Python

Use this reference for Functions, Jobs and Schedules written in Python. For a container with your own web server, see Services.

Project files

my-function/
├── handler.py
└── requirements.txt

Zip these at the root of the archive, not inside a folder. List your packages in requirements.txt, one per line:

requests

The handler

Define a function named handler in handler.py:

def handler(context, payload):
    # your code
    pass

• payload is the JSON input as a dict: the request body for a Function, the run's input for a Job.
• context is reserved for information from the runtime. Don't rely on its contents yet.

Function example

A Function can return anything. This shape is the convention:

def handler(context, payload):
    name = (payload or {}).get("name")
    if not name:
        return {
            "statusCode": 400,
            "body": {"error": "name is required"},
            "headers": {"content-type": "application/json"},
        }
    return {
        "statusCode": 200,
        "body": {"greeting": f"Hello, {name}"},
        "headers": {"content-type": "application/json"},
    }

Use status codes as in the Node.js reference.

Job and Schedule example

A Job or Schedule doesn't need to return anything:

def handler(context, payload):
    print("Import started")
    count = import_customers((payload or {}).get("source"))
    print(f"Imported {count} customers")

If an exception escapes the handler, the run is marked Failed. If you catch it, the run succeeds and handling it is up to you. See Jobs.

Settings, secrets and logs

• Read settings with os.environ.get("NAME"). See Configuration and secrets.
• Read secrets from the Secrets API at runtime, never from your ZIP; see Configuration and secrets.
• print output goes to the Process Log.

Deploy

gf deploy --name my-function --workload-type function   # or job, or schedule with --schedule

The CLI detects Python from requirements.txt or pyproject.toml. Or create the Coderunner in the Client Portal and upload the ZIP; see Functions.