Haiku SQLite RAG

A SQLite-based Retrieval-Augmented Generation (RAG) system built for efficient document storage, chunking, and hybrid search capabilities.

Features

Document Management: Store and manage documents with automatic content parsing
Smart Updates: Intelligent file/URL monitoring with MD5-based change detection
Hybrid Search: Full-text search (FTS5) combined with vector embeddings
Multi-format Support: Parse 40+ file formats including PDF, DOCX, HTML, Markdown, and more
Web Content: Direct URL ingestion with automatic content type detection
Vector Embeddings: Uses sqlite-vec for efficient similarity search
Automatic Chunking: Intelligent document segmentation for better retrieval

Installation

uv pip install haiku.rag

or for development, checkout the repository and then,

# Install dependencies
uv sync

# Activate virtual environment
source .venv/bin/activate

Quick Start

from pathlib import Path
from haiku.rag.client import HaikuRAG

# Initialize client with database path
client = HaikuRAG("path/to/database.db")
# Or use in-memory database for testing
client = HaikuRAG(":memory:")

# Create document from text
doc = await client.create_document(
    content="Your document content here",
    uri="doc://example",
    metadata={"source": "manual", "topic": "example"}
)

# Create document from file (auto-parses content)
doc = await client.create_document_from_source("path/to/document.pdf")

# Create document from URL
doc = await client.create_document_from_source("https://example.com/article.html")

# Retrieve documents
doc = await client.get_document_by_id(1)
doc = await client.get_document_by_uri("file:///path/to/document.pdf")

# List all documents with pagination
docs = await client.list_documents(limit=10, offset=0)

# Update document content
doc.content = "Updated content"
await client.update_document(doc)

# Delete document
await client.delete_document(doc.id)

# Clean up
client.close()

Smart Document Updates

The system automatically tracks file changes using MD5 hashes:

# First call - creates new document
doc1 = await client.create_document_from_source("document.txt")

# Second call - no changes, returns existing document (no processing)
doc2 = await client.create_document_from_source("document.txt")
assert doc1.id == doc2.id

# After file modification - automatically updates existing document
# File content changed...
doc3 = await client.create_document_from_source("document.txt")
assert doc1.id == doc3.id  # Same document
assert doc3.content != doc1.content  # Updated content

Supported File Formats

The system supports 40+ file formats through MarkItDown:

Documents: PDF, DOCX, PPTX, XLSX
Web: HTML, XML
Text: TXT, MD, CSV, JSON, YAML
Code: PY, JS, TS, C, CPP, JAVA, GO, RS, and more
Media: MP3, WAV (transcription)

Document Metadata

Documents automatically include metadata:

doc = await client.create_document_from_source("example.pdf")
print(doc.metadata)
# {
#   "contentType": "application/pdf",
#   "md5": "abc123...",
#   "custom_field": "value"  # Your custom metadata
# }

Contributing

Fork the repository
Create a feature branch
Add tests for new functionality
Ensure all tests pass: pytest
Run type checking: pyright
Run linting: ruff check
Submit a pull request

3.4 KiB Raw Blame History