Datasette

To run the Datasette data exploration tool using Unit:

1. Install Unit with a Python 3.6+ language module.

2. Create a virtual environment to install Datasette’s PIP package, for instance:

   console

   Copy

   $ cd /path/to/app/  # Path to the application directory; use a real path in your configuration
   $ python --version  # Make sure your virtual environment version matches the module version
         Python X.Y.Z  # Major version, minor version, and revision number
   $ python -m venv venv  # Arbitrary name of the virtual environment
   $ source venv/bin/activate  # Name of the virtual environment from the previous command
   $ pip install datasette
   $ deactivate
   

   $ cd /path/to/app/  # Path to the application directory; use a real path in your configuration
   $ python --version  # Make sure your virtual environment version matches the module version
         Python X.Y.Z  # Major version, minor version, and revision number
   $ python -m venv venv  # Arbitrary name of the virtual environment
   $ source venv/bin/activate  # Name of the virtual environment from the previous command
   $ pip install datasette
   $ deactivate
   

3. Running Datasette on Unit requires a wrapper to expose the application object as the ASGI callable. Let’s use the following basic version, saving it as /path/to/app/asgi.py:

   python

   Copy

   import glob
   from datasette.app import Datasette
   
   application = Datasette(glob.glob('*.db')).app()

   import glob
   from datasette.app import Datasette
   
   application = Datasette(glob.glob('*.db')).app()

4. Change ownership:

   Run the following command (as root) so Unit can access the application directory (If the application uses several directories, run the command for each one):

   console

   Copy

   # chown -R unit:unit /path/to/app/  # User and group that Unit's router runs as by default
      

   # chown -R unit:unit /path/to/app/  # User and group that Unit's router runs as by default
      

   The unit:unit user-group pair is available only with official packages , Docker images, and some third-party repos. Otherwise, account names may differ; run the ps aux | grep unitd command to be sure.

   For further details, including permissions, see the security checklist.

5. Next, prepare the Datasette configuration for Unit (use real values for type, home, and path):

   json

   Copy

   {
      "listeners": {
         "*:80": {
            "pass": "applications/datasette"
         }
      },
   
      "applications": {
         "datasette": {
            "type": "python 3.Y",
            "_comment_type": "Must match language module version and virtual environment version",
            "path": "/path/to/app/",
            "_comment_path": "Path to the ASGI module",
            "home": "/path/to/app/venv/",
            "_comment_home": "Path to the virtual environment, if any",
            "module": "asgi",
            "_comment_module": "ASGI module filename with extension omitted",
            "callable": "app",
            "_comment_callable": "Name of the callable in the module to run"
         }
      }
   }

   {
      "listeners": {
         "*:80": {
            "pass": "applications/datasette"
         }
      },
   
      "applications": {
         "datasette": {
            "type": "python 3.Y",
            "_comment_type": "Must match language module version and virtual environment version",
            "path": "/path/to/app/",
            "_comment_path": "Path to the ASGI module",
            "home": "/path/to/app/venv/",
            "_comment_home": "Path to the virtual environment, if any",
            "module": "asgi",
            "_comment_module": "ASGI module filename with extension omitted",
            "callable": "app",
            "_comment_callable": "Name of the callable in the module to run"
         }
      }
   }

6. Upload the updated configuration.

   Assuming the JSON above was added to config.json. Run the following command as root:

   console

   Copy

   # curl -X PUT --data-binary @config.json --unix-socket \
         /path/to/control.unit.sock \  # Path to Unit's control socket in your installation
         http://localhost/config/      # Path to the config section in Unit's control API
      

   # curl -X PUT --data-binary @config.json --unix-socket \
         /path/to/control.unit.sock \  # Path to Unit's control socket in your installation
         http://localhost/config/      # Path to the config section in Unit's control API
      

   The control socket path may vary; run unitd -h or see Startup and shutdown for details.

   After a successful update, Datasette should be available on the listener’s IP address and port: