DocsChat

RAG chat system: ingest documents (PDF, DOCX, TXT, MD), embed them in ChromaDB, and chat with them using any LLM.

Exposes both a FastAPI REST API and a Typer CLI.

Features

• Multi-format ingestion — PDF, DOCX, TXT, Markdown (max 50MB per file)
• Pluggable embeddings — OpenAI or local sentence-transformers
• Multi-LLM support — OpenAI, Gemini, Ollama
• Vector search — ChromaDB with persistent storage + standalone search endpoint
• Multi-turn chat — Conversation history with query rephrasing
• Streaming — SSE (API) and Rich Live (CLI) with configurable timeout
• REST API — Full CRUD for collections, documents, and conversations
• Input validation — Question length, collection names, file types, provider names
• Rate limiting — Per-IP sliding window (default 30 req/min)
• Connection pooling — Singleton SQLite connection with WAL mode

Quick Start

# Clone and install
git clone https://github.com/brolyroly007/docschat.git
cd docschat
pip install -r requirements.txt

# Configure
cp .env.example .env
# Edit .env with your API keys

# Ingest documents
python -m cli.app ingest ./docs/
python -m cli.app ingest report.pdf -c myproject

# Chat
python -m cli.app chat                         # Interactive REPL
python -m cli.app chat -q "What is X?"         # Single question
python -m cli.app chat -c myproject -p ollama  # Specific collection + provider

# Manage collections
python -m cli.app collections
python -m cli.app collections delete myproject

# System status
python -m cli.app status

# Start API server
python app.py

API Endpoints

Method	Endpoint	Description
POST	/api/chat	Send message (streaming SSE)
POST	/api/search	Vector search without LLM (returns raw results)
POST	/api/ingest	Upload + ingest files
GET	/api/collections	List collections
DELETE	/api/collections/{name}	Delete collection
GET	/api/collections/{name}/documents	List documents
DELETE	/api/collections/{name}/documents/{id}	Delete individual document
GET	/api/conversations	List conversations
GET	/api/conversations/{id}	Get conversation history
DELETE	/api/conversations/{id}	Delete conversation
GET	/api/health	Health check

Configuration

All settings are configured via .env file or environment variables. See .env.example for all options.

Key settings:

Variable	Default	Description
OPENAI_API_KEY	—	OpenAI API key
GEMINI_API_KEY	—	Google Gemini API key
DEFAULT_PROVIDER	openai	Default LLM provider
EMBEDDING_PROVIDER	openai	openai or local
CHUNK_SIZE	1000	Text chunk size
TOP_K	5	Number of chunks to retrieve
STREAM_TIMEOUT	120	SSE stream timeout in seconds
RATE_LIMIT_RPM	30	Max requests per minute per IP

Development

# Install dev dependencies
pip install -r requirements-dev.txt

# Run tests
pytest tests/ -v

# Lint
ruff check .

# Format
ruff format .

Tech Stack

• FastAPI — REST API
• Typer + Rich — CLI
• ChromaDB — Vector store
• aiosqlite — Async SQLite
• Pydantic Settings — Configuration
• Loguru — Logging

Input Validation

The API validates all inputs:

• Question: 1-5000 characters
• Collection names: alphanumeric, hyphens, underscores only (^[a-zA-Z0-9_-]+$)
• Provider: must be openai, gemini, or ollama
• top_k: 1-20
• File uploads: max 50MB, only .pdf, .docx, .txt, .md

Invalid requests return HTTP 422 with details.

License

MIT