batmanscode/debatable

LLM to address sales objections. Can be used as a template for input output APIs with LLMs.

debatable

LLM to address sales objections

---

debatable.py - python funcs to generate suggestions

api.py - REST API wrapper around debatable.py

app.py - streamlit app using debatable.py

---

Note: the following will have to be saved as an environment variables:

• OPENAI_API_KEY
• DETA_PROJECT_KEY: Key for deta.space's "Base" NoSQL DB (to collect stats)

Run a FastAPI server with

uvicorn api:app --reload # reload flag for development

Run streamlit app with

streamlit run app.py

---

Using Codespaces/VSCode

Everything you need will be installed when you open Codespaces/VSCode; specified in .devcontainer/

Notes for Codespaces:

• Aider coding assistant will automatically be installed if you want to use it
  • https://github.com/paul-gauthier/aider
  • To use it, in the terminal use aider
  • Personally I like to use it with aider --no-git
• Jupyter (will be added to the devcontainer later, not used now)
  • Currently Jupyter notebook doesn’t work on Codespaces for an unknown reason, or at least I can’t, so you’ll have to use JupyterLab
  • Open with jupyter lab --NotebookApp.allow_origin='*' --NotebookApp.ip='0.0.0.0'
  • For more info on using see https://code.visualstudio.com/docs/datascience/notebooks-web
• Streamlit
  • ./.devcontainer/streamlit.sh is the equivalent to running streamlit with streamlit run app.py --server.enableCORS false --server.enableXsrfProtection false
  • Open with streamlit run app.py --server.enableCORS false --server.enableXsrfProtection false
  • For more info see https://discuss.streamlit.io/t/how-to-make-streamlit-run-on-codespaces/24526/6
• FastAPI
  • --reload doesn't work in codespaces, will have to refresh to see changes