Create LLM Functions

[DreamsDragon]

The LLM code should take

1. Meta data about the data we have like column names and file names
2. Decide if it is an update on existing graph or new graph
3. Give existing code if any to LLM along with message history to generate new code
4. If any errors give error giving code with the error back to LLM to fix it
5. The final product should be a python function that we can pass the data into as dataframes and output is a plotly plot

[DreamsDragon]

You can create a folder called notebooks in root directory and add your notebooks to that. Please dont push any keys even by mistake. You can use environment variables in notebooks to avoid this