Parses Chinese Union Version (UNV) biblical text with Strong's Numbers into structured semantic groups according to SPECIFICATION_v1.8.md. Includes automatic generic compound detection supporting all types of compounds (מִן, לִפְנֵי, etc.). Use when the user requests parsing UNV+SN verses, batch processing biblical text, or analyzing Strong's number groupings.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
Parses Chinese Union Version (UNV) biblical text with Strong's Numbers into structured semantic groups according to SPECIFICATION_v1.8.md. Includes automatic generic compound detection supporting all types of compounds (מִן, לִפְנֵי, etc.). Use when the user requests parsing UNV+SN verses, batch processing biblical text, or analyzing Strong's number groupings.
allowed-tools
Read, Write, Bash, Grep, Glob
UNV+SN Backparse Skill (Specification v1.8)
This skill parses Chinese Union Version (UNV) biblical text with Strong's Numbers into structured semantic groups according to SPECIFICATION_v1.8.md.
v1.7 Feature: Automatic compound preposition detection - resolves 94% of qb_qp_mismatch errors by detecting מִן (04480) compounds directly from qp.php wform field.
v1.7.2 Enhancement: Multi-token compound detection - automatically skips 900x prefixes to detect complex compounds like <04480><09001><06440> (מִלִּפְנֵי).
v1.8 Enhancement: Generic compound detection - supports all types of compound prepositions, not just מִן. Detects לִפְנֵי (<09001><06440>) and other 900x-starting compounds by checking both wform and remark fields in qp.php.
When to Use This Skill
Activate this skill when the user:
Requests parsing of specific biblical verses (e.g., "parse Genesis 1:1", "parse all verses in Exodus 2")
Wants to batch process a range of verses
Needs to analyze Strong's number groupings in UNV text
Asks about the parsing results or output format
Requests verification of parsed output
Core Architecture
The system uses a three-stage pipeline:
Data Retrieval (fetch_text.sh) - Fetches from FHL API endpoints
Parsing (parse_verse_v1_8.py via run_parser_temp.py) - Transforms raw data into structured groups
v1.7: Automatically detects compound prepositions from qp.php wform
v1.8: Generic detection for all compound types
Output Generation - Saves to output/{Book}/{Chapter}/{verse} (text format)
Parsing Workflow
Single Verse Parsing
# Parse a single verse (e.g., Genesis 1:1)
python run_parser_temp.py 1 1
# View output without writing to disk
python run_parser_temp.py --no-write 1 1
Batch Parsing
Follow the Batch_Parsing_SOP.md workflow:
Create Output Directories
mkdir -p output/{Book}/{Chapter}/
Determine Verse Range
If only book is provided, automatically determine starting point
For existing book: continue from last processed verse
For new book: start from chapter 1, verse 1
Iterate and Process Each Verse
for verse in {START..END}; do
python run_parser_temp.py {chapter} $versedone
[2025-11-25 01:57:42] Gen 3:14 | qb_qp_mismatch | Strong's number <03212> from qb.php not found in qp.php records. | KJV also uses <03212>
[2025-11-24 23:40:07] Gen 2:2 | prep_noun_compound | Prep+noun compound detected: <04480><03605> = מִכָּל (介系詞 מִן + 名詞，單陽附屬形) - not merged per config
[2025-11-25 01:50:43] Gen 3:16 | dangling_p900x | 900x prefix <09002> had no following Strong's number to attach to.
Important User Presentation Rules
CRITICAL: When showing parsed verse results to the user:
Present all three sections in order (Parsed Text → Raw Source → Morphology Notes)
Display sections as-is with no inserted commentary
After all three sections, you MAY add English explanations if helpful
DO NOT insert English translations or bullet points within the sections
DO NOT add commentary between sections
Testing Strategy
Verified Test Cases (SPECIFICATION_v1.8.md §7):
Gen 1:2 - Brace preposition right-attach + construct state
Gen 1:4 - Object marker handling with multiple {<0853>}
Gen 1:5 - FHL profile mapping with inferred vs explicit prefixes
Gen 3:5 - Verb left-attach exception for infinitive complement
Gen 4:16 (v1.7.2) - Multi-token compound <04480><09001><06440> detection
Gen 6:11 (v1.8) - לִפְנֵי compound <09001><06440> detection
Validate parsed output against expected groupings in spec §7.
