Calendar.Gregorian, tradition=Tradition.Slavic, language='en'):
+    def __new__(cls, year, month, day, calendar=Calendar.Gregorian, tradition=Tradition.Slavic, language='en', translation=None):
         if cls is Day:
             cls = _DAY_CLASSES[tradition]
         return super().__new__(cls)
 
-    def __init__(self, year, month, day, calendar=Calendar.Gregorian, tradition=Tradition.Slavic, language='en'):
+    def __init__(self, year, month, day, calendar=Calendar.Gregorian, tradition=Tradition.Slavic, language='en', translation=None):
         self.gregorian_date = date(year, month, day)
 
         if calendar == Calendar.Gregorian:
@@ -123,6 +123,7 @@ def __init__(self, year, month, day, calendar=Calendar.Gregorian, tradition=Trad
         self.calendar = calendar
         self.pyear = _YEAR_CLASSES[tradition](pyear, calendar)
         self.language = language
+        self.translation = translation
 
     async def ainitialize(self):
         """Do the expensive stuff here to keep it out of the constructor."""
@@ -440,7 +441,7 @@ async def aget_readings(self, fetch_content=False):
         if hasattr(self, 'readings'):
             if fetch_content:
                 for reading in self.readings:
-                    await reading.pericope.aget_passage(language=self.language)
+                    await reading.pericope.aget_passage(language=self.language, translation=self.translation)
 
             return self.readings
 
@@ -505,7 +506,7 @@ async def aget_readings(self, fetch_content=False):
         self.readings = []
         for reading in rows:
             if fetch_content:
-                await reading.pericope.aget_passage(language=self.language)
+                await reading.pericope.aget_passage(language=self.language, translation=self.translation)
 
             if -42 < self.pdist < -7 and self.feast_level < 7 and reading.source == 'Matins Gospel':
                 # Place Lenten Matins Gospel at the top
@@ -524,7 +525,7 @@ async def aget_abbreviated_readings(self, fetch_content=False):
         if hasattr(self, 'abbreviated_readings'):
             if fetch_content:
                 for reading in self.abbreviated_readings:
-                    await reading.pericope.aget_passage(language=self.language)
+                    await reading.pericope.aget_passage(language=self.language, translation=self.translation)
 
             return self.abbreviated_readings
 
@@ -586,7 +587,7 @@ async def aget_abbreviated_readings(self, fetch_content=False):
 
         if fetch_content:
             for reading in readings:
-                await reading.pericope.aget_passage(language=self.language)
+                await reading.pericope.aget_passage(language=self.language, translation=self.translation)
 
         self.abbreviated_readings = readings
         return readings