Acryl Executor

0.3.10 · active · verified Fri Apr 17

Acryl Executor is a Python library used internally by Acryl Agents and the DataHub ecosystem to define and execute modular tasks. It provides core abstractions like `BaseTask` and `Executor` for building extensible task-based workflows. As a sub-package of the larger DataHub project, its releases are often coupled with DataHub's evolution, though its own versioning is independent. The current version is 0.3.10, and it follows an active release cadence driven by DataHub development.

Common errors

```
TypeError: Can't instantiate abstract class MySimpleTask with abstract method execute
```
cause A class inheriting from `acryl.executor.sdk.tasks.base_task.BaseTask` must implement the `execute` method as defined in the abstract base class.

fix Ensure your custom task class includes a definition for `def execute(self, config: Dict[str, Any]) -> TaskResult: ...` with the correct signature and return type.
```
ModuleNotFoundError: No module named 'acryl.executor.sdk'
```
cause The `acryl-executor` library is not installed, or the import path is incorrect (e.g., trying `acryl_executor` instead of `acryl`).

fix Run `pip install acryl-executor`. Verify all imports use `from acryl.executor...`.
```
TypeError: execute_task() missing 1 required positional argument: 'config'
```
cause The `execute_task` method on `Executor` expects two arguments: a `BaseTask` instance and a `config` dictionary.

fix Ensure you call `executor.execute_task(my_task_instance, {'key': 'value'})` providing a dictionary for task configuration, even if it's an empty dictionary.

Warnings

gotcha Acryl Executor is primarily an internal library for the DataHub ecosystem. While available on PyPI, its public API (especially pre-1.0 versions like 0.x.x) may evolve rapidly with breaking changes that are not always explicitly documented for `acryl-executor` itself, but rather as part of broader DataHub releases.
Fix: Always pin to exact versions (e.g., `acryl-executor==0.3.10`) and review DataHub's main release notes for related changes when upgrading. Be prepared to adapt custom task implementations.
breaking Given `acryl-executor` is in its 0.x.x series, changes to `BaseTask` method signatures or `TaskResult` structure might occur in minor versions (e.g., `0.3.10` to `0.4.0`) without adhering to strict semantic versioning. This means a minor version bump could introduce breaking changes.
Fix: Thoroughly test custom tasks against new `acryl-executor` versions before deploying. Consult the DataHub GitHub repository commit history for `acryl-executor` changes if experiencing unexpected behavior after an upgrade.
gotcha The `Executor` provides synchronous task execution by default. For long-running or asynchronous tasks, `acryl-executor` is typically integrated with an external orchestration layer (e.g., DataHub Actions framework), which handles queuing, scheduling, and concurrency. Running `executor.execute_task()` directly will block the calling thread.
Fix: For production scenarios requiring async or background execution, integrate `acryl-executor` within a suitable task runner like `datahub-actions` or a custom worker pool. Do not assume `Executor` handles concurrency internally.

Install

pip install acryl-executor Install latest version

Imports

BaseTask
wrong:
```
from acryl_executor.sdk.tasks.base_task import BaseTask
```
correct:
```
from acryl.executor.sdk.tasks.base_task import BaseTask
```
The top-level package is `acryl`, not `acryl_executor`, despite the PyPI package name containing a hyphen.

TaskResult

from acryl.executor.sdk.tasks.base_task import TaskResult

Executor
wrong:
```
from acryl_executor.sdk.executor import Executor
```
correct:
```
from acryl.executor.sdk.executor import Executor
```
The top-level package is `acryl`, not `acryl_executor`, despite the PyPI package name containing a hyphen.

Quickstart

This quickstart demonstrates how to define a custom task by inheriting from `BaseTask`, implement its `execute` method, and then run it using the `Executor`. It shows how to pass configuration to the task and retrieve results. The example also illustrates how to safely access environment variables for sensitive data.

import os
from typing import Dict, Any, NamedTuple
from abc import ABC, abstractmethod

from acryl.executor.sdk.tasks.base_task import BaseTask, TaskResult
from acryl.executor.sdk.executor import Executor


class MySimpleTaskResult(NamedTuple):
    status: str
    message: str


class MySimpleTask(BaseTask):
    def execute(self, config: Dict[str, Any]) -> MySimpleTaskResult:
        name = config.get('name', 'World')
        # Example of using an environment variable for configuration (e.g., for API keys)
        secret_key = os.environ.get('MY_SECRET_KEY', 'default_secret')
        print(f"Executing MySimpleTask for {name} with secret: {secret_key[:4]}...")
        return MySimpleTaskResult(status='SUCCESS', message=f'Hello, {name}!')

# Configure with actual environment variable or a placeholder for quick run
os.environ['MY_SECRET_KEY'] = os.environ.get('MY_SECRET_KEY', 'some_default_value_for_testing')

# 1. Instantiate the Executor
executor = Executor()

# 2. Define a task instance
my_task = MySimpleTask()

# 3. Define task configuration
task_config = {"name": "DataHub User"}

# 4. Execute the task
print("\n--- Executing MySimpleTask ---")
result = executor.execute_task(my_task, task_config)

print(f"Task Result: {result.status}, Message: {result.message}")

# Example with different config
task_config_2 = {"name": "Acryl Team"}
print("\n--- Executing MySimpleTask with different config ---")
result_2 = executor.execute_task(my_task, task_config_2)
print(f"Task Result: {result_2.status}, Message: {result_2.message}")

view raw JSON →