Skip to content

UPDATE_TOKEN.md

Latest commit

 

History

History
111 lines (79 loc) · 3.59 KB

UPDATE_TOKEN.md

File metadata and controls

111 lines (79 loc) · 3.59 KB

Updating Your TabPFN Cloud Function Token

This document explains how to use the update_function.py script to update your deployed cloud function with a new TabPFN API token and switch from mock mode to real predictions.

Prerequisites

  • Python 3.6 or higher
  • tabpfn-client package installed (pip install tabpfn-client)
  • Google Cloud SDK installed and configured
  • A TabPFN account (sign up at tabpfn.github.io if needed)

Usage

The script provides several options for updating your function:

1. Basic Usage (Interactive)

Simply run the script with no arguments to be prompted for all necessary information:

python update_function.py

This will:

  • Detect your current GCP project
  • Prompt for your TabPFN email and password
  • Retrieve an API token
  • Save the token to ~/.tabpfn_token for future use
  • Update your cloud function to use the real TabPFN API

2. Provide TabPFN Credentials

python update_function.py --email your.email@example.com

You'll be prompted securely for your password.

3. Use an Existing Token

If you already have a TabPFN API token:

python update_function.py --token "your-api-token"

4. Specify GCP Project and Region

python update_function.py --project your-project-id --region us-central1

5. Keep Using Mock Mode (but with a Valid Token)

python update_function.py --use-mock true

Full Options Reference

usage: update_function.py [-h] [--project PROJECT] [--region REGION]
                         [--function FUNCTION] [--email EMAIL]
                         [--password PASSWORD] [--token TOKEN]
                         [--use-mock {true,false}] [--use-gcs {true,false}]
                         [--no-save-token]

options:
  -h, --help            show this help message and exit
  --project PROJECT     Google Cloud project ID
  --region REGION       GCP region where the function is deployed
  --function FUNCTION   Function name
  --email EMAIL         TabPFN account email
  --password PASSWORD   TabPFN account password (not recommended - use prompt)
  --token TOKEN         Use existing TabPFN token (skip login)
  --use-mock {true,false}
                        Whether to use mock predictions (default: false)
  --use-gcs {true,false}
                        Whether to use GCS for storage (default: false)
  --no-save-token       Don't save token to ~/.tabpfn_token

How It Works

  1. The script logs in to TabPFN using your credentials and obtains an API token
  2. It retrieves the current configuration of your deployed cloud function
  3. It updates the environment variables with the new token and settings
  4. It redeploys only the configuration (not the entire function)

Troubleshooting

If you encounter issues:

  1. Authentication errors: Make sure you're using the correct TabPFN credentials
  2. Permission errors: Ensure your Google Cloud account has permission to update the function
  3. Token not working: Try logging in to the TabPFN website to verify your account status
  4. Cloud Function not found: Check the function name and region with gcloud functions list

Security Notes

  • The script stores your TabPFN token in ~/.tabpfn_token with restricted permissions (readable only by you)
  • Never include your TabPFN password in command line arguments (use the prompt)
  • The token is securely transmitted to Google Cloud and stored as an environment variable

Testing After Update

After updating your function, you can test it with the test_deployment.py script:

python test_deployment.py --url "https://your-function-url" --verbose