How half-pace/multimodal-agentic-ai-framework-for-virtual-reality-museum-navigation works
This repository is a document-ingestion and handloom-knowledge preparation project; the README names a virtual-reality museum, but the available implementation is focused on preparing knowledge content rather than museum navigation. The principal workflow discovers raw documents, extracts and cleans text, forms sections, tracks document versions, and can generate OKF knowledge files. Chunking is an additional capability. The graph reflects the shown ingestion calls and documented component behavior; it does not assume an unshown application or runtime connection to chunking or OKF generation.
Main entry points: processing.py Ingestion Processor · extraction.py PDF Extraction · okf.py OKF Generator · manifest.py Document Registry
Components 15
Ingestion
- PDF Extraction
script_ingestion/extraction.py - Text Cleaning
script_ingestion/cleaning.py - File Discovery
script_ingestion/ingestion/discovery.py - Ingestion Processor
script_ingestion/ingestion/processing.py - Document Registry
script_ingestion/ingestion/manifest.py - Manifest File
knowledge_base/document_manifest.json
Knowledge Modeling
- Document Identity
script_ingestion/knowledge/document.py - Normalized Documents
script_ingestion/knowledge/normalized.py - Document Sections
script_ingestion/knowledge/section.py
Knowledge Outputs
- OKF Generator
script_ingestion/knowledge/okf.py - OKF Knowledge Files
knowledge_base/04_okf/processes - Text Chunking
script_ingestion/knowledge/chunking.py
Knowledge Sources
- Raw Documents
knowledge_base/01_raw_data - Markdown Text
- OKF Process Sources
knowledge_base/04_okf/processes
Connections 13
- Raw Documents → PDF Extraction: extracts PDFs
script_ingestion/extraction.py - PDF Extraction → Markdown Text: writes Markdown
script_ingestion/extraction.py - Ingestion Processor → File Discovery: discovers files
script_ingestion/ingestion/processing.py - Ingestion Processor → PDF Extraction: extracts PDF text
script_ingestion/ingestion/processing.py - Ingestion Processor → Text Cleaning: normalizes text
script_ingestion/ingestion/processing.py - Ingestion Processor → Document Sections: creates sections
script_ingestion/ingestion/processing.py - Ingestion Processor → Normalized Documents: builds documents
script_ingestion/ingestion/processing.py - Ingestion Processor → Document Identity: hashes and creates
script_ingestion/ingestion/processing.py - Ingestion Processor → Document Registry: checks and updates
script_ingestion/ingestion/processing.py - Document Registry → Manifest File: loads and saves
script_ingestion/ingestion/manifest.py - OKF Generator → Document Sections: formats sections
script_ingestion/knowledge/okf.py - OKF Generator → OKF Knowledge Files: writes knowledge files
script_ingestion/knowledge/okf.py - Text Chunking → OKF Process Sources: can parse OKF
script_ingestion/knowledge/chunking.py