Issue 376 facilitate saving feedback annotations

[Flix6x]

This PR offers better support for saving and counting annotations on the level of generic assets. It also implements a Marshmallow validation field for generic asset ids.

closes #376

[coveralls]

Pull Request Test Coverage Report for Build 1928181827

• 31 of 70 (44.29%) changed or added relevant lines in 5 files are covered.
• 374 unchanged lines in 21 files lost coverage.
• Overall coverage decreased (-1.5%) to 69.104%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
flexmeasures/data/queries/annotations.py	12	16	75.0%
flexmeasures/data/schemas/generic_assets.py	4	9	44.44%
flexmeasures/data/models/generic_assets.py	8	16	50.0%
flexmeasures/data/models/annotations.py	5	27	18.52%

Files with Coverage Reduction	New Missed Lines	%
flexmeasures/data/models/legacy_migration_utils.py	1	92.86%
flexmeasures/data/services/users.py	1	85.63%
flexmeasures/ui/charts/latest_state.py	1	79.31%
flexmeasures/ui/utils/plotting_utils.py	1	46.2%
flexmeasures/cli/data_delete.py	2	26.11%
flexmeasures/conftest.py	2	98.74%
flexmeasures/data/models/generic_assets.py	4	71.08%
flexmeasures/data/schemas/sensors.py	4	78.79%
flexmeasures/ui/views/analytics.py	5	44.35%
flexmeasures/data/services/asset_grouping.py	6	80.46%

Totals
Change from base Build 1900760815:	-1.5%
Covered Lines:	6670
Relevant Lines:	9174

---

💛 - Coveralls

[nhoening]

Can you elaborate why this function should expect a DataFrame and not, say, a dict?

[Flix6x]

To be honest, I just happened to be handling a DataFrame with annotations in our GripOnGas plugin. There, the choice for a DataFrame makes sense because of the slicing and sorting functionality of the timing columns. So here, I designed just what I needed, a function to save annotations that accepts input in the form I already had.

[nhoening]

My take after thinking about this again: These functions are not for saving finished Annotations (the best way then would be to pass Annotation objects), but to create and save multiple annotations from raw data (for which a data frame is a good representation).

There are method names to better represent this fact. I also looked up the similar method TimedBelief.add (which in similar fashion gets a BeliefsDataFrame).

I thus believe Annotation.add would be a better choice for a name here.

We can then (in this PR or in follow-ups, I'd add TODOs) also use similar parameters:

• expunge_session
• allow_overwrite
• bulk_save_objects
• commit_transaction (you have this, requires just a renaming)

[nhoening]

Unsurprisingly, I have a renaming request.

[nhoening]

My take after thinking about this again: These functions are not for saving finished Annotations (the best way then would be to pass Annotation objects), but to create and save multiple annotations from raw data (for which a data frame is a good representation).

There are method names to better represent this fact. I also looked up the similar method TimedBelief.add (which in similar fashion gets a BeliefsDataFrame).

I thus believe Annotation.add would be a better choice for a name here.

We can then (in this PR or in follow-ups, I'd add TODOs) also use similar parameters:

• expunge_session
• allow_overwrite
• bulk_save_objects
• commit_transaction (you have this, requires just a renaming)