Handle GET requests for current progress status

Handle Server-Sent Events stream for real-time progress

Rename an existing session title

Handle chat within a specific session. Intelligently routes between direct LLM (fast) and RAG pipeline (document-aware).

Initialize the SQLite database with required tables

Initialize SQLite database for storing PDF content

Initialize the global PDF processor

Inspect LanceDB table to extract metadata for older indexes. Returns the inferred metadata or empty dict if inspection fails.

Show chunk_id, truncated text and score for the first *k* rows.

Run complete system health check

Main entry point for the script.

Main entry point for the demo script.

Main entry point.

Main function to initialize and start the server

Test the Ollama client

Run all validation checks.

Print a comprehensive map of which models are used where.

Return *docs* with their `text` field pruned sentence-wise. If the model could not be initialised we simply echo the input.

Pull a model if not available

Processes and indexes documents based on the pipeline's configuration. Accepts legacy keyword *documents* as an alias for *file_paths

Starts the advanced RAG API server.

Runs the agentic RAG pipeline for a given query. Returns the result as a JSON string.

Run the complete batch indexing demo.

Enhanced indexing function with real-time progress tracking

Interactive document selection.

Setup centralized logging with colors.

Loads and displays the knowledge graph.

Start all services in order.

Test the chat endpoint

Test conversation with history

Test the health endpoint

Test the created index with a sample query.

Test Ollama connectivity from within Docker container

Test the built-in validation function.

Switch embedding model at runtime and clear cached objects so they re-initialize.

Validate that all configurations are consistent.

Validate and filter document paths.

Async variant that calls the Ollama client asynchronously.