*** This repo has been retired, please see https://github.com/csymonds/chatbot ***

This is a chatGPT bot with long term memory using Pinecone

It also uses the Whisper API to do speech to text

Setup

1. If you don’t have Python installed, install it from here

   You will need to create an account on Pinecone and follow their registration procedures.

   Pinecone also has documentation you can follow for integration with OpenAI if you want to try your own.

   Documentation for the OpenAI chatGPT API is available here.

2. Clone this repository

   $ git clone [email protected]:csymonds/codebot.git

3. Navigate into the project directory

   $ cd codebot

Virtual environment and dependency installation

4. Virtualize (Note: Windows users will see venv/Scripts/activate)

   $ python -m venv venv
   $ . venv/bin/activate
   

5. Install the library dependencies

   $ pip3 install -r requirements.txt

   In addition to the library dependencies, there are some additional runtime dependencies like pytorch

   The below install version assumes you have CUDA 11.7 installed.

   The necessary CUDA installers are found here.

   Additional PyTorch installers are found here.

   $ pip3 install keyboard torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu117

6. Add keys

   Step 1: Make a copy of the example environment variables files

   $ cp example_key_openai.txt key_openai.txt
   $ cp example_key_pinecone.txt key_pinecone.txt

   Step 2: Copy in your key to the respective file

   Add your OpenAI API key to the newly created key_openai.txt file

   Add your Pinecone API key to the newly created key_pinecone.txt file

7. Deactivate

   $ deactivate
   

Tips: sounddevice defaults to your default communications devices in order to record/playback

To check/modify what those defaults are, use

python -m sounddevice

PLEASE NOTE: You will get a warning about model parameters missing when you run this. It is harmless, and can be suppressed by setting the ignore_warnings parameter to True