"""Recount ten KJV words (hell, Satan, Enoch, 666, heaven, angels, fasting, forgiveness, heart, purgatory) in the pinned TrustBible KJV text. Python 3 standard library only. The compressed input is hash-checked before anything is written. Every rule is the Bible word counter's whole-word rule: any case, possessives included, hyphen-joined forms excluded. """ import argparse import csv import gzip import hashlib import io import json from pathlib import Path import re import urllib.error import urllib.request SOURCE_URL = "https://trustbible.org/data/kjv.jsonl.gz" SOURCE_SHA256 = "0c3b8afefa4efaa8bd72576682c8ad060e4ccb499e81943a9678de49ab9416ad" STANDALONE = r"(? 0 for k in ("chapter", "verse")), "Invalid verse reference.") key = (verse["book"], verse["chapter"], verse["verse"]) require(key not in keys, f"Duplicate verse: {key}") keys.add(key) require(len({v["book"] for v in records}) == 66, "Expected 66 books.") return records def reference(verse): return f"{verse['book']} {verse['chapter']}:{verse['verse']}" def context_url(verse): slug = verse["book"].lower().replace(" ", "-") return f"https://trustbible.org/en/bible/{slug}/{verse['chapter']}/#verse-{verse['verse']}" def pair(counts): return {"occurrences": sum(counts), "verses": sum(n > 0 for n in counts)} def recount_word(key, config, kjv, books): ot = set(books[:39]) refs = [reference(v) for v in kjv] sense = config.get("sense_filtered", {}) for allowed in sense.values(): require(set(allowed) <= set(refs), f"{key}: unknown sense reference.") forms, form_values, all_sense_values = {}, {}, {} for name, phrase in config["forms"].items(): regex = re.compile(pattern_for(phrase), re.IGNORECASE) matches = [list(regex.finditer(v["text"])) for v in kjv] raw = [len(m) for m in matches] allowed = set(sense.get(name, refs)) kept = [n if ref in allowed else 0 for n, ref in zip(raw, refs)] form_values[name], all_sense_values[name] = kept, raw casing = {} for verse_matches, ref in zip(matches, refs): if ref in allowed: for match in verse_matches: casing[match.group()] = casing.get(match.group(), 0) + 1 forms[name] = { "phrase": phrase, "pattern": pattern_for(phrase), "flags": "re.IGNORECASE", "totals": pair(kept), "casing": dict(sorted(casing.items())), "possessives": sum(v["text"][m.end():m.end() + 1] in ("'", "’") for v, ms, ref in zip(kjv, matches, refs) if ref in allowed for m in ms), **({"all_senses": pair(raw), "excluded_other_sense": pair([r - k for r, k in zip(raw, kept)])} if name in sense else {}), } headline = [sum(form_values[name][i] for name in config["headline"]) for i in range(len(kjv))] every = [sum(form_values[name][i] for name in config["forms"]) for i in range(len(kjv))] per_book = {book: pair([n for v, n in zip(kjv, headline) if v["book"] == book]) for book in books} hits = [i for i, n in enumerate(headline) if n] most = max(headline) related = {} for label, phrase in config["related"].items(): regex = re.compile(pattern_for(phrase), re.IGNORECASE) related[label] = {"phrase": phrase, "pattern": pattern_for(phrase), "flags": "re.IGNORECASE", **pair([len(regex.findall(v["text"])) for v in kjv]), "testaments": {t: pair([len(regex.findall(v["text"])) for v in kjv if (v["book"] in ot) == (t == "OT")]) for t in ("OT", "NT")}} result = { "question": config["question"], "headline_forms": config["headline"], "headline": pair(headline), "all_forms": pair(every), "testaments": {t: pair([n for v, n in zip(kjv, headline) if (v["book"] in ot) == (t == "OT")]) for t in ("OT", "NT")}, "forms": forms, "per_book": per_book, "books_ranked": [b for b, _ in sorted(((b, p) for b, p in per_book.items() if p["occurrences"]), key=lambda item: (-item[1]["occurrences"], books.index(item[0])))], "books_with_zero": [b for b, p in per_book.items() if p["occurrences"] == 0], "first": refs[hits[0]] if hits else None, "last": refs[hits[-1]] if hits else None, "most_in_one_verse": {"occurrences": most, "verses": [refs[i] for i in hits if headline[i] == most]} if hits else None, "verses": [{"reference": refs[i], "occurrences": headline[i], "context_url": context_url(kjv[i])} for i in hits], "related_counts_not_in_headline": related, } if sense: result["sense_rule"] = ("“fast” and “fastest” count only in the listed verses, where they mean going without food; " "elsewhere “fast” means firm or fixed (“hold fast”, “stand fast”, “fast asleep”).") result["sense_references"] = sorted(set().union(*map(set, sense.values())), key=refs.index) result["excluded_other_sense"] = [{"reference": refs[i], "form": name, "occurrences": all_sense_values[name][i] - form_values[name][i]} for