Taskwarrior bindings for Python

Interact with your local taskwarrior task list from Python.

This is heavily inspired by @ralphbean's taskw library (that I'm also a contributor to), but breaks with many past decisions so as to have a cleaner internal API, a saner method of interacting with Taskwarrior itself, type annotations, and much improved maintainability.

Installation

pip install taskwarrior

You can also install the in-development version with:

pip install https://github.com/coddingtonbear/taskwarrior/archive/master.zip

Quickstart

>>> from taskwarrior import Client
>>> client = Client()
>>> tasks = client.filter(status='pending')
[Task(annotations=None, depends=None, description='Wake up', due=None, end=None, entry=datetime.datetime(2022, 1, 24, 4, 28, 11, tzinfo=tzutc()), id=1, imask=None, mask=None, modified=datetime.datetime(2022, 1, 24, 4, 28, 51, tzinfo=tzutc()), parent=None, project=None, recur=None, scheduled=None, start=None, status='pending', tags=['alarm'], until=None, urgency=0.8, uuid=UUID('a39ea0fa-682a-4815-9556-8b6785ee301c'), wait=None)]
>>> 

Examples

Finding Tasks

The filter method is used for finding tasks:

>>> from taskwarrior import Client
>>> client = Client()
>>> client.filter(status='pending')
[Task(annotations=None, depends=None, description='Wake up', due=None, end=None, entry=datetime.datetime(2023, 1, 24, 4, 28, 11, tzinfo=tzutc()), id=1, imask=None, mask=None, modified=datetime.datetime(2022, 1, 24, 4, 28, 51, tzinfo=tzutc()), parent=None, project=None, recur=None, scheduled=None, start=None, status='pending', tags=['alarm'], until=None, urgency=0.8, uuid=UUID('a39ea0fa-682a-4815-9556-8b6785ee301c'), wait=None)]

It accepts as parameters any of the following:

• Any number of keyword arguments:
  • Double-underscores are transformed into a . when sending the filter to Taskwarrior so as to allow you to use filters like description.contains or due.after.
  • Example of filter keyword arguments include:
    • status='pending'
    • description__contains='Some string'
    • due__after='yesterday'
• Any number of filter dictionaries:
  • These work much the same way as keyword arguments described above, but allow you to specify these as a raw dictionary instead.
• Any number of raw strings:
  • E.g. for filtering to tasks having only a certain tag, you can provide as a value +mytag.
• Any number of Q objects:
  • You can use Q objects to perform complex or and and queries using exactly the same keyword arguments, filter dictionaries, or string values descrribed above.
  • See "Using Q Objects" below for more information.

Each of these parameters are and-ed together should more than one parameter be provided. If you need to use or expressions, see "Using Q Objects" below.

If you are expecting to retrieve just a single task, you can use the .get method, too:

>>> from taskwarrior import Client
>>> client = Client()
>>> client.get(uuid="a39ea0fa-682a-4815-9556-8b6785ee301c")
Task(annotations=None, depends=None, description='Wake up', due=None, end=None, entry=datetime.datetime(2023, 1, 24, 4, 28, 11, tzinfo=tzutc()), id=1, imask=None, mask=None, modified=datetime.datetime(2022, 1, 24, 4, 28, 51, tzinfo=tzutc()), parent=None, project=None, recur=None, scheduled=None, start=None, status='pending', tags=['alarm'], until=None, urgency=0.8, uuid=UUID('a39ea0fa-682a-4815-9556-8b6785ee301c'), wait=None)

if either zero or more than one task is returned from your query, an exception will be raised.

Retrieving tasks

>>> from taskwarrior import Client
>>> client = Client()
>>> client.get(uuid="a39ea0fa-682a-4815-9556-8b6785ee301c")
Task(annotations=None, depends=None, description='Wake up', due=None, end=None, entry=datetime.datetime(2023, 1, 24, 4, 28, 11, tzinfo=tzutc()), id=1, imask=None, mask=None, modified=datetime.datetime(2022, 1, 24, 4, 28, 51, tzinfo=tzutc()), parent=None, project=None, recur=None, scheduled=None, start=None, status='pending', tags=['alarm'], until=None, urgency=0.8, uuid=UUID('a39ea0fa-682a-4815-9556-8b6785ee301c'), wait=None)

See "Finding Tasks" for more details -- this allows all of the functionality described there, except that it asserts that only a single task be returned. Note that you can use any combination of fields you might like for retrieving your single task, but uuid and id are the ones most likely to be of use to you.

Counting tasks

>>> from taskwarrior import Client
>>> client = Client()
>>> client.count(status='pending')
1

This allows all of the same filtering logic that you can find described in "Finding tasks" above; see that section for more details.

Adding Tasks

>>> from taskwarrior import Client, Task
>>> client = Client()
>>> task = Task(description="my new task")
>>> client.add(task)
>>> task.uuid
UUID('29d06231-525f-4a62-9e9f-dd0f680aaaff')'

Changing tasks

>>> from taskwarrior import Client
>>> import pytz
>>> client = Client()
>>> task = client.get(uuid="a39ea0fa-682a-4815-9556-8b6785ee301c")
>>> task.due = datetime.datetime(2029, 1, 1, 10, 0, tzinfo=pytz.timezone('America/Los_Angeles'))
>>> client.modify(task)

Just alter the fields on the object directly using native python datatypes and pass your altered object to modify and your task will be updated immediately.

Being Flexible

>>> from taskwarrior import Client
>>> client = Client(
        config_filename="/path/to/my/taskrc",
        config_overrides={
            "some": {
                "config": "values",
            }
        },
        task_bin="/path/to/bin/task",
    )

You can provide the following parameters when instantiating your client:

• config_filename: (Default: ~/.taskrc) The path of the taskrc file to use.
• config_overrides: A dictionary object representing configuration overrides to use when interacting with Taskwarrior. Nested dictionaries will be encoded into dotted configuration paths using the key name in their parent dictionary.
• task_bin: The path to the task binary to use.

Using Q Objects

Q objects (inspired by Django's objects of the same name can be used for building complex logical queries for filtering your tasks.

Q objects accept all of the same parameter types described in "Finding tasks" above, but can also be or or and-ed together using | or &:

>>> from taskwarrior import Client, Q
>>> import pytz
>>> client = Client()
>>> client.filter(
        Q(status='pending') | Q(
            status='waiting',
            due__before=pytz.timezone('America/Los_Angeles').localize(
                datetime.datetime.utcnow() + datetime.timedelta(days=7)
            )
        )
    )

How does this differ from taskw?

• This is a much younger library and may still have bugs.
• This interacts with Taskwarrior over its export and import ("plumbing") interfaces instead of interacting with its "porcelain" interfaces like modify. This makes this library internally a lot simpler and less likely to break as Taskwarrior's command-line API is changed in the future.
• This has far fewer options and controls. Mostly: those options and controls are of limited value and high maintenance cost anyway, so it's likely you won't even notice.
• This has much more sophisticated filtering capabilities.
• This uses a slightly simpler API for retrieving tasks -- rather than handing you a 2-tuple of values (id, {data}) this just hands you a data object (or list of them).
• This uses a third-party library (Pydantic) for data serialization/deserialization so as to remove those responsibilities from the library itself (and hopefully make it somewhat easier to maintain).
• This supports only modern versions of Taskwarrior. Specifically: only versions of Taskwarrior newer than 2.5.0 are supported.
• This supports only currently supported versions of Python. Specifically: only versions of Python currently receiving security updates are supported -- that means Python 3.7+ at the moment.