Skip to content

pixano_inference.main

CLI entrypoint for starting the Pixano Inference server.

serve(host='127.0.0.1', port=7463, config=None, module_path=None)

Start the Pixano Inference server.

Examples:

Start the server

pixano-inference --host 0.0.0.0 --port 7463

Start with Python config

pixano-inference --host 0.0.0.0 --port 7463 --config models.py

Start with custom model modules

pixano-inference --module-path /path/to/my-models --config my_config.py

Source code in pixano_inference/main.py 
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
@app.command()
def serve(
    host: Annotated[str, typer.Option(help="Pixano Inference app URL host")] = "127.0.0.1",
    port: Annotated[int, typer.Option(help="Pixano Inference app URL port")] = 7463,
    config: Annotated[
        Optional[Path], typer.Option(exists=True, help="Path to Python config file (.py) for model deployments")
    ] = None,
    module_path: Annotated[
        Optional[list[Path]],
        typer.Option(help="Directory to add to Python path for custom model modules. Can be repeated."),
    ] = None,
):
    """Start the Pixano Inference server.

    Examples:
        # Start the server
        pixano-inference --host 0.0.0.0 --port 7463

        # Start with Python config
        pixano-inference --host 0.0.0.0 --port 7463 --config models.py

        # Start with custom model modules
        pixano-inference --module-path /path/to/my-models --config my_config.py
    """
    if module_path:
        for p in module_path:
            resolved = str(p.resolve())
            if resolved not in sys.path:
                sys.path.insert(0, resolved)

    from .ray import InferenceServer, RayServeConfig

    ray_config = RayServeConfig(host=host, port=port)
    server = InferenceServer(config=ray_config)

    if config is not None:
        server.register_from_config(config)

    server.start(host=host, port=port, blocking=True)