sg-python-api

sg-python-api is a Python client for Autodesk Flow Production Tracking, written for pipeline tools and scripts built on Autodesk's shotgun_api3 client. It keeps the same import name and documented API, so an existing integration can switch without rewriting its calls and gain shared connections, bounded network waits, and retries that do not repeat an ambiguous write. It also adds a native asynchronous client.

This documentation uses SG for Flow Production Tracking, formerly named Shotgun and ShotGrid.

Why use it?

• Reuse connections: share one synchronous client across threads, or one async client across tasks on the same event loop.
• Bound network waits: default connect and read timeouts help applications recover from an unavailable service.
• Handle writes carefully: retries account for whether an operation is safe to repeat. An ambiguous write is left for the application to reconcile.
• Test without a site: synchronous and asynchronous mockgun clients provide an in-memory API for application tests.

sg-python-api is an independent implementation of the documented API of Autodesk's Python client 3.10.3. Some behavior differs on purpose, so validate your application's workflows against a test site before replacing its production client. The migration guide walks through the switch, and the compatibility reference lists the differences.

Installation

You need Python 3.11 or newer and an SG site running server version 7.10.0 or newer. Authenticated requests to a remote site require HTTPS.

sg-python-api and Autodesk's shotgun-api3 both install the shotgun_api3 package, so install exactly one of them in a Python environment. If the official client is already installed, follow the migration steps to replace it.

From a checkout, install into the Python environment that runs your application:

python -m pip install .

The base installation provides the synchronous Shotgun client and both mockgun test clients. To use the native asynchronous client, AsyncShotgun, install the async extra, which adds the httpx2 HTTP library:

python -m pip install '.[async]'

Quick start

These examples read one project from your site. They take the site URL and script credentials from environment variables so the API key stays out of source code; the client does not read these variables itself. Set SHOTGUN_URL to your site's HTTPS URL, and set SHOTGUN_SCRIPT_NAME and SHOTGUN_API_KEY to a script account that can read projects.

Synchronous client

import os

from shotgun_api3 import Shotgun

sg = Shotgun(
    os.environ['SHOTGUN_URL'],
    script_name=os.environ['SHOTGUN_SCRIPT_NAME'],
    api_key=os.environ['SHOTGUN_API_KEY'],
)
try:
    project = sg.find_one('Project', [], ['name'])
    print(project)
finally:
    sg.close()

The script prints a project dictionary containing type, id, and name, or None when no project is visible to that account. You can share one Shotgun client across threads; close() releases its pooled connections when you no longer need it.

Asynchronous client

With the async extra installed and the same environment variables set, run:

import asyncio
import os

from shotgun_api3 import AsyncShotgun


async def main() -> None:
    async with AsyncShotgun(
        os.environ['SHOTGUN_URL'],
        script_name=os.environ['SHOTGUN_SCRIPT_NAME'],
        api_key=os.environ['SHOTGUN_API_KEY'],
    ) as sg:
        project = await sg.find_one('Project', [], ['name'])
        print(project)


asyncio.run(main())

The output matches the synchronous example. The async with block closes the client even when an operation fails. Without the async extra, importing AsyncShotgun raises ImportError.

Share an AsyncShotgun only among tasks on one event loop, and cancel its work by cancelling the task. A cancelled write may already have reached the server, so check its outcome before repeating it. The async reference covers shutdown and cancellation in detail.

Documentation

Development

See the development guide for the required tools, CI, and dependency locks. With those tools installed, run:

make setup
make check

make check also runs the live-site tests, which create and delete records, whenever the SG_TEST_* variables described in the development guide are set.

License

sg-python-api is available under the MIT License. It is an independent project and is not affiliated with or endorsed by Autodesk. Flow Production Tracking is a product of Autodesk, Inc.