jason810496 commented on code in PR #61153: URL: https://github.com/apache/airflow/pull/61153#discussion_r2839544644
########## airflow-core/src/airflow/executors/workloads/callback.py: ########## @@ -0,0 +1,158 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. +"""Callback workload schemas for executor communication.""" + +from __future__ import annotations + +from enum import Enum +from importlib import import_module +from pathlib import Path +from typing import TYPE_CHECKING, Literal +from uuid import UUID + +import structlog +from pydantic import BaseModel, Field, field_validator + +from airflow.executors.workloads.base import BaseDagBundleWorkload, BundleInfo + +if TYPE_CHECKING: + from airflow.api_fastapi.auth.tokens import JWTGenerator + from airflow.models import DagRun + from airflow.models.callback import Callback as CallbackModel, CallbackKey + +log = structlog.get_logger(__name__) + + +class CallbackFetchMethod(str, Enum): + """Methods used to fetch callback at runtime.""" + + # For future use once Dag Processor callbacks (on_success_callback/on_failure_callback) get moved to executors + DAG_ATTRIBUTE = "dag_attribute" + + # For deadline callbacks since they import callbacks through the import path + IMPORT_PATH = "import_path" + + +class CallbackDTO(BaseModel): Review Comment: Naming is hard, but we haven’t introduced the term “DTO” in the core codebase so far. Would it be better to name it “DataModel” instead? This also align with the `datamodels` module naming in API. ########## airflow-core/src/airflow/executors/workloads/callback.py: ########## @@ -0,0 +1,158 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. +"""Callback workload schemas for executor communication.""" + +from __future__ import annotations + +from enum import Enum +from importlib import import_module +from pathlib import Path +from typing import TYPE_CHECKING, Literal +from uuid import UUID + +import structlog +from pydantic import BaseModel, Field, field_validator + +from airflow.executors.workloads.base import BaseDagBundleWorkload, BundleInfo + +if TYPE_CHECKING: + from airflow.api_fastapi.auth.tokens import JWTGenerator + from airflow.models import DagRun + from airflow.models.callback import Callback as CallbackModel, CallbackKey + +log = structlog.get_logger(__name__) + + +class CallbackFetchMethod(str, Enum): + """Methods used to fetch callback at runtime.""" + + # For future use once Dag Processor callbacks (on_success_callback/on_failure_callback) get moved to executors + DAG_ATTRIBUTE = "dag_attribute" + + # For deadline callbacks since they import callbacks through the import path + IMPORT_PATH = "import_path" + + +class CallbackDTO(BaseModel): + """Schema for Callback with minimal required fields needed for Executors and Task SDK.""" + + id: str # A uuid.UUID stored as a string + fetch_method: CallbackFetchMethod + data: dict + + @field_validator("id", mode="before") + @classmethod + def validate_id(cls, v): + """Convert UUID to str if needed.""" + if isinstance(v, UUID): + return str(v) + return v + + @property + def key(self) -> CallbackKey: + """Return callback ID as key (CallbackKey = str).""" + return self.id + + +class ExecuteCallback(BaseDagBundleWorkload): + """Execute the given Callback.""" + + callback: CallbackDTO + + type: Literal["ExecuteCallback"] = Field(init=False, default="ExecuteCallback") + + @classmethod + def make( + cls, + callback: CallbackModel, + dag_run: DagRun, + dag_rel_path: Path | None = None, + generator: JWTGenerator | None = None, + bundle_info: BundleInfo | None = None, + ) -> ExecuteCallback: + """Create an ExecuteCallback workload from a Callback ORM model.""" + if not bundle_info: + bundle_info = BundleInfo( + name=dag_run.dag_model.bundle_name, + version=dag_run.bundle_version, + ) + fname = f"executor_callbacks/{callback.id}" # TODO: better log file template + + return cls( + callback=CallbackDTO.model_validate(callback, from_attributes=True), + dag_rel_path=dag_rel_path or Path(dag_run.dag_model.relative_fileloc or ""), + identity_token=cls.generate_token(str(callback.id), generator), + log_path=fname, + bundle_info=bundle_info, + ) + + +def execute_callback_workload( Review Comment: May I ask whether the callback will eventually be executed within TaskSDK instead of within the Scheduler/Executor process in the long term? Executing the callback in the Scheduler/Executor process does not seem very secure, even though it does not violate the AF3 security model, since the Scheduler/Executor process can directly access the database. From my perspective, all synchronous workloads will be executed within TaskSDK. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
