vec-cache 0.1.1

No project description provided

Sure! Below is a base README.md for your Python package:

---

VecCache

VecCache is a Python package that provides an easy interface to cache embeddings of texts using OpenAI's API and allows for fast similarity searches over the cached embeddings.

Features

• Utilize OpenAI's API to generate text embeddings.
• Cache the embeddings efficiently using faiss.
• Quick similarity searches over the cached embeddings.

Installation

pip install vec_cache

Note: You need to have an OpenAI API key to use this package.

Usage

Basic Usage

from vec_cache import VecCache

# Initialize the cache with your OpenAI API key. If openai_api_key is not provided, .env or environment variable is loaded.
cache = VecCache(ttl=3600, openai_api_key="YOUR_OPENAI_API_KEY")

# Store a text in the cache
cache.store("Hello, World!")

# Search for the most similar text in the cache
result = cache.search("Hi, World!")
print(result)  # Outputs: "Hello, World!"

Storing with Precomputed Vectors

If you have precomputed vectors, you can store them directly:

vector = [0.1, 0.2, ...]  # Some precomputed vector
cache.store_with_vector("Some text", vector)

Searching with Distance

If you wish to obtain the distance along with the result:

result, distance = cache.search("Hi, World!", return_with_distance=True)
print(result)   # Outputs: "Hello, World!"
print(distance) # Outputs the L2 distance

Configuration

VecCache accepts the following parameters:

• ttl: Time to live for cached vectors.
• openai_api_key: Your OpenAI API key.
• embedding_model_name: Name of the OpenAI embedding model. Default is "text-embedding-ada-002".
• vector_size: Size of the vector. Default is 1536.

Contributions

We welcome contributions! Please submit a pull request or open an issue if you'd like to improve the package.

Using pre-commit for a Smoother Contribution Experience

To ensure consistent code quality and avoid common pitfalls, we use pre-commit hooks in our development process. Before you submit your pull request, it's recommended that you set up pre-commit to ensure your changes adhere to our coding standards and guidelines.

Steps to set up pre-commit:

1. Install pre-commit: If you haven't already, install packages including pre-commit:

   poetry install
   

2. Activate pre-commit for the repository: Navigate to the root directory of this repository and run:

   pre-commit install
   

License

This project is licensed under the MIT License.

---

Feel free to customize the content based on your package's details and specific requirements.

Of course! Here's a message you can include in your README.md file to encourage contributors to use pre-commit:

---

Contributing to [Your Project Name]

Using pre-commit for a Smoother Contribution Experience

To ensure consistent code quality and avoid common pitfalls, we use pre-commit hooks in our development process. Before you submit your pull request, it's recommended that you set up pre-commit to ensure your changes adhere to our coding standards and guidelines.

Steps to set up pre-commit:

1. Install pre-commit: If you haven't already, install pre-commit:

   pip install pre-commit
   

2. Activate pre-commit for the repository: Navigate to the root directory of this repository and run:

   pre-commit install
   

3. Run pre-commit on all files (optional): If you want to run the hooks on all the files in the repository (useful for the first-time setup):

   pre-commit run --all-files
   

With pre-commit set up, before every commit, the hooks will automatically check your changes and ensure they meet the repository's guidelines. If any issues are found, pre-commit will notify you, and in some cases, it might auto-fix the problems.

We appreciate your contribution and effort, and by setting up pre-commit, you're helping us maintain the quality and consistency of the project. Thank you!

---

Make sure to customize the placeholder [Your Project Name] with the actual name of your project. Adjust any other details as needed to better fit the specifics of your project or your desired messaging style.