4 articles
An internal NLP pipeline that turns screenplay PDFs into structured data: extraction, character psychology modeling, and a resilient orchestration engine.
Deconstructing Personality with AI
A technical deep dive into prompt engineering, multi-pass architecture, and the 'Big Five' personality synthesis in scenario-parser.
Event-Driven Architecture & Graph Theory
A deep dive into the event-driven architecture, concurrency control, and graph theory algorithms powering scenario-parser's analysis engine.
Solving the Screenplay Parsing Challenge
How scenario-parser transforms raw PDF binaries into semantic screenplay data using PyMuPDF4LLM, heuristic state machines, and fuzzy string matching.
3-Stage Competitive Parsing Architecture
A complete system architecture overview of scenario-parser's 3-stage competitive parsing pipeline: extraction, 4 parallel parsing strategies with quality selection, and graph-theory + LLM hybrid analysis.