Update web/core/scripture_normalizer.py

This commit is contained in:
Joshua Laymon 2025-08-14 02:23:30 +00:00
parent cbb60cc3cd
commit 815d9c5af3

View File

@ -1,233 +1,223 @@
# core/scripture_normalizer.py
from __future__ import annotations
import re
from typing import Dict, List, Tuple, Optional
from typing import Dict, List, Optional, Tuple
# =========================
# --------------------------
# Canonical book abbreviations
# =========================
# House style (edit as you like). Keys are normalized to lower-case without periods.
# --------------------------
BOOK_CANON: Dict[str, str] = {
# ----- OT -----
"genesis": "Gen.", "gen": "Gen.", "ge": "Gen.", "gn": "Gen.", "genesisesis": "Gen.",
"exodus": "Ex.", "ex": "Ex.", "exo": "Ex.",
"leviticus": "Lev.", "lev": "Lev.", "le": "Lev.",
"numbers": "Num.", "num": "Num.", "nu": "Num.", "nums": "Num.",
"deuteronomy": "Deut.", "deut": "Deut.", "deu": "Deut.", "dt": "Deut.", "deutronomy": "Deut.",
"joshua": "Josh.", "josh": "Josh.", "jos": "Josh.",
"judges": "Judg.", "judg": "Judg.", "jdg": "Judg.",
"ruth": "Ruth", "ru": "Ruth",
"1 samuel": "1 Sam.", "i samuel": "1 Sam.", "1 sam": "1 Sam.", "1sam": "1 Sam.", "1sa": "1 Sam.",
"2 samuel": "2 Sam.", "ii samuel": "2 Sam.", "2 sam": "2 Sam.", "2sam": "2 Sam.", "2sa": "2 Sam.",
"1 kings": "1 Ki.", "i kings": "1 Ki.", "1 ki": "1 Ki.", "1kgs": "1 Ki.", "1 kgs": "1 Ki.", "1kings": "1 Ki.", "1ki": "1 Ki.",
"2 kings": "2 Ki.", "ii kings": "2 Ki.", "2 ki": "2 Ki.", "2kgs": "2 Ki.", "2 kgs": "2 Ki.", "2kings": "2 Ki.", "2ki": "2 Ki.",
"1 chronicles": "1 Chron.", "i chronicles": "1 Chron.", "1 chron": "1 Chron.", "1ch": "1 Chron.", "1chron": "1 Chron.",
"2 chronicles": "2 Chron.", "ii chronicles": "2 Chron.", "2 chron": "2 Chron.", "2ch": "2 Chron.", "2chron": "2 Chron.",
"ezra": "Ezra", "ezr": "Ezra",
# OT
"genesis": "Gen.", "gen": "Gen.", "genesisesis": "Gen.",
"exodus": "Ex.", "ex": "Ex.",
"leviticus": "Lev.", "lev": "Lev.",
"numbers": "Num.", "num": "Num.", "nums": "Num.",
"deuteronomy": "Deut.", "deut": "Deut.", "deutronomy": "Deut.", "deu": "Deut.",
"joshua": "Josh.", "josh": "Josh.",
"judges": "Judg.", "judg": "Judg.",
"ruth": "Ruth",
"1 samuel": "1 Sam.", "i samuel": "1 Sam.", "1 sam": "1 Sam.", "1sam": "1 Sam.",
"2 samuel": "2 Sam.", "ii samuel": "2 Sam.", "2 sam": "2 Sam.", "2sam": "2 Sam.",
"1 kings": "1 Ki.", "i kings": "1 Ki.", "1 ki": "1 Ki.", "1kgs": "1 Ki.", "1 kgs": "1 Ki.",
"2 kings": "2 Ki.", "ii kings": "2 Ki.", "2 ki": "2 Ki.", "2kgs": "2 Ki.", "2 kgs": "2 Ki.",
"1 chronicles": "1 Chron.", "i chronicles": "1 Chron.", "1 chron": "1 Chron.",
"2 chronicles": "2 Chron.", "ii chronicles": "2 Chron.", "2 chron": "2 Chron.",
"ezra": "Ezra",
"nehemiah": "Neh.", "neh": "Neh.",
"esther": "Esth.", "esth": "Esth.", "es": "Esth.",
"esther": "Esth.", "esth": "Esth.",
"job": "Job",
"psalm": "Ps.", "psalms": "Ps.", "ps": "Ps.", "psm": "Ps.", "pss": "Ps.",
"proverbs": "Prov.", "prov": "Prov.", "pr": "Prov.",
"ecclesiastes": "Eccl.", "eccles": "Eccl.", "eccl": "Eccl.", "ecc": "Eccl.", "ec": "Eccl.",
"song of solomon": "Song", "song of songs": "Song", "song": "Song", "so": "Song", "sos": "Song",
