FC group metadata not being stored with correct window information

[kkappler]

In tests/synthetic/test_fourier_coefficients.py FCs are added to an mth5 file, the file is written, and later those fcs are to be accessed.

This test passed when there was a boxcar window, but is not running as intended with a hamming window.

Fixes needed:

• The file written should have "hamming" and not "boxcar"
  The freshly minted mth5
  ls -l
  -rw-rw-r-- 1 kkappler kkappler 537852 Nov 18 17:05 test2.h5
  Around line 376 of process_mth5.py is called

save_fourier_coefficients(
                dec_level_config, i_dec_level, row, run_obj, stft_obj
            )

Within save_fourier_coefficients, the critical lines are:

fc_group = station_obj.fourier_coefficients_group.add_fc_group(
            run_obj.metadata.id
        )
decimation_level_metadata = dec_level_config.to_fc_decimation()
fc_decimation_level = fc_group.add_decimation_level(
            f"{i_dec_level}",
            decimation_level_metadata=decimation_level_metadata,
        )
fc_decimation_level.from_xarray(stft_obj)
fc_decimation_level.update_metadata()
fc_group.update_metadata()

After this exectutes, the metadata looks correct:

fc_decimation_level.metadata
Out[2]:
{
"decimation": {
"anti_alias_filter": "default",
"channels_estimated": [
"ex",
"ey",
"hx",
"hy",
"hz"
],
"decimation_factor": 1,
"decimation_level": 0,
"harmonic_indices": [
-1
],
"hdf5_reference": "",
"id": "0",
"method": "fft",
"min_num_stft_windows": 2,
"mth5_type": "FCDecimation",
"pre_fft_detrend_type": "linear",
"prewhitening_type": "first difference",
"recoloring": true,
"sample_rate_decimation": 1.0,
"time_period.end": "1980-01-01T11:04:00+00:00",
"time_period.start": "1980-01-01T00:00:00+00:00",
"window.clock_zero_type": "ignore",
"window.num_samples": 128,
"window.overlap": 32,
"window.type": "hamming"
}
}

The window .type is "hamming"

However, if I call
tmp = fc_group.get_decimation_level("0")

The window is boxcar:

tmp.metadata
Out[6]:
{
"decimation": {
"anti_alias_filter": "default",
"channels_estimated": [
"ex",
"ey",
"hx",
"hy",
"hz"
],
"decimation_factor": 1,
"decimation_level": 0,
"harmonic_indices": [
-1
],
"hdf5_reference": "",
"id": null,
"method": "fft",
"min_num_stft_windows": 2,
"mth5_type": "FCDecimation",
"pre_fft_detrend_type": "linear",
"prewhitening_type": "first difference",
"recoloring": true,
"sample_rate_decimation": 1.0,
"time_period.end": "1980-01-01T00:00:00+00:00",
"time_period.start": "1980-01-01T