DurationField output format

[sevdog]

Description

As explained in #8527 this PR will provide the ability to choose the output format for DurationFields.

Since Python builtim timedelta object does not have any format function we are going to allow only formats which are already defined in Django codebase in the package django.utils.duration:

• duration_string
• duration_iso_string

[tomchristie]

Suggested change

	Format strings may either be the special string `'iso-8601'`, which indicates that [ISO 8601][iso8601] style intervals should be used (eg `'P4DT1H15M20S'`), or the special string `'standard'`, which indicates that Django interval format `'[DD] [HH:[MM:]]ss[.uuuuuu]'` hould be used (eg: `'4 1:15:20'`).
	Format strings may either be the special string `'iso-8601'`, which indicates that [ISO 8601][iso8601] style intervals should be used (eg `'P4DT1H15M20S'`), or the special string `'standard'`, which indicates that Django interval format `'[DD] [HH:[MM:]]ss[.uuuuuu]'` should be used (eg: `'4 1:15:20'`).

[sevdog]

Sorry for the typo 😅

I have fixed it locally, will update as soon as I have other elements to push because I would like to keep the history clear without having to perform too many rebase.

[tomchristie]

Okay. (Incidentally, all pull requests are squashed, so that doesn't matter at all from my perspective.)

[tomchristie]

If None, then DurationField serializer fields will return Python timedelta objects, and the duration encoding will be determined by the renderer.

What output style does this produce with the current JSONRenderer?

[sevdog]

Using the current JSONRenderer, which uses JSONEncoder it will output the string value of "total seconds" (doc).

django-rest-framework/rest_framework/utils/encoders.py

Lines 39 to 40 in 129890a

	elif isinstance(obj, datetime.timedelta):
	return str(obj.total_seconds())

Using the same example value will be:

from datetime import timedelta
d = timedelta(days=4, hours=1, minutes=15, seconds=20)
print(str(d.total_seconds())
# 350120.0

[tomchristie]

Thanks - good to talk this through from a docs-first perspective.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[pietzschke]

Can we reopen this?

[pietzschke]

I think it's not possible to contribute to the pr to solve the conflicts, @sevdog would you be able to do it?
Otherwise i will create a new one and do the fixes.

[sevdog]

Just rebased the PR, the "conflict" was a very silly issue in the doc file 🤷‍♂️

CI is failing due to a change in django-main branch, it should be addressed in an other PR.