fix: aborting of last Actor/task run

[vdusek]

Problem description

Aborting the last Actor run does not work

• Resulting in:

apify_client._errors.ApifyApiError: We have bad news: there is no API endpoint at this URL. 
Did you specify it correctly?

• Code to reproduce it:

# sync version
from apify_client import ApifyClient

TOKEN = '...'
ACTOR_ID = '...'

def actor_run_abort(apify_client: ApifyClient, actor_id: str) -> None:
    actor_client = apify_client.actor(actor_id)
    actor_client.call(wait_secs=1)
    last_run = actor_client.last_run(status='RUNNING', origin='API')
    aborted_info = last_run.abort()
    print(f'aborted_info: {aborted_info}')
    
if __name__ == '__main__':
    apify_client = ApifyClient(TOKEN)
    actor_run_abort(apify_client, ACTOR_ID)   

# async version
import asyncio
from apify_client import ApifyClientAsync

TOKEN = '...'
ACTOR_ID = '...'

async def actor_run_abort_async(apify_client: ApifyClientAsync, actor_id: str) -> None:
    actor_client = apify_client.actor(actor_id)
    await actor_client.call(wait_secs=1)
    last_run = await actor_client.last_run(status='RUNNING', origin='API')
    aborted_info = await last_run.abort()
    print(f'aborted_info: {aborted_info}')

if __name__ == '__main__':
    apify_client = ApifyClientAsync(TOKEN)
    asyncio.run(actor_run_abort_async(apify_client, ACTOR_ID))

Aborting of the last task run does not work

• Resulting in:

apify_client._errors.ApifyApiError: We have bad news: there is no API endpoint at this URL. 
Did you specify it correctly?

• Code to reproduce it:

# sync version
from apify_client import ApifyClient

TOKEN = '...'
TASK_ID = '...'

def task_run_abort(apify_client: ApifyClient, task_id: str) -> None:
    task_client = apify_client.task(task_id)
    task_client.call(wait_secs=1)
    last_run = task_client.last_run(status='RUNNING', origin='API')
    aborted_info = last_run.abort()
    print(f'aborted_info: {aborted_info}')

if __name__ == '__main__':
    apify_client = ApifyClient(TOKEN)
    task_run_abort(apify_client, TASK_ID)

# async version
import asyncio
from apify_client import ApifyClientAsync

TOKEN = '...'
TASK_ID = '...'

async def task_run_abort_async(apify_client: ApifyClientAsync, task_id: str) -> None:
    task_client = apify_client.task(task_id)
    await task_client.call(wait_secs=1)
    last_run = await task_client.last_run(status='RUNNING', origin='API')
    aborted_info = await last_run.abort()
    print(f'aborted_info: {aborted_info}')

if __name__ == '__main__':
    apify_client = ApifyClientAsync(TOKEN)
    asyncio.run(task_run_abort_async(apify_client, TASK_ID))

Related issues

• The aborting of the last task run was reported in No Endpoint for RunsClient.abort() #190. The aborting of the last Actor run does not work as well as was described above.

Solution

• Using additional API call to be able to call https://api.apify.com/v2/acts/{actorId}/runs/{runId}/abort in both cases.
• In the async version it results in last_run method being async (should not be released as patch version).
• Copied from source code (Actor run):

# Note:
# The API does not provide a direct endpoint for aborting the last Actor run using a URL like:
# https://api.apify.com/v2/acts/{actor_id}/runs/last/abort
# To achieve this, we need to implement a workaround using the following URL format:
# https://api.apify.com/v2/acts/{actorId}/runs/{runId}/abort

• Copied from source code (task run):

# Note:
# The API does not provide a direct endpoint for aborting the last task run using a URL like:
# https://api.apify.com/v2/actor-tasks/{task_id}/runs/last/abort
# To achieve this, we need to implement a workaround using the following URL format:
# https://api.apify.com/v2/acts/{actorId}/runs/{runId}/abort

Testing

• As we do not have a proper testing framework here, it was tested just manually with the code examples provided at the beginning.

[github-actions]

⚠️ Pull Request Tookit has failed!

Pull request is neither linked to an issue or epic nor labeled as adhoc!

[B4nan]

why a feature release when you only have fixes in it?

[vdusek]

From PR description:

In the async version it results in last_run method being async (should not be released as patch version).

So, since we are making additional requests, the last_run method must be made async. This change is technically breaking, although the original implementation was broken. Therefore, I have released it as a new minor version. This means it will not be updated automatically if you are using a tilde pin like apify-client~=1.6.0.

[B4nan]

Ok, then please mention it in the changelog too. I would still go with a patch release personally, especially since it was broken before anyway, but it's not a huge deal. I don't know how common it is to lock versions via ~, but I am quite sure people do not do that to get around breaking changes.