A small, fast, and easy package for semantic searching using artificial intelligence.
Project description
DeepSearchKit
A small, fast, and easy package that allows easy semantic searching using artificial intelligence. It is based on and similar to the closed-source DeepSearch system.
Installation
You can install the latest stable version from the registry:
pip3 install deepsearchkit
You can install the very latest version directly from the Git repository, however certain features may not work:
pip3 install git+https://github.com/fakerybakery/DeepSearchKit
Features
- CPU, CUDA, and MPS support (enhanced GPU acceleration)!
- Simple usage
Usage
Documentation is available here.
Todo
- Integrate DeepSearch into DeepSearchKit
- Open-source DeepSearch
- Add Web Interface (from DeepSearch)
- Add document search demo
- Add document chat demo
- Add upsert feature (txtai#251)
- Add more data support, e.g. parquet
- Custom prompt/data format for multiple columns in JSON
- Custom progress callback for indexing
- Make some example projects
- Chat with a folder using open-sourced conversational models
- Search an entire directory
- Allow easy publishing with a
.dskpkg
file - compressed DeepSearchKit package that includes the index, the data, and some attributes (name, author, license, etc)- Maybe in the future: "DSK Hub" - hub for DSK packages
Credits
We would like to thank the authors of the following open source projects:
Disclaimer
By using/contributing to this software, you agree to the disclaimer.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
deepsearchkit-1.0.tar.gz
(3.7 kB
view hashes)
Built Distribution
Close
Hashes for deepsearchkit-1.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 83ff418826a3f3df0e854a43189774e9173bf791830bfb803af32d0626127a60 |
|
MD5 | 7abf455fafe5251a49e4703532642aeb |
|
BLAKE2b-256 | 2d9273c940a5dd8b0c462467c89fe8b7f7d93a7bef054fd1048f7ca3075e46d5 |