Deep Search Toolkit

Interact with the Deep Search platform for new knowledge explorations and discoveries

The Deep Search Toolkit is a Python SDK and CLI allowing users to interact with the Deep Search platform. The Toolkit provides easy-to-use functionalities for several common processes such as document conversion, graph creation and querying.

Learn about IBM Deep Search

Quick links

• Documentation
• Deep Search Examples

Install

To set up, just install deepsearch-toolkit with your packaging tool.

With poetry:

poetry add deepsearch-toolkit

With pip:

pip install deepsearch-toolkit

Extras

Optional functionality can be installed as package "extras". To install all extras, use deepsearch-toolkit[all] with your packaging tool.

Install as toolkit developer

If you are a Deep Search Toolkit developer, set up as follows:

poetry install --all-extras

Requirements

Python 3.8+

Start using the toolkit

Set up a profile

For details, check Profiles.

deepsearch profile config

Convert a document

For details, check Document conversion.

deepsearch documents convert -p 1234567890abcdefghijklmnopqrstvwyz123456 -u https://arxiv.org/pdf/2206.00785.pdf

The output should look like:

Submitting input:     : 100%|██████████████████████████████| 1/1 [00:01<00:00,  1.52s/it]
Converting input:     : 100%|██████████████████████████████| 1/1 [00:33<00:00, 33.80s/it]
Downloading result:   : 100%|██████████████████████████████| 1/1 [00:01<00:00,  1.11s/it]
Total online documents             1
Successfully converted documents   1

Get help and support

Please feel free to connect with us using the discussion section.

Contributing

Please read Contributing to Deep Search Toolkit for details.

References

If you use Deep Search in your projects, please consider citing the following:

@software{Deep Search Toolkit,
author = {Deep Search Team},
month = {6},
title = {{Deep Search Toolkit}},
url = {https://github.com/DS4SD/deepsearch-toolkit},
version = {main},
year = {2022}
}

License

The Deep Search Toolkit codebase is under MIT license. For individual model usage, please refer to the model licenses found in the original packages.