lightning-sdk 0.1.49__py3-none-any.whl → 2025.11.5__py3-none-any.whl

lightning_sdk/cli/run.py

@@ -1,225 +0,0 @@
-from typing import TYPE_CHECKING, Dict, Optional
-
-from lightning_sdk.job import Job
-from lightning_sdk.machine import Machine
-from lightning_sdk.mmt import MMT
-from lightning_sdk.teamspace import Teamspace
-
-if TYPE_CHECKING:
-    from lightning_sdk.cli.legacy import _LegacyLightningCLI
-
-_MACHINE_VALUES = tuple([machine.value for machine in Machine])
-
-
-class _Run:
-    """Run async workloads on the Lightning AI platform."""
-
-    def __init__(self, legacy_run: Optional["_LegacyLightningCLI"] = None) -> None:
-        if legacy_run is not None:
-            self.app = legacy_run.app
-            self.model = legacy_run.model
-
-        # Need to set the docstring here for f-strings to work.
-        # Sadly this is the only way to really show options as f-strings are not allowed as docstrings directly
-        # and fire does not show values for literals, just that it is a literal.
-        docstr_job = f"""Run async workloads using a docker image or a compute environment from your studio.
-
-        Args:
-            name: The name of the job. Needs to be unique within the teamspace.
-            machine: The machine type to run the job on. One of {", ".join(_MACHINE_VALUES)}.
-            command: The command to run inside your job. Required if using a studio. Optional if using an image.
-                If not provided for images, will run the container entrypoint and default command.
-            studio: The studio env to run the job with. Mutually exclusive with image.
-            image: The docker image to run the job with. Mutually exclusive with studio.
-            teamspace: The teamspace the job should be associated with. Defaults to the current teamspace.
-            org: The organization owning the teamspace (if any). Defaults to the current organization.
-            user: The user owning the teamspace (if any). Defaults to the current user.
-            cloud_account: The cloud account to run the job on.
-                Defaults to the studio cloud account if running with studio compute env.
-                If not provided will fall back to the teamspaces default cloud account.
-            env: Environment variables to set inside the job.
-            interruptible: Whether the job should run on interruptible instances. They are cheaper but can be preempted.
-            image_credentials: The credentials used to pull the image. Required if the image is private.
-                This should be the name of the respective credentials secret created on the Lightning AI platform.
-            cloud_account_auth: Whether to authenticate with the cloud account to pull the image.
-                Required if the registry is part of a cloud provider (e.g. ECR).
-            artifacts_local: The path of inside the docker container, you want to persist images from.
-                CAUTION: When setting this to "/", it will effectively erase your container.
-                Only supported for jobs with a docker image compute environment.
-            artifacts_remote: The remote storage to persist your artifacts to.
-                Should be of format <CONNECTION_TYPE>:<CONNECTION_NAME>:<PATH_WITHIN_CONNECTION>.
-                PATH_WITHIN_CONNECTION hereby is a path relative to the connection's root.
-                E.g. efs:data:some-path would result in an EFS connection named `data` and to the path `some-path`
-                within it.
-                Note that the connection needs to be added to the teamspace already in order for it to be found.
-                Only supported for jobs with a docker image compute environment.
-            entrypoint: The entrypoint of your docker container. Defaults to `sh -c` which
-                just runs the provided command in a standard shell.
-                To use the pre-defined entrypoint of the provided image, set this to an empty string.
-                Only applicable when submitting docker jobs.
-        """
-        # TODO: the docstrings from artifacts_local and artifacts_remote don't show up completely,
-        # might need to switch to explicit cli definition
-        self.job.__func__.__doc__ = docstr_job
-
-        # Need to set the docstring here for f-strings to work.
-        # Sadly this is the only way to really show options as f-strings are not allowed as docstrings directly
-        # and fire does not show values for literals, just that it is a literal.
-        docstr_mmt = f"""Run async workloads on multiple machines using a docker image.
-
-        Args:
-            name: The name of the job. Needs to be unique within the teamspace.
