BDF-Public

Public repository for the BDF Chatbot developed by ICF

Requirements

Python version >= 3.10

Before using the app, need to obtain various authentications and set them in .env file such as the following

• ChainLit secret string that is used to sign the authentication tokens
• AWS access and secret keys
• AWS Knowledge Base ID
• Name of the table where conversations history will be saved if using authentication

Installation

1. Clone Bioinsight app code
2. cd BDF-Chainlit-Chatbot
3. pip install -r requirements.txt
4. Install bdi-kit library. Bioinsight app is integrated with bdi-kit for data harmonization. For further instructions on installation and usage please refer here: https://github.com/VIDA-NYU/bdi-kit
5. Due to GitHub repository file size limitations we added only abbriviated version of MIDRC distributed subjects data file (first 1500 rows). To obtain the full version for proper MIDRC integration, please refer to the following repository generated by our collaborators https://github.com/BioIntelligence-Lab/VoxelInsight specifically, you will need the following file: /Data/midrc_distributed_subjects.csv copied to the following location /BDF-Public/bioinsight_ai/data_sources/cancer_research_data_commons/imaging_data_commons

If running the app in local environment we suggest to create a virtual environment and run the app from there:

Running

chainlit run .\chatbot-server.py