pydatatask.session module

A session is a tool for managing multiple live resources. Async resource manager routines can be registered, and in their place will be left a callable which will return the live resource while the session is opened.

For example:

session = pydatatask.Session()

@session.resource
async def my_file_dst():
    with open('/tmp/my_file_dst', 'wb') as fp:
        yield fp

@session.resource
async def my_file_src():
    with open('/tmp/my_file_src', 'rb') as fp:
        yield fp

async with session:
    my_file_dst().write(my_file_src().read())

The advantage over using normal contextlib context managers is that this produces a function reference which can be passed into other locations in sync contexts with the promise that it will not be called until the session is opened.

If a session is passed to a pipeline, it will be opened and closed when the pipeline is opened and closed.

Sessions cannot be opened more than once. But this doesn’t have to be the way! If you have a use case, complain in a GitHub issue, and I’ll see what can be done.

class pydatatask.session.Session

Bases: object

The session class. See module docs for usage information.

resource(manager: Callable[[], AsyncGenerator[T, None]]) → Callable[[], T]

Decorator for resource managers. Should be called with an async function that will yield exactly one object, the live constructed resource, and then tear that resource down on completion.

async open()

Open the session, initializing all the resources.

This is automatically called when entering an async with session: block.

async close()

Close the session, tearing down all the resources.

This is automatically called when exiting an async with session: block.