-            num_machines: The number of Machines to run on. Defaults to 2 Machines
-            machine: The machine type to run the job on. One of {", ".join(_MACHINE_VALUES)}. Defaults to CPU
-            command: The command to run inside your job. Required if using a studio. Optional if using an image.
-                If not provided for images, will run the container entrypoint and default command.
-            studio: The studio env to run the job with. Mutually exclusive with image.
-            image: The docker image to run the job with. Mutually exclusive with studio.
-            teamspace: The teamspace the job should be associated with. Defaults to the current teamspace.
-            org: The organization owning the teamspace (if any). Defaults to the current organization.
-            user: The user owning the teamspace (if any). Defaults to the current user.
-            cloud_account: The cloud account to run the job on.
-                Defaults to the studio cloud account if running with studio compute env.
-                If not provided will fall back to the teamspaces default cloud account.
-            env: Environment variables to set inside the job.
-            interruptible: Whether the job should run on interruptible instances. They are cheaper but can be preempted.
-            image_credentials: The credentials used to pull the image. Required if the image is private.
-                This should be the name of the respective credentials secret created on the Lightning AI platform.
-            cloud_account_auth: Whether to authenticate with the cloud account to pull the image.
-                Required if the registry is part of a cloud provider (e.g. ECR).
-            artifacts_local: The path of inside the docker container, you want to persist images from.
-                CAUTION: When setting this to "/", it will effectively erase your container.
-                Only supported for jobs with a docker image compute environment.
-            artifacts_remote: The remote storage to persist your artifacts to.
-                Should be of format <CONNECTION_TYPE>:<CONNECTION_NAME>:<PATH_WITHIN_CONNECTION>.
-                PATH_WITHIN_CONNECTION hereby is a path relative to the connection's root.
-                E.g. efs:data:some-path would result in an EFS connection named `data` and to the path `some-path`
-                within it.
-                Note that the connection needs to be added to the teamspace already in order for it to be found.
-                Only supported for jobs with a docker image compute environment.
-            entrypoint: The entrypoint of your docker container. Defaults to `sh -c` which
-                just runs the provided command in a standard shell.
-                To use the pre-defined entrypoint of the provided image, set this to an empty string.
-                Only applicable when submitting docker jobs.
-        """
-        # TODO: the docstrings from artifacts_local and artifacts_remote don't show up completely,
-        # might need to switch to explicit cli definition
-        self.mmt.__func__.__doc__ = docstr_mmt
-
-    # TODO: sadly, fire displays both Optional[type] and Union[type, None] as Optional[Optional]
-    # see https://github.com/google/python-fire/pull/513
-    # might need to move to different cli library
-    def job(
-        self,
-        name: Optional[str] = None,
-        machine: Optional[str] = None,
-        command: Optional[str] = None,
-        studio: Optional[str] = None,
-        image: Optional[str] = None,
-        teamspace: Optional[str] = None,
-        org: Optional[str] = None,
-        user: Optional[str] = None,
-        cloud_account: Optional[str] = None,
-        env: Optional[Dict[str, str]] = None,
-        interruptible: bool = False,
-        image_credentials: Optional[str] = None,
-        cloud_account_auth: bool = False,
-        artifacts_local: Optional[str] = None,
-        artifacts_remote: Optional[str] = None,
-        entrypoint: str = "sh -c",
-    ) -> None:
-        if not name:
-            from datetime import datetime
-
-            timestr = datetime.now().strftime("%Y-%m-%d-%H-%M-%S")
