BSB / MSB Enriched USX

USX editions of the Berean Standard Bible (BSB) and the Majority Standard Bible (MSB), enriched in place with:

1. Inline Strong's numbers on every translated word.
2. Words-of-Jesus (red-letter) markup.

Both editions are the same translation; the MSB differs from the BSB only in the New Testament, where it follows the Majority Text (e.g. Romans 8:1 and the Matthew 6:13 doxology carry the longer Byzantine readings). The Old Testament is identical between the two.

Layout

bsb/                     Berean Standard Bible
  USX_1/*.usx            66 enriched book files
  styles.xml, versification.vrs, eng_en.ldml
msb/                     Majority Standard Bible
  USX_1/*.usx            66 enriched book files
  styles.xml, versification.vrs, eng_en.ldml, metadata.xml, license.xml

bsb_tables.tsv           BSB interlinear table (word -> Strong's), OT + NT
msb_nt_tables.tsv        MSB NT interlinear table (Majority Text)
words_of_jesus.jsonl     per-verse red-letter word runs for the BSB

add_strongs.py           inserts inline Strong's numbers
add_words_of_jesus.py    adds words-of-Jesus (red-letter) markup to the BSB

Enrichments

1. Strong's numbers (add_strongs.py)

Every translated word/phrase is wrapped in the USX wordlist character style with a strong attribute. Identifiers use the spec format: H/G prefix + the number zero-padded to five digits (Hebrew & Aramaic → H, Greek → G).

<verse number="1" .../><char style="w" strong="H07225">In the beginning</char>
<char style="w" strong="H00430">God</char> <char style="w" strong="H01254">created</char> …

The mapping comes from the interlinear tables, whose English text reconstructs the verse exactly. The OT and the BSB NT come from bsb_tables.tsv; the MSB NT comes from msb_nt_tables.tsv. Resolution is word-aligned per verse, so punctuation, quotation marks and footnotes stay outside the tags.

• BSB: 381,950 elements · MSB: 383,324 elements
• Every verse matched a table; ~22 words per edition are left untagged (stray source artifacts such as a literal vvv marker).

2. Words of Jesus (add_words_of_jesus.py)

Jesus's spoken words are wrapped in the red-letter character style, <char style="wj">, nested outside the Strong's markup (with the opening/closing quotation marks pulled inside the span).

<char style="wj">“<char style="w" strong="G03107">Blessed are</char>
  <char style="w" strong="G03588">the</char> …</char>

The MSB already ships with red-letter markup, so only the BSB needed it. The plain "bsb2usfm" BSB has none; the publisher's richer BSB export carries it and its NT text matches ours word-for-word, so the spans transfer by word index (captured in words_of_jesus.jsonl) with no fuzzy alignment. Spans are split at paragraph boundaries to stay well-formed.

• BSB: 2,302 wj spans added · MSB: 2,319 (pre-existing, untouched).

Regenerating

The scripts edit the USX files in place and only ever add markup — existing scripture text is preserved byte-for-byte (verified by XML well-formedness plus a scripture-word diff). All are idempotent: a file that already carries the markup is skipped, so re-running is safe.

python3 add_strongs.py                  # add Strong's numbers (BSB + MSB)
python3 add_words_of_jesus.py --apply   # add red-letter markup to the BSB

(Omit --apply to dry-run and print stats only.)

add_words_of_jesus.py reads the committed words_of_jesus.jsonl. To regenerate that file from a wj-marked BSB export, run python3 add_words_of_jesus.py --generate '<export>/USX_1/*.usx'.

Order matters: run add_strongs.py first — the red-letter nesting keys off the Strong's markup it adds.