Support fetching all completed tasks

[programmerPhysicist]

Question description

I get a HTTPError when I attempt to run get_completed_items() off the api object. I get the following error:

is this a bug, or am I doing something wrong?

[brunorosilva]

Hi, you seem to be using get_completed_items() without any project nor section filter, which is not supported by this api version (insane, I know). They should give other options to get all completed items, as it was possible before the version update.

Call to endpoint without project_id field

Call to endpoint with project_id field

There's no direct fix to this, but there's a workaround for non-archived projects. If you know all of the project_ids you can call it with get_all_completed_items(project_id="1231924012").

@lefcha do you guys have any plans to add back the call without the project_id filter? Or maybe give us a way to get archived projects.

[lefcha]

Yes, while the public API has a way to get all completed tasks, ie.

https://developer.todoist.com/sync/v9/#get-all-completed-items

Unfortunately we haven't yet implemented very much from the auxiliary functions of the Sync API in this SDK. Your best option for now would be to directly send a request to the https://api.todoist.com/sync/v9/completed/get_all endpoint.

There's also an API endpoint to get archived projects https://api.todoist.com/sync/v9/projects/get_archived, in case you find it useful, documented at:

https://developer.todoist.com/sync/v9/#get-archived-projects

This SDK is primarily about implementing the REST API, and we don't plan to implement all the auxiliary functions of the Sync API at the moment.

[oscarclivio]

Same problem here.

[brunorosilva]

@oscarclivio did you try to call the get_completed_items() using a project_id argument?

If you really want to get all of the completed items or you do not know the project_ids you're supposed to use, use the sync api - as lefcha answered me. If you're not sure how to do this, feel free to use this code on my working branch code

[oscarclivio]

Thanks @brunorosilva for helping. I do yes and I iterate over project ids. But your solution is faster and returns non-root completed tasks unlike get_completed_items() (weird!).

[brunorosilva]

@oscarclivio Glad I could help!

Yeah, that happens because their implementation of get_completed_items() calls to https://api.todoist.com/sync/v9/archive/items?project_id=<project_id>, and each call takes ~500ms to get 100 completed items (the limit for each call) for a single project as the example below. The endpoint I'm using is https://api.todoist.com/sync/v9/completed/get_all, which takes ~700ms to get 200 completed items (the limit), so this is why it's faster.

Their approach is not wrong, it is in line with their argument for having two APIs. Rest providing simpler use cases and Sync for data heavy use cases.

Below are some images and comparisons I made.

Using /archive
See the time below for one project -> 570ms

Using /completed

[programmerPhysicist]

I came up with the following implementation:

def dict_to_Task(obj, url):
    obj['comment_count'] = obj['note_count']
    obj['is_completed'] = (obj['completed_at'] != '')
    obj['created_at'] = "unknown"
    obj['creator_id'] = obj['user_id']
    obj['description'] = obj['content']
    obj['priority'] = ''
    obj['url'] = url
    return Task.from_dict(obj)

def get_all_completed_items(api):
    url = get_sync_url('completed/get_all')
    completed_items = get(_session, url, _token)
    tasks = completed_items['items']
    return [dict_to_Task(obj, url) for obj in tasks]

This probably could be added to the api, with a few tweaks.

[goncalossilva]

Hey, all! Our SDK is currently working as intended, but I'll repurpose this issue as a feature request for ways to get all completed tasks. It would indeed be a welcome addition to it.