-            name = f"job-{timestr}"
-
-        if machine is None:
-            # TODO: infer from studio
-            machine = "CPU"
-        machine_enum = Machine[machine.upper()]
-
-        resolved_teamspace = Teamspace(name=teamspace, org=org, user=user)
-
-        if cloud_account is None:
-            cloud_account = resolved_teamspace.default_cloud_account
-        machine_enum = Machine(machine.upper())
-
-        Job.run(
-            name=name,
-            machine=machine_enum,
-            command=command,
-            studio=studio,
-            image=image,
-            teamspace=resolved_teamspace,
-            org=org,
-            user=user,
-            cloud_account=cloud_account,
-            env=env,
-            interruptible=interruptible,
-            image_credentials=image_credentials,
-            cloud_account_auth=cloud_account_auth,
-            artifacts_local=artifacts_local,
-            artifacts_remote=artifacts_remote,
-            entrypoint=entrypoint,
-        )
-
-    # TODO: sadly, fire displays both Optional[type] and Union[type, None] as Optional[Optional]
-    # see https://github.com/google/python-fire/pull/513
-    # might need to move to different cli library
-    def mmt(
-        self,
-        name: Optional[str] = None,
-        num_machines: int = 2,
-        machine: Optional[str] = None,
-        command: Optional[str] = None,
-        image: Optional[str] = None,
-        teamspace: Optional[str] = None,
-        org: Optional[str] = None,
-        user: Optional[str] = None,
-        cloud_account: Optional[str] = None,
-        env: Optional[Dict[str, str]] = None,
-        interruptible: bool = False,
-        image_credentials: Optional[str] = None,
-        cloud_account_auth: bool = False,
-        artifacts_local: Optional[str] = None,
-        artifacts_remote: Optional[str] = None,
-        entrypoint: str = "sh -c",
-    ) -> None:
-        if name is None:
-            from datetime import datetime
-
-            timestr = datetime.now().strftime("%Y-%m-%d-%H-%M-%S")
-            name = f"mmt-{timestr}"
-
-        if machine is None:
-            # TODO: infer from studio
-            machine = "CPU"
-        machine_enum = Machine[machine.upper()]
-
-        resolved_teamspace = Teamspace(name=teamspace, org=org, user=user)
-        if cloud_account is None:
-            cloud_account = resolved_teamspace.default_cloud_account
-
-        if image is None:
-            raise RuntimeError("Image needs to be specified to run a multi-machine job")
-
-        MMT.run(
-            name=name,
-            num_machines=num_machines,
-            machine=machine_enum,
-            command=command,
-            studio=None,
-            image=image,
-            teamspace=resolved_teamspace,
-            org=org,
-            user=user,
-            cloud_account=cloud_account,
-            env=env,
-            interruptible=interruptible,
-            image_credentials=image_credentials,
-            cloud_account_auth=cloud_account_auth,
-            artifacts_local=artifacts_local,
-            artifacts_remote=artifacts_remote,
-            entrypoint=entrypoint,
-        )

lightning_sdk/cli/serve.py

@@ -1,218 +0,0 @@
-import os
-import subprocess
-import warnings
-from pathlib import Path
-from typing import Optional, Union
-
-from rich.console import Console
-from rich.progress import Progress, SpinnerColumn, TextColumn, TimeElapsedColumn
-from rich.prompt import Confirm
-
-
-class _LitServe:
-    """Serve a LitServe model.
-
-    Example:
-        lightning serve api server.py  # serve locally
-        lightning serve api server.py --cloud  # deploy to the cloud
-
-    You can deploy the API to the cloud by running `lightning serve api server.py --cloud`.
-    This will generate a Dockerfile, build the image, and push it to the image registry.
-    Deploying to the cloud requires pre-login to the docker registry.
-    """
-
-    def api(
-        self,
-        script_path: Union[str, Path],
-        easy: bool = False,
-        cloud: bool = False,
-        repository: Optional[str] = None,
-        non_interactive: bool = False,
-    ) -> None:
-        """Deploy a LitServe model script.