"isaiah": "Isa.", "isa": "Isa.", "is": "Isa.",
"jeremiah": "Jer.", "jer": "Jer.", "je": "Jer.",
"lamentations": "Lam.", "lam": "Lam.", "la": "Lam.",
"ezekiel": "Ezek.", "ezek": "Ezek.", "eze": "Ezek.", "ezk": "Ezek.",
"daniel": "Dan.", "dan": "Dan.", "da": "Dan.",
"hosea": "Hos.", "hos": "Hos.", "ho": "Hos.",
"joel": "Joel", "joe": "Joel", "jl": "Joel",
"amos": "Amos", "am": "Amos",
"obadiah": "Obad.", "obad": "Obad.", "ob": "Obad.",
"psalms": "Ps.", "psalm": "Ps.", "ps": "Ps.", "pss": "Ps.",
"proverbs": "Prov.", "prov": "Prov.",
"ecclesiastes": "Eccl.", "eccles": "Eccl.", "eccl": "Eccl.",
"song of solomon": "Song", "song of songs": "Song", "song": "Song",
"isaiah": "Isa.", "isa": "Isa.", # IMPORTANT: direct alias so 'Isa.' never becomes '1 Sam.'
"jeremiah": "Jer.", "jer": "Jer.",
"lamentations": "Lam.", "lam": "Lam.",
"ezekiel": "Ezek.", "ezek": "Ezek.",
"daniel": "Dan.", "dan": "Dan.",
"hosea": "Hos.", "hos": "Hos.",
"joel": "Joel",
"amos": "Amos",
"obadiah": "Obad.", "obad": "Obad.",
"jonah": "Jon.", "jon": "Jon.",
"micah": "Mic.", "mic": "Mic.",
"nahum": "Nah.", "nah": "Nah.",
"habakkuk": "Hab.", "hab": "Hab.",
"zephaniah": "Zeph.", "zeph": "Zeph.", "zep": "Zeph.",
"zephaniah": "Zeph.", "zeph": "Zeph.",
"haggai": "Hag.", "hag": "Hag.",
"zechariah": "Zech.", "zech": "Zech.", "zec": "Zech.",
"zechariah": "Zech.", "zech": "Zech.",
"malachi": "Mal.", "mal": "Mal.",
# ----- NT -----
"matthew": "Matt.", "matt": "Matt.", "mt": "Matt.", "mat": "Matt.",
"mark": "Mark", "mrk": "Mark", "mk": "Mark", "mr": "Mark",
# NT
"matthew": "Matt.", "matt": "Matt.", "mt": "Matt.",
"mark": "Mark", "mk": "Mark", "mrk": "Mark", "mr": "Mark",
"luke": "Luke", "lk": "Luke",
"john": "John", "jn": "John", "jo": "John", "joh": "John",
"acts": "Acts", "act": "Acts", "ac": "Acts",
"romans": "Rom.", "rom": "Rom.", "ro": "Rom.", "rm": "Rom.",
"1 corinthians": "1 Cor.", "i corinthians": "1 Cor.", "1 cor": "1 Cor.", "1cor": "1 Cor.", "1co": "1 Cor.", "1 corinthians": "1 Cor.",
"2 corinthians": "2 Cor.", "ii corinthians": "2 Cor.", "2 cor": "2 Cor.", "2cor": "2 Cor.", "2co": "2 Cor.", "2 corinthians": "2 Cor.",
"galatians": "Gal.", "gal": "Gal.", "ga": "Gal.",
"ephesians": "Eph.", "eph": "Eph.", "eph.": "Eph.",
"philippians": "Phil.", "phil": "Phil.", "php": "Phil.", "phi": "Phil.", "philippians216": "Phil.",
"john": "John", "jn": "John", "joh": "John",
"acts": "Acts",
"romans": "Rom.", "rom": "Rom.",
"1 corinthians": "1 Cor.", "i corinthians": "1 Cor.", "1 cor": "1 Cor.", "1co": "1 Cor.",
"2 corinthians": "2 Cor.", "ii corinthians": "2 Cor.", "2 cor": "2 Cor.", "2co": "2 Cor.",
"galatians": "Gal.", "gal": "Gal.",
"ephesians": "Eph.", "eph": "Eph.",
"philippians": "Phil.", "phil": "Phil.", "philippians 216": "Phil.", # import glitch seen
"colossians": "Col.", "col": "Col.",
"1 thessalonians": "1 Thess.", "i thessalonians": "1 Thess.", "1 thess": "1 Thess.", "1 th": "1 Thess.", "1thess": "1 Thess.",
"2 thessalonians": "2 Thess.", "ii thessalonians": "2 Thess.", "2 thess": "2 Thess.", "2 th": "2 Thess.", "2thess": "2 Thess.",
