# PRAXIS Scripture word-study data

These publicly accessible chapter files are reference data, separate from application code. No authentication, proprietary claim or additional reuse restriction is imposed on them. `manifest.json` identifies every upstream revision, archive SHA-256, source edition payload hash, omitted reference, attribution and chapter path. The adapted Greek annotation data remains CC BY-SA 3.0.

## Hebrew annotations

Original work of the Open Scriptures Hebrew Bible available at https://github.com/openscriptures/morphhb

License: CC BY 4.0, https://creativecommons.org/licenses/by/4.0/

The underlying Westminster Leningrad Codex text is public domain. Lemma values in this dataset are augmented Strong identifiers (including prefix and sense information), not English definitions. The source morphology codes describe Hebrew and Biblical Aramaic words. `/` separates morphemes.

## Greek annotations

Tauber, J. K., ed. (2017) MorphGNT: SBLGNT Edition. Version 6.12 [Data set]. https://github.com/morphgnt/sblgnt DOI: 10.5281/zenodo.376200

Morphological parsing, lemmatization and these adaptations: CC BY-SA 3.0, https://creativecommons.org/licenses/by-sa/3.0/

Underlying text: The Greek New Testament: SBL Edition. Copyright © 2010 Society of Biblical Literature and Logos Bible Software. Current text license CC BY 4.0, https://creativecommons.org/licenses/by/4.0/ (https://github.com/LogosBible/SBLGNT).

## Changes and verification

The source datasets have been repackaged into chapter JSON. Only complete verses whose running letters and diacritics match the previously imported edition are included. Comparison ignores punctuation, spacing and Hebrew morpheme separators; Hebrew paragraph markers are compared as source running-text markers. Unicode canonical decomposition is used for comparison only. Existing Scripture text and source word spelling are preserved. Direct Hebrew words are used; alternative readings inside notes are excluded. Annotation-source word forms omit punctuation and retain Hebrew morpheme separators. Text differences or missing lemma/morphology values withhold the entire verse. This is not an English interlinear, dictionary, critical apparatus or theological interpretation.

Each verse records a SHA-256 of the exact imported Scripture text. The client checks that hash against the verse it is displaying before presenting annotations. A changed database verse therefore cannot silently inherit stale annotations. See `scripts/prepareScriptureStudy.py` in the source repository for the transformation.

Word arrays contain `[word, lemma, morphology, partOfSpeech]`; the last field is empty for Hebrew, where part of speech is encoded within morphology. Source code guides: https://github.com/openscriptures/morphhb/blob/master/parsing/HebrewMorphologyCodes.html and https://github.com/morphgnt/sblgnt#parsing-code