示例#1
0
def run_merge_job(
    vector,
    target,
    entity_rows=None,
    timestamp_column=None,
    run_config=None,
    drop_columns=None,
):
    name = vector.metadata.name
    if not name:
        raise mlrun.errors.MLRunInvalidArgumentError(
            "feature vector name must be specified")
    if not target or not hasattr(target, "to_dict"):
        raise mlrun.errors.MLRunInvalidArgumentError(
            "target object must be specified")
    name = f"{name}_merger"
    run_config = run_config or RunConfig()
    if not run_config.function:
        function_ref = vector.spec.function.copy()
        if function_ref.is_empty():
            function_ref = FunctionReference(name=name, kind="job")
        if not function_ref.url:
            function_ref.code = _default_merger_handler
        run_config.function = function_ref

    function = run_config.to_function(
        "job", mlrun.mlconf.feature_store.default_job_image)
    function.metadata.project = vector.metadata.project
    function.metadata.name = function.metadata.name or name
    task = new_task(
        name=name,
        params={
            "vector_uri": vector.uri,
            "target": target.to_dict(),
            "timestamp_column": timestamp_column,
            "drop_columns": drop_columns,
        },
        inputs={"entity_rows": entity_rows},
    )
    task.spec.secret_sources = run_config.secret_sources
    task.set_label("job-type",
                   "feature-merge").set_label("feature-vector", vector.uri)
    task.metadata.uid = uuid.uuid4().hex
    vector.status.run_uri = task.metadata.uid
    vector.save()

    run = function.run(
        task,
        handler=run_config.handler or "merge_handler",
        local=run_config.local,
        watch=run_config.watch,
    )
    logger.info(f"feature vector merge job started, run id = {run.uid()}")
    return RemoteVectorResponse(vector, run)
示例#2
0
def run_merge_job(
    vector,
    target,
    entity_rows=None,
    timestamp_column=None,
    local=None,
    watch=None,
    drop_columns=None,
    function=None,
    secrets=None,
    auto_mount=None,
):
    name = vector.metadata.name
    if not name:
        raise mlrun.errors.MLRunInvalidArgumentError(
            "feature vector name must be specified")
    if not target or not hasattr(target, "to_dict"):
        raise mlrun.errors.MLRunInvalidArgumentError(
            "target object must be specified")
    name = f"{name}_merger"
    if not function:
        function_ref = vector.spec.function
        if not function_ref.to_dict():
            function_ref = FunctionReference(name=name, kind="job")
        function_ref.image = (function_ref.image
                              or mlrun.mlconf.feature_store.default_job_image)
        if not function_ref.url:
            function_ref.code = _default_merger_handler
        function = function_ref.to_function()

    if auto_mount:
        function.apply(mlrun.platforms.auto_mount())

    function.metadata.project = vector.metadata.project
    task = new_task(
        name=name,
        params={
            "vector_uri": vector.uri,
            "target": target.to_dict(),
            "timestamp_column": timestamp_column,
            "drop_columns": drop_columns,
        },
        inputs={"entity_rows": entity_rows},
    )
    if secrets:
        task.with_secrets("inline", secrets)
    task.metadata.uid = uuid.uuid4().hex
    vector.status.run_uri = task.metadata.uid
    vector.save()

    run = function.run(task, handler="merge_handler", local=local, watch=watch)
    logger.info(f"feature vector merge job started, run id = {run.uid()}")
    return RemoteVectorResponse(vector, run)
示例#3
0
文件: base.py 项目: AlonMaor14/mlrun
 def _generate_task(self):
     return new_task(name=self.name,
                     project=self.project,
                     artifact_path=self.artifact_path)