"1 timothy": "1 Tim.", "i timothy": "1 Tim.", "1 tim": "1 Tim.", "1ti": "1 Tim.", "1 timothy": "1 Tim.",
"2 timothy": "2 Tim.", "ii timothy": "2 Tim.", "2 tim": "2 Tim.", "2ti": "2 Tim.", "2 timothy": "2 Tim.",
"titus": "Titus", "tit": "Titus", "ti.": "Titus",
"philemon": "Philem.", "philem": "Philem.", "phm": "Philem.",
"1 thessalonians": "1 Thess.", "i thessalonians": "1 Thess.", "1 thess": "1 Thess.",
"2 thessalonians": "2 Thess.", "ii thessalonians": "2 Thess.", "2 thess": "2 Thess.",
"1 timothy": "1 Tim.", "i timothy": "1 Tim.", "1 tim": "1 Tim.", "1ti": "1 Tim.",
"2 timothy": "2 Tim.", "ii timothy": "2 Tim.", "2 tim": "2 Tim.", "2ti": "2 Tim.",
"titus": "Titus", "tit": "Titus",
"philemon": "Philem.", "philem": "Philem.",
"hebrews": "Heb.", "heb": "Heb.",
"james": "Jas.", "jas": "Jas.", "jam": "Jas.", "jms": "Jas.",
"1 peter": "1 Pet.", "i peter": "1 Pet.", "1 pet": "1 Pet.", "1pe": "1 Pet.",
"2 peter": "2 Pet.", "ii peter": "2 Pet.", "2 pet": "2 Pet.", "2pe": "2 Pet.",
"1 john": "1 John", "i john": "1 John", "1jn": "1 John", "1 jo": "1 John",
"2 john": "2 John", "ii john": "2 John", "2jn": "2 John", "2 jo": "2 John",
"3 john": "3 John", "iii john": "3 John", "3jn": "3 John", "3 jo": "3 John",
"jude": "Jude", "jud": "Jude",
"revelation": "Rev.", "rev": "Rev.", "re": "Rev.",
"james": "Jas.", "jas": "Jas.",
"1 peter": "1 Pet.", "i peter": "1 Pet.", "1 pet": "1 Pet.",
"2 peter": "2 Pet.", "ii peter": "2 Pet.", "2 pet": "2 Pet.",
"1 john": "1 John", "i john": "1 John",
"2 john": "2 John", "ii john": "2 John",
"3 john": "3 John", "iii john": "3 John",
"jude": "Jude",
"revelation": "Rev.", "rev": "Rev.",
}
def _variants() -> Dict[str, str]:
base = dict(BOOK_CANON)
numbered = [
("samuel", "Sam."), ("kings", "Ki."), ("chronicles", "Chron."),
("corinthians", "Cor."), ("thessalonians", "Thess."),
("timothy", "Tim."), ("peter", "Pet."), ("john", "John"),
]
for n, prefix in (("1", "1"), ("i", "1"), ("2", "2"), ("ii", "2"), ("3", "3"), ("iii", "3")):
for name, abbr in numbered:
base[f"{n} {name}"] = f"{prefix} {abbr}"
base[f"{n}{name}"] = f"{prefix} {abbr}"
base[f"{n}{name[:2]}"] = f"{prefix} {abbr}"
base["ps"] = "Ps."
base["prov"] = "Prov."
base["eccles"] = "Eccl."
base["deut."] = "Deut."
base["mt"] = "Matt."
base["mk"] = "Mark"
base["lk"] = "Luke"
base["jn"] = "John"
base["ti"] = "Tim."
base["co"] = "Cor."
return base
# Short bare tokens that must *never* be treated as “roman numeral + book”
# (prevents 'Isa.' => '1 Sam.' and similar).
NEVER_PREFIX_NUMERAL = {"isa", "isaiah", "job", "joel", "amos", "nah", "hag", "mal", "rom", "gal", "eph", "tit", "heb", "jas", "jude"}
BOOK_MAP = _variants()
def _canon_key(s: str) -> str:
s = (s or "").strip().lower()
s = s.replace(".", " ")
s = re.sub(r"\s+", " ", s)
return s
def _direct_lookup(key: str) -> Optional[str]:
"""Try direct lookups with and without spaces (e.g., 'i samuel', '1co')."""
if key in BOOK_CANON:
return BOOK_CANON[key]
nospace = key.replace(" ", "")
if nospace in BOOK_CANON:
return BOOK_CANON[nospace]
return None
# Split helpers
CRUFT_RE = re.compile(r"\b(read|see|chap(?:ter)?|ch)\b\.?", re.I)
BOOK_RE = re.compile(
r"""
^\s*
(?:
(?P<num>[1-3]|i{1,3})\s*
)?
