iota

iota - a minimal local embedding database.

Motivation

This project was done with the aim of reproducing some of my favourite features from existing vector stores while maintaining minimalism and simplicity.

Important

This is by no means scalable, but should suffice for smaller projects.

Installation

Install the package via PyPI:

pip install iotadb

Usage

Here is a very simple example:

from iotadb import IotaDB, Document

# Define a list of documents
docs = [
    Document(text="That is a happy dog"),
    Document(text="That is a very happy person"),
    Document(text="Today is a sunny day")
]

# Create a collection
db = IotaDB()
db.create_collection(name="my_collection", documents=docs)

# Query documents within your collection
results = db.search("That is a happy person", return_similarities=True)

for doc, score in results:
    print(f"Text: {doc.text}")
    print(f"similarity: {score:.3f}n")

More examples can be found in the /examples directory.

Features

• Simple interface: Easy-to-use API for database operations.
• Lightweight implementation: Minimal resource utilization.
• Local storage: Stores embeddings locally for fast and retrieval.
• Fast Indexing: Efficient embedding indexing for storage and retrieval.

Use cases

• Query with Natural Language: Search for relevant documents using simple natural language queries.
• Contextual Summarization: Integrate documents into LLM contexts like GPT-3 for data-augmented tasks.
• Similarity Search: Find similar items/documents based on their embeddings.

Contributing

Interested in contributing? Head over to the Contribution Guide for more details.