i in range(len(kjv)) for name in sense if all_sense_values[name][i] - form_values[name][i]] if key == "enoch": result["people"] = {who: pair([headline[refs.index(r)] for r in rs]) | {"references": rs} for who, rs in ENOCH_PEOPLE.items()} require(sum(p["occurrences"] for p in result["people"].values()) == result["headline"]["occurrences"], "Every Enoch occurrence must be assigned to one person.") rows = [] for i, verse in enumerate(kjv): candidates = {name: all_sense_values[name][i] for name in config["forms"]} if not any(candidates.values()): continue rows.append({ "reference": refs[i], "testament": "OT" if verse["book"] in ot else "NT", "book": verse["book"], **{name: form_values[name][i] for name in config["forms"]}, "headline_occurrences": headline[i], "other_sense_occurrences": sum(all_sense_values[name][i] - form_values[name][i] for name in config["forms"]), "kjv_text": verse["text"], "context_url": context_url(verse), }) return result, rows def recount(kjv): books = list(dict.fromkeys(v["book"] for v in kjv)) require(books[38:40] == ["Malachi", "Matthew"], "Unexpected OT/NT ordering.") words, ledgers = {}, {} for key, config in WORDS.items(): words[key], ledgers[key] = recount_word(key, config, kjv, books) return words, ledgers def csv_bytes(key, rows): fields = ["reference", "testament", "book", *WORDS[key]["forms"], "headline_occurrences", "other_sense_occurrences", "kjv_text", "context_url"] output = io.StringIO(newline="") writer = csv.DictWriter(output, fieldnames=fields, lineterminator="\n") writer.writeheader() writer.writerows(rows) return output.getvalue().encode("utf-8-sig"), fields def main(): parser = argparse.ArgumentParser(description=__doc__) parser.add_argument("--input", type=Path, help="Offline copy of the exact pinned KJV gzip.") parser.add_argument("--output-dir", type=Path, default=Path(".")) args = parser.parse_args() try: kjv = load_source(args.input) words, ledgers = recount(kjv) files = {} for key, rows in ledgers.items(): if not rows: words[key]["audit"] = None continue data, fields = csv_bytes(key, rows) name = f"{key}-count-audit.csv" files[name] = data words[key]["audit"] = {"url": f"https://trustbible.org/data/{name}", "sha256": hashlib.sha256(data).hexdigest(), "rows": len(rows), "columns": fields, "encoding": "UTF-8 with BOM; comma-delimited, LF"} method = { "dataset": "Whole-word counts for ten frequently searched words in a pinned English KJV", "method_version": "1.0", "verified_on": "2026-11-01", "source": {"url": SOURCE_URL, "sha256": SOURCE_SHA256, "upstream": "https://github.com/aruljohn/Bible-kjv", "imported_on": "2026-09-06", "printing": "Unknown", "verse_records": len(kjv), "scope": "Verse text only; 66-book Protestant canon, no headings, notes, Apocrypha or metadata."}, "rule": ("Whole word or phrase, any case (re.IGNORECASE): a letter, digit, underscore or ASCII hyphen beside the match " "disqualifies it; possessives (’s) stay included. Same rule as https://trustbible.org/en/bible-word-counter/. " "An occurrence is one match; a verse counts once however many matches it holds."), "words": words, "reproduction": {"script_url": "https://trustbible.org/data/count-words.py", "script_sha256": hashlib.sha256(Path(__file__).read_bytes()).hexdigest(), "requires": "Python 3 standard library only; internet for the default input download.", "command": "python count-words.py --output-dir results", "offline_command": "python count-words.py --input kjv.jsonl.gz --output-dir results", "outputs": [*files, "word-counts-method.json"]}, "limits": ["A word count is not a count of every reference to a person, place or idea: pronouns, titles and synonyms are not counted.", "English words are not Hebrew or Greek source words; one English word can translate several source words and the reverse."], } args.output_dir.mkdir(parents=True, exist_ok=True) for name, data in files.items(): (args.output_dir / name).write_bytes(data) (args.output_dir / "word-counts-method.json").write_text(json.dumps(method, ensure_ascii=False, indent=2) + "\n", encoding="utf-8", newline="\n") except (OSError, ValueError, KeyError, urllib.error.URLError) as error: parser.exit(1, f"Recount failed: {error}\n") print("PASS: source hash and unique verse inventory verified.") for key, word in words.items(): print(f"{key}: {word['headline']['occurrences']:,} occurrences / {word['headline']['verses']:,} verses " f"(all forms {word['all_forms']['occurrences']:,} / {word['all_forms']['verses']:,})") print(f"Output: {args.output_dir.resolve()}") if __name__ == "__main__": main()