\s*
(?P<book>[A-Za-z\.]+(?:\s+[A-Za-z\.]+){0,2})
""",
re.X,
)
C_V_RE = re.compile(
r"""
(?:
(?P<ch>\d+)
(?:
:(?P<vs>[\d,\-\u2013\u2014\s]+)
)?
)
""",
re.X,
)
def _clean_text(s: str) -> str:
s = s.replace("\xa0", " ").replace("\u2009", " ").replace("\u202f", " ")
s = CRUFT_RE.sub("", s)
s = s.replace("..", ".").replace("", "-").replace("", "-")
s = re.sub(r"\s{2,}", " ", s)
s = s.replace("", "-").replace("", "-")
s = re.sub(r"\s+", " ", s)
return s.strip(" ;,.\t\r\n ")
def _canon_key(raw: str) -> str:
key = raw.lower().strip()
key = key.replace(".", "")
key = re.sub(r"\s+", " ", key)
return key
BOOK_PREFIX_RE = re.compile(
r"""
^\s*
(?:(?P<num>[1-3]|i{1,3})\s*)? # optional leading number (1/2/3 or i/ii/iii)
(?P<book>[A-Za-z\.]+(?:\s+[A-Za-z\.]+){0,2}) # 1-3 words for the book
""",
re.X,
)
def _canon_book(book_raw: str, num: Optional[str] = None) -> Optional[str]:
C_V_RE = re.compile(r"(?P<ch>\d+)(?::(?P<vs>[\d,\-\s]+))?")
def _canon_book(book_raw: str, num: Optional[str]) -> Optional[str]:
"""
Map a raw book token to our canonical abbreviation.
Strategy:
1) Try exact/alias matches (with dots removed)
2) Try the same again with ALL spaces removed (so 'i sa' -> 'isa')
3) Only then, try the 'leading 1/2/3 or i/ii/iii + rest' heuristic.
Decide the canonical book abbreviation.
Rules:
1) Always try direct alias matches first (with/without spaces).
2) If the *book* piece looks like a short token that could be misread as a roman
numeral + another book (e.g., 'Isa'), *never* do numeral heuristics.
3) Otherwise, allow '1/2/3 + co/cor/ti/tim/pet/peter/jn/john' style heuristics.
"""
key = _canon_key(book_raw) # lower, trimmed, dots removed, single spaces
key_nospace = key.replace(" ", "") # used to catch 'i sa' -> 'isa'
key = _canon_key(book_raw)
# Step 1: direct lookups
direct = _direct_lookup(key)
if direct:
return direct
# 1) direct lookup (e.g., 'isa', 'isaiah', 'ps', '1 corinthians', etc.)
if key in BOOK_MAP:
return BOOK_MAP[key]
if key_nospace in BOOK_MAP:
return BOOK_MAP[key_nospace]
# Guard: tokens like 'isa' must not be split into 'i' + 'sa'
bare = key.replace(" ", "")
if bare in NEVER_PREFIX_NUMERAL:
return _direct_lookup(bare)
# 2) special explicit “number + short key” combos
if num and key in ("co", "cor"):
return f"{num} Cor."
if num and key in ("ti", "tim"):
return f"{num} Tim."
# Step 3: numeral + short token combos
if num:
n = {"i": "1", "ii": "2", "iii": "3"}.get(num.lower(), num)
short = key.replace(" ", "")
# common short targets where number matters
map_num = {
"co": "Cor.", "cor": "Cor.",
"ti": "Tim.", "tim": "Tim.",
"pe": "Pet.", "pet": "Pet.", "peter": "Pet.",
"jo": "John", "jn": "John", "joh": "John", "john": "John",
"sa": "Sam.", "sam": "Sam.", "samuel": "Sam.",
"ki": "Ki.", "kgs": "Ki.", "kings": "Ki.",
"chronicles": "Chron.", "chron": "Chron.",
"thessalonians": "Thess.", "thess": "Thess.",
}
if short in map_num:
return f"{n} {map_num[short]}"
# 3) heuristic: split a leading number/roman from the rest (e.g., 'i john' or '2ki')
m = re.match(r"(?P<n>[1-3]|i{1,3})(?P<rest>[A-Za-z].*)$", key_nospace)
if m:
n = {"i": "1", "ii": "2", "iii": "3"}.get(m.group("n"), m.group("n"))
rest = m.group("rest")
if rest in BOOK_MAP:
base = BOOK_MAP[rest]
# If base already begins with a number (rare in our map), replace it with n
if base[0] in "123":
return f"{n} {base.split(' ', 1)[1]}"
return f"{n} {base}"
# 4) lastresort: try the raw key against canonical map
return BOOK_CANON.get(key_nospace) or BOOK_CANON.get(key)
# Fallback
return None
def _parse_segment(seg: str, last_book: Optional[str]) -> Tuple[Optional[str], Optional[str], bool]:
original = seg.strip()
s = _clean_text(seg)
if not s:
"""
Parse one semicolon-delimited piece.
