How rohit318/unstructured-rag works
This is a Streamlit unstructured-document RAG application. Its evidenced workflow is model creation, upload dispatch, file extraction/chunking and enrichment, then graph storage; users can browse model records and manage stored content. The home page describes query retrieval and LLM answers, but utils/chat_model.py is empty in the supplied source, so that documented capability is shown without inferred internal wiring. Source excerpts are partial, so omitted links may exist.
Main entry points: data_model.py Upload and model view · data_model_creation.py Model records · neo4j_ingestion.py Graph persistence · index.py Page navigation
Components 14
User interface
- Page navigation
index.py - Upload and model view
utils/data_model.py - Home overview
utils/home.py - Chat assistant
utils/chat_model.py
Ingestion
- Upload dispatcher
utils/data_model_ingestion.py - File processing
utils/data_pipeline.py - Image analysis
utils/img_to_txt.py
Model management
- Model sidebar
utils/data_model_sidebar.py - Model details
utils/data_model_details.py - Model records
utils/data_model_creation.py
Persistence
- Graph persistence
utils/neo4j_ingestion.py - Neo4j graph
- Image and table files
Other components
- User
Connections 22
- User → Page navigation: opens app
- Page navigation → Home overview: dispatches
- Page navigation → Upload and model view: dispatches
- Page navigation → Chat assistant: dispatches
- User → Upload and model view: uploads files
- Upload and model view → Model sidebar: renders
- Upload and model view → Model details: renders
- Upload and model view → Upload dispatcher: submits files
- Upload and model view → Model records: records submission
- Upload dispatcher → File processing: dispatches files
- File processing → Image analysis: enriches images
- File processing → Graph persistence: stores elements
- Graph persistence → Neo4j graph: reads and writes
- Graph persistence → Image and table files: tracks file paths
- Model records → Neo4j graph: updates model metadata
- Model records → Image and table files: manages artifacts
- Model sidebar → Model records: reads records
- Model sidebar → Graph persistence: lists models
- Model details → Model records: reads and edits
- Model details → Upload dispatcher: adds files
- Model details → Graph persistence: reads counts
- User → Chat assistant: asks questions