This is an experiment to prompt an LLM to provide suggestions on how to improve decision matrix criteria.

pip|conda install openai google-api-python-client google-auth-httplib2 google-auth-oauthlib

Download (and store at project root) a credentials.json file from your Google Cloud Project which has Google Sheets API enabled.

Add your OpenAI API key to your environment.

export OPENAI_API_KEY="<your key>"

Modify aisheets.py to specify the sheet ID & your DM's range.

# Replace with your sheet's ID which can be found in its URL.
SPREADSHEET_ID = "..."
# Replace w/ a range that includes your DM's data, for example Sheet20!A1:Z
RANGE_NAME = "..."

Use a number 1-3 to indicate the step to execute.

1. Generate suggestions from the original prompt
2. Generate an optimized prompt, using the user's review of the suggestions
3. Generate suggestions from the optimized prompt

python3 aisheets.py <1|2|3>