Returns (book_canon, cv, preserve_original_if_unparsed).
"""
raw = _clean_text(seg)
if not raw:
return (None, None, False)
m = BOOK_RE.match(s)
book = None
rest = s
m = BOOK_PREFIX_RE.match(raw)
book_part = None
num = None
rest = raw
if m:
num = (m.group("num") or "").strip()
raw_book = (m.group("book") or "").strip()
raw_joined = f"{num} {raw_book}".strip()
canon = _canon_book(raw_joined or raw_book, num=num or None)
if canon:
book = canon
rest = s[m.end():].strip(",;: .")
else:
return (None, original, True)
# What the regex thinks is a numeric prefix and a book word blob:
num = (m.group("num") or "").strip() or None
book_text = (m.group("book") or "").strip()
rest = raw[m.end():].strip(" :,;.")
if not book:
if last_book:
book = last_book
else:
return (None, original, True)
# Try to resolve the book using robust rules
book_part = _canon_book(book_text if not num else f"{num} {book_text}", num=num)
rest = re.sub(r"\s+", "", rest)
rest = re.sub(r":\s*", ":", rest)
# If regex split out a bogus 'i' from something like "Isa.", fix by trying full token directly.
if not book_part:
whole = _canon_book(book_text, None)
if whole:
book_part = whole
rest = raw[m.end():].strip(" :,;.")
# no book found → inherit previous if we have verses; otherwise preserve as-is
if not book_part:
# If there's obvious chapter/verse, but no book, we cannot link → preserve as-is.
if C_V_RE.search(rest):
return (None, None, True)
# Or the whole thing might already be a known book alone:
whole = _canon_book(raw, None)
if whole:
return (whole, None, False)
return (None, None, True)
# normalize chapter/verse
rest = rest.replace(" ", "")
if not rest:
return (book, None, False)
return (book_part, None, False)
if C_V_RE.search(rest):
cv = rest.replace(" ", "")
else:
m2 = re.search(r"\d+(?::[\d,\-]+)?", rest)
cv = m2.group(0).replace(" ", "") if m2 else None
mcv = C_V_RE.search(rest)
if mcv:
cv = rest
cv = cv.replace(" ", "")
return (book_part, cv, False)
if not cv:
return (book, None, False)
cv = cv.replace("", "-").replace("", "-")
cv = re.sub(r",\s*", ",", cv)
return (book, cv, False)
# not a recognizable cv → keep original piece untouched
return (None, None, True)
def normalize_scripture_field(text: str) -> Tuple[str, List[str]]:
"""
Normalize a whole scripture_raw string.
Returns (normalized_text, warnings).
Unknown/unparseable chunks are preserved verbatim.
"""
warnings: List[str] = []
if not text:
return ("", warnings)
@ -238,13 +228,10 @@ def normalize_scripture_field(text: str) -> Tuple[str, List[str]]:
for piece in pieces:
book, cv, preserve = _parse_segment(piece, last_book)
if preserve:
out.append(cv or piece.strip())
continue
if book and cv:
out.append(f"{book} {cv}")
last_book = book
if preserve:
# Keep the original chunk
out.append(_clean_text(piece))
continue
if book and not cv:
@ -252,9 +239,14 @@ def normalize_scripture_field(text: str) -> Tuple[str, List[str]]:
last_book = book
continue
if piece.strip():
out.append(piece.strip())
if not book and cv:
warnings.append(f"Missing book for '{piece.strip()}'")
continue
norm = "; ".join(x.strip() for x in out if x.strip())
if book and cv:
out.append(f"{book} {cv}")
last_book = book
norm = "; ".join(s for s in (o.strip(" ;,") for o in out) if s)
norm = re.sub(r"\s+", " ", norm).strip(" ;,")
return (norm, warnings)