Improved CSG filling strategies

[JGuetschow]

Pull request

Please confirm that this pull request has done the following:

• Tests added
• Documentation added (where applicable)
• Description in a {pr}.thing.md file in the directory changelog added - see changelog/README.md for details

Description

Add more filling strategies for the csg

• local trend based filling: similar to primap1 method
• local least squares filling: improvement over global least squares with configurable matching area

Important

Merge #286 first

[codecov]

Codecov Report

Attention: Patch coverage is 99.19355% with 4 lines in your changes missing coverage. Please review.

Project coverage is 96.94%. Comparing base (09060da) to head (2bcbcec).

✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
primap2/csg/_strategies/gaps.py	97.72%	3 Missing ⚠️
primap2/csg/_wrapper.py	96.15%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #288      +/-   ##
==========================================
+ Coverage   96.70%   96.94%   +0.23%     
==========================================
  Files          49       54       +5     
  Lines        4710     5205     +495     
==========================================
+ Hits         4555     5046     +491     
- Misses        155      159       +4     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[JGuetschow]

Ready for review. Docs still to come, but that should be quick as the class docstring is quite detailed.

[mikapfl]

Suggested change

	) -> xr.DataArray:
	) -> xr.DataArray:
	"""Fill a gap in a timeseries using data from another timeseries and factors.
	Parameters
	----------
	ts: xr.DataArray
	Timeseries with a gap to be filled.
	fill_ts: xr.DataArray
	Timeseries from which to take the data to fill in the gap in ts.
	gap: Gap
	The gap that should be filled.
	factor: np.array
	Numpy array containing two floats, the scaling factor to use at the left side of the gap
	and the scaling factor to use at the right side of the gap. The data from fill_ts will be
	multiplied with the factor before filling the gap in ts. If the factors differ, a linear
	interpolation of the scaling factor between both values will be performed.
	Returns
	-----------
	result
	DataArray with the same size and shape as ts, but with the given gap filled from fill_ts.
	"""