/create-text

Deterministic public-domain text corpus chunker with seeded selector.

Quick Start

# Build text banks from extracted datalake corpus
./run.sh build

# Build from Wikipedia (specific categories)
./run.sh build --source wikipedia --categories "Cybersecurity" "Avionics"

# Get 16 requirement-style text chunks from defense domain, seed 42
./run.sh select --content-type requirement --domain defense --count 16 --seed 42

# Get prose paragraphs from any domain
./run.sh select --content-type prose --count 8 --seed 7