Chainlit Chatbot

A Chatbot which analyzes stock analysts' forecast on the stock market
The chatbot can store stock articles found online, and extract information such as market outlook from the articles.

Copy the code

Run the following command in your local folder after logging in with your Github account:

git clone https://github.com/madzai/chatbot_llm.git

Install required Python packages

Create and activate a virtual environment. Then run the following code in the parent folder:

pip install -r requirements.txt

Update the variables in .env file

Update variable KEY_TOGETHERAI (you need a Together AI account) for the Chatbot to run.
Update variable SERPER_API_KEY (you need a Serper account) for the Chatbot to be able to do web search for new stock articles.

Run the code

Run the chatbot with the following code in the parent folder:

chainlit run app.py

The chatbot will be available in http://localhost:8000/

Update the data

Search for newest stock articles in order to get the most updated market information.
Use function Update_all_data() and run the code in webSearch.py to search for new stock articles and store for use.

python webSearch.py

About the stock article data

The stock articles are stored in guru_urls.json.
File guru_source.json defines the data source and stock analysts to search for when updating the data.