Simple AI assistant built with Python backend and Streamlit UI talking to Open AI model via Langchain.

Prerequisite

• Rename the .env.example to .env.

• Google Cloud Credentials:

  • Set up Google Cloud Console credentials.
  • Download OAuth credentials JSON file.
  • Update the /credentials.json file content with the downloaded file content.

• Asana Account:

  • Create Asana account.
  • Follow this Doc to create the API token and update the .env file.

• Open AI:

  • Create an Open AI account if you don't have one.
  • Get the API key and update the env file.

Local Development

This is built using Python virtual environment. Assuming python is installed, I have provided the requirement.txt with all the dependencies. You can either create a new virtual env or install the deps directly.

# install dependencies
pip install -r requirements.txt

# run locally
streamlit run main-langchian.py

---

NOTE: The first time you run the script, you'll need to authenticate to google. You need to provide the permissions to access the Google drive.

---

Workflows Tools

We have 3 Langchain tools created:

• create_asana_task - Tool for creating task in Asana. This uses Asana APIs.
• google_doc_tool - This tool is use to read content of the Google Document.
• google_drive_lister - This tool is use for listing the folders/files of a Google drive folder.

User Flow

User can perform certain task:

• Read content of Google Drive folder.
• Read the content of Google document.
• Can summarize or create task from the document in Asana.
• Or create new task in Asana.

Common MIME types for filtering:

• Google Docs: 'application/vnd.google-apps.document'
• Google Sheets: 'application/vnd.google-apps.spreadsheet'
• Google Slides: 'application/vnd.google-apps.presentation'
• PDFs: 'application/pdf'