-
-        Args:
-            script_path: Path to the script to serve
-            easy: If True, generates a client for the model
-            cloud: If True, deploy the model to the Lightning Studio
-            repository: Optional Docker repository name (e.g., 'username/model-name')
-            non_interactive: If True, do not prompt for confirmation
-        Raises:
-            FileNotFoundError: If script_path doesn't exist
-            ImportError: If litserve is not installed
-            subprocess.CalledProcessError: If the script fails to run
-            IOError: If client.py generation fails
-        """
-        console = Console()
-        script_path = Path(script_path)
-        if not script_path.exists():
-            raise FileNotFoundError(f"Script not found: {script_path}")
-        if not script_path.is_file():
-            raise ValueError(f"Path is not a file: {script_path}")
-
-        try:
-            from litserve.python_client import client_template
-        except ImportError:
-            raise ImportError(
-                "litserve is not installed. Please install it with `pip install lightning_sdk[serve]`"
-            ) from None
-
-        if easy:
-            client_path = Path("client.py")
-            if client_path.exists():
-                console.print("Skipping client generation: client.py already exists", style="blue")
-            else:
-                try:
-                    client_path.write_text(client_template)
-                    console.print("✅ Client generated at client.py", style="bold green")
-                except OSError as e:
-                    raise OSError(f"Failed to generate client.py: {e!s}") from None
-
-        if cloud:
-            tag = repository if repository else "litserve-model"
-            return self._handle_cloud(script_path, console, tag=tag, non_interactive=non_interactive)
-
-        try:
-            subprocess.run(
-                ["python", str(script_path)],
-                check=True,
-                text=True,
-            )
-        except subprocess.CalledProcessError as e:
-            error_msg = f"Script execution failed with exit code {e.returncode}\nstdout: {e.stdout}\nstderr: {e.stderr}"
-            raise RuntimeError(error_msg) from None
-
-    def _handle_cloud(
-        self,
-        script_path: Union[str, Path],
-        console: Console,
-        tag: str = "litserve-model",
-        non_interactive: bool = False,
-    ) -> None:
-        try:
-            import docker
-        except ImportError:
-            raise ImportError("docker-py is not installed. Please install it with `pip install docker`") from None
-
-        try:
-            client = docker.from_env()
-            client.ping()
-        except docker.errors.DockerException as e:
-            raise RuntimeError(f"Failed to connect to Docker daemon: {e!s}. Is Docker running?") from None
-
-        dockerizer = _Docker()
-        path = dockerizer.api(script_path, port=8000, gpu=False, tag=tag)
-
-        console.clear()
-        if non_interactive:
-            console.print("[italic]non-interactive[/italic] mode enabled, skipping confirmation prompts", style="blue")
-
-        console.print(f"\nPlease review the Dockerfile at [u]{path}[/u] and make sure it is correct.", style="bold")
-        correct_dockerfile = True if non_interactive else Confirm.ask("Is the Dockerfile correct?", default=True)
-        if not correct_dockerfile:
-            console.print("Please fix the Dockerfile and try again.", style="red")
-            return
-
-        with Progress(
-            SpinnerColumn(),
-            TextColumn("[progress.description]{task.description}"),
-            TimeElapsedColumn(),
-            console=console,
-            transient=False,
-        ) as progress:
-            build_task = progress.add_task("Building Docker image", total=None)
-            build_status = client.api.build(
-                path=os.path.dirname(path), dockerfile=path, tag=tag, decode=True, quiet=False
-            )
-            for line in build_status:
-                if "error" in line:
-                    progress.stop()
-                    console.print(f"\n[red]{line}[/red]")
-                    return
-                if "stream" in line and line["stream"].strip():
-                    console.print(line["stream"].strip(), style="bright_black")
-                    progress.update(build_task, description="Building Docker image")
-
-            progress.update(build_task, description="[green]Build completed![/green]")
-
-            push_task = progress.add_task("Pushing to registry", total=None)
-            console.print("\nPushing image...", style="bold blue")
-            push_status = client.api.push(tag, stream=True, decode=True)
-            for line in push_status:
-                if "error" in line:
-                    progress.stop()
-                    console.print(f"\n[red]{line}[/red]")
-                    return
-                if "status" in line:
-                    console.print(line["status"], style="bright_black")
-                    progress.update(push_task, description="Pushing to registry")
-
-            progress.update(push_task, description="[green]Push completed![/green]")
-
-        console.print(f"\n✅ Image pushed to {tag}", style="bold green")
-        console.print(
-            "Soon you will be able to deploy this model to the Lightning Studio!",
-        )
-        # TODO: Deploy to the cloud
-
-
-class _Docker:
-    """Generate a Dockerfile for a LitServe model."""
-
-    def api(self, server_filename: str, port: int = 8000, gpu: bool = False, tag: str = "litserve-model") -> str:
-        """Generate a Dockerfile for the given server code.
-
-        Args:
-            server_filename: The path to the server file. Example sever.py or app.py.
-            port: The port to expose in the Docker container.
-            gpu: Whether to use a GPU-enabled Docker image.
-            tag: Docker image tag to use in examples.
-        """
-        import litserve as ls
-        from litserve import docker_builder
-
-        console = Console()
-        requirements = ""
-        if os.path.exists("requirements.txt"):
-            requirements = "-r requirements.txt"
-        else:
-            warnings.warn(
-                f"requirements.txt not found at {os.getcwd()}. "
-                f"Make sure to install the required packages in the Dockerfile.",
-                UserWarning,
-            )
-
-        current_dir = Path.cwd()
-        if not (current_dir / server_filename).is_file():
-            raise FileNotFoundError(f"Server file `{server_filename}` must be in the current directory: {os.getcwd()}")
-
-        version = ls.__version__
-        if gpu:
-            run_cmd = f"docker run --gpus all -p {port}:{port} {tag}:latest"
-            docker_template = docker_builder.CUDA_DOCKER_TEMPLATE
-        else:
-            run_cmd = f"docker run -p {port}:{port} {tag}:latest"
-            docker_template = docker_builder.DOCKERFILE_TEMPLATE
-        dockerfile_content = docker_template.format(
-            server_filename=server_filename,
-            port=port,
-            version=version,
-            requirements=requirements,
-        )
-        with open("Dockerfile", "w") as f:
-            f.write(dockerfile_content)
-
-        success_msg = f"""[bold]Dockerfile created successfully[/bold]
-Update [underline]{os.path.abspath("Dockerfile")}[/underline] to add any additional dependencies or commands.
-
-[bold]Build the container with:[/bold]
-> [underline]docker build -t {tag} .[/underline]
-
-[bold]To run the Docker container on the machine:[/bold]
-> [underline]{run_cmd}[/underline]
-
-[bold]To push the container to a registry:[/bold]
-> [underline]docker push {tag}[/underline]
-"""
-        console.print(success_msg)
-        return os.path.abspath("Dockerfile")

lightning_sdk/cli/stop.py

@@ -1,37 +0,0 @@
-from typing import Optional
-
-from lightning_sdk.cli.job_and_mmt_action import _JobAndMMTAction
-
-
-class _Stop(_JobAndMMTAction):
-    """Stop resources on the Lightning AI platform."""
-
-    def job(self, name: Optional[str] = None, teamspace: Optional[str] = None) -> None:
-        """Stop a job.
-
-        Args:
-            name: the name of the job. If not specified can be selected interactively.
-            teamspace: the name of the teamspace the job lives in.
-                Should be specified as {teamspace_owner}/{teamspace_name} (e.g my-org/my-teamspace).
-                If not specified can be selected interactively.
-
-        """
-        job = super().job(name=name, teamspace=teamspace)
-
-        job.stop()
-        print(f"Successfully stopped {job.name}!")
-
-    def mmt(self, name: Optional[str] = None, teamspace: Optional[str] = None) -> None:
-        """Stop a multi-machine job.
-
-        Args:
-            name: the name of the job. If not specified can be selected interactively.
-            teamspace: the name of the teamspace the job lives in.
-                Should be specified as {teamspace_owner}/{teamspace_name} (e.g my-org/my-teamspace).
-                If not specified can be selected interactively.
-
-        """
-        mmt = super().mmt(name=name, teamspace=teamspace)
-
-        mmt.stop()
-        print(f"Successfully stopped {mmt.name}!")