Sunset API versions 1 and 2

[Flix6x]

This PR sunsets API versions 1, 1.1, 1.2, 1.3 and 2.0.

• Move worthwhile tests to API v3
• Remove all other tests for sunset API versions
• Add sunset notice to changelog
• Remove obsolete documentation (keeping a reference to snapshot on readthedocs.org)
• Remove obsolete script for manual simulations against sunset API versions
• Add service listing to v3
• Document public endpoints in v3 (authentication, version listing and service listing)
• Responses with 410 (Gone) status codes * (see Support blackout tests #651)

* I'm thinking towards supporting FM hosts in running their own blackout tests, by switching a dedicated config setting, then actually removing the endpoints in flexmeasures==0.14, which will lead to 404 (Not Found). To be discussed in the next TSC.

[coveralls]

Pull Request Test Coverage Report for Build 4812786450

• 19 of 45 (42.22%) changed or added relevant lines in 4 files are covered.
• 798 unchanged lines in 35 files lost coverage.
• Overall coverage decreased (-9.2%) to 55.619%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
flexmeasures/api/v3_0/public.py	14	40	35.0%

Files with Coverage Reduction	New Missed Lines	%
flexmeasures/auth/decorators.py	1	69.23%
flexmeasures/data/models/assets.py	1	69.95%
flexmeasures/data/models/time_series.py	1	68.92%
flexmeasures/data/schemas/sensors.py	1	76.19%
flexmeasures/data/services/users.py	1	82.17%
flexmeasures/data/utils.py	1	76.39%
flexmeasures/api/common/schemas/sensors.py	2	73.21%
flexmeasures/api/v1_2/routes.py	2	87.18%
flexmeasures/api/v1_3/routes.py	2	87.01%
flexmeasures/conftest.py	2	97.66%

Totals
Change from base Build 4787293963:	-9.2%
Covered Lines:	6306
Relevant Lines:	10211

---

💛 - Coveralls

[nhoening]

Great work!

Coverage fell by 9.2%, but the report shows that's mostly because you left the implementations (why?) and in api_utils and validators, where I expect a lot of cleanup later.

[Flix6x]

Coverage fell by 9.2%, but the report shows that's mostly because you left the implementations (why?) and in api_utils and validators, where I expect a lot of cleanup later.

I left them in anticipation of supporting blackout tests, and will remove them immediately after release. Technically, I could have left the tests in place until then (and only update the docs and add the blackout test support before release). We could still do it like that, or accept the hit in coverage, which would show longer on Pypi (until the 0.14 release), and only very short on GitHub (until the main branch is updated with the PR that removes the endpoint implementations).

[nhoening]

Coverage fell by 9.2%, but the report shows that's mostly because you left the implementations (why?) and in api_utils and validators, where I expect a lot of cleanup later.

I left them in anticipation of supporting blackout tests, and will remove them immediately after release. Technically, I could have left the tests in place until then (and only update the docs and add the blackout test support before release). We could still do it like that, or accept the hit in coverage, which would show longer on Pypi (until the 0.14 release), and only very short on GitHub (until the main branch is updated with the PR that removes the endpoint implementations).

I'm fine with a short-term low coverage. It's not the most important metric.