The dataset viewer is not available for this split.
Error code: FeaturesError
Exception: ArrowInvalid
Message: Schema at index 1 was different:
# versification scheme registry; standard = KJV. shoresh versification.build: string
__index_level_0__: string
__index_level_1__: string
__index_level_2__: string
vs
# hebrew versification -> KJV standard; single-verse diffs only (identity elsewhere). Source: STEPBible TVTMS (CC-BY). shoresh versification.build: string
__index_level_0__: string
__index_level_1__: string
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/split/first_rows.py", line 243, in compute_first_rows_from_streaming_response
iterable_dataset = iterable_dataset._resolve_features()
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 4379, in _resolve_features
features = _infer_features_from_batch(self.with_format(None)._head())
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2661, in _head
return next(iter(self.iter(batch_size=n)))
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2839, in iter
for key, pa_table in ex_iterable.iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2377, in _iter_arrow
yield from self.ex_iterable._iter_arrow()
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 580, in _iter_arrow
yield new_key, pa.Table.from_batches(chunks_buffer)
~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^
File "pyarrow/table.pxi", line 5039, in pyarrow.lib.Table.from_batches
c_table = GetResultValue(
File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
return check_status(status)
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
raise convert_status(status)
pyarrow.lib.ArrowInvalid: Schema at index 1 was different:
# versification scheme registry; standard = KJV. shoresh versification.build: string
__index_level_0__: string
__index_level_1__: string
__index_level_2__: string
vs
# hebrew versification -> KJV standard; single-verse diffs only (identity elsewhere). Source: STEPBible TVTMS (CC-BY). shoresh versification.build: string
__index_level_0__: string
__index_level_1__: stringNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Versification (V1) — KJV-standard scheme registry
Maps any Bible tradition's verse references to one standard: KJV. Versification is a per-version
property — a Bible not numbered like KJV carries a versification scheme, and its refs normalize to
KJV via that scheme's diffs. Built by shoresh/versification/build.py from STEPBible TVTMS
(CC-BY, the "Expanded Version": SourceType | SourceRef | StandardRef(=KJV) | Action).
Files
schemes.tsv— the registry:scheme, tvtms_sourcetype, n_diffs, standard.kjvis the standard (0 diffs).schemes/<scheme>.tsv— one file per non-standard scheme:source_ref, standard_ref, action. Diffs only — a ref not listed maps to KJV by identity. Refs areBOOK ch:v(USFM codes); a Psalm superscription is the verse tokentitle(HebrewPSA 3:1 → PSA 3:title).
Schemes built now (the ones our corpora + consumers need):
| scheme | tradition | our corpus | diffs |
|---|---|---|---|
hebrew |
Masoretic / WLC | BHSA spine | 2,031 |
lxx |
Septuagint / Rahlfs | lxx.db |
5,386 |
Resolver (shoresh data.py)
to_standard(scheme, ref) / from_standard(scheme, ref) — identity for kjv/unlisted; endpoint
/versify/{scheme}/{book}/{chapter}/{verse}.
Consumers
- X1 quotations — OT refs normalized to KJV (e.g. LXX
PSA 15:8 → PSA 16:8,JOL 3:1 → JOL 2:28); retired the old hardcoded Psalm map + thevrs=lxx?flag (0 remaining). - Aligner — tag each target Bible with its scheme; Psalm superscription offset (
hebrew) lets Psalm verses align.
For the aligner (build_corpus): target → Hebrew-source composition
The aligner's source is Hebrew (WLC); KJV is the pivot standard. To bring a target Bible verse (scheme X) onto the Hebrew source numbering so verse-by-verse matching lines up:
kjv_ref = to_standard(X, target_ref)— apply scheme X's forward map (source_ref→standard_ref; identity if the target is KJV-numbered, or uselxx.tsvif it follows LXX numbering).hebrew_ref = from_standard(hebrew, kjv_ref)— applyhebrew.tsvin reverse (standard_ref → source_ref).
So target_ref → kjv → hebrew_ref. Concretely for Psalms: a titled Psalm's superscription is Hebrew
v1, so KJV PSA 3:1 ↔ Hebrew PSA 3:2 (the hebrew.tsv row PSA 3:2 → PSA 3:1, read in reverse).
Identity for any ref not listed. This unblocks Psalm alignment for all methods (eflomal/gloss/neural),
not just neural. Tables are UTF-8 TSV, BOOK ch:v (USFM); title = a superscription verse token.
Scope (v1)
Single-verse remaps only (identity elsewhere). Out of scope, flagged for a later pass: ranges /
LXX sub-verses / concatenations (; - ! in TVTMS — long verses, merges/splits like the LXX 9|10
Psalm merge), the Latin/other traditions, and the NT sub-verse micro-differences. The single-verse
map covers the Psalm superscriptions + OT chapter/verse offsets, which is what X1 + the aligner need.
Per-version metadata (external — the design point)
Each Bible version declares its versification scheme; this is external metadata, not derivable
from content. Our own corpora auto-tag (hebrew, lxx, else kjv); externally-sourced Bibles
(the aligner's targets, helloAO/PKF) must supply it (default kjv, flag unknowns). Lives in the
bibles recipe layer — see internal-docs/bibles-recipe-layer.md.
Rebuild
python -m versification.build # downloads pinned TVTMS -> resources/versification/
- Downloads last month
- -