Support adding a row from a template

[ITAYC0HEN]

Hey 👋
This pull request implements a naive way to add a new row to a collection where its origin is a template (another block_id).

The implementation is based on a sequence of API calls:

• Enqueue a new task of type duplicateBlock (API: v3/enqueueTask)
• Loop N times every X seconds (or milliseconds) and check the status of the task (API: v3/getTasks)

The usage is very straightforward:

cv = client.get_collection_view("...")
cv.collection.add_row(sourceBlockId="75dd203a-EXAMPLE-a12b-ba492c37a1f6")

And of course, more parameters can be added:

cv.collection.add_row(
    sourceBlockId="75dd203a-EXAMPLE-a12b-ba492c37a1f6",
    name="The Hen House: Episode #123",
    type="Podcast"
)

Note:
The v3/getTasks actually supports getting the status of multiple tasks. I used it to get the status of a single one. We can consider improving it to support multiple tasks but didn't find it useful in the current design.

closes #194

[jamalex]

This is awesome! Very useful.

I agree that longer-term it would be nice to have a more general interface for task management, but unless we were going to make things pretty async, the current approach seems great.

A couple small suggestions.

[jamalex]

Would this be long enough for a huge template doc (with lots of children)? Worth upping the tries to 10, do you think?

[ITAYC0HEN]

done.
We can also consider passing interval and tries as an optional variable to add_row(...) so it can be configured when calling the function that uses wait_for_task(...) and not only when directly calling wait_for_task(...).

So currently, users can't control the intervals and tries of add_row(...) and I am not sure they should

[ag0n1k]

Hi there!

My next step is use template on creation, did some CR, if u don't mind. ;)

[ag0n1k]

Sorry for the duplication, it's all new for me)

[ag0n1k]

Suggested change

	data = self.post(
	"getTasks",
	{
	"taskIds": [task_id]
	}
	).json()
	data = self.post("getTasks", dict(taskIds=[task_id]).json()

[ag0n1k]

That is the same with

if results is None:

Or something I don't know =)

[ag0n1k]

Suggested change

	print("Task takes more time than expected. Specify 'interval' or 'tries' to wait more.")
	logger.debug("Task takes more time than expected. Specify 'interval' or 'tries' to wait more.")

[ag0n1k]

i would rather use dict() instead of {}, especially when the keys are constants (not variables). But this decision up to @jamalex

[ag0n1k]

Suggested change

	if len(results) == 1:
	state = results[0].get("state")
	return state
	return None
	try:
	task = results.pop()
	return task.get("state", None)
	except IndexError:
	logger.error("There is no task {}".format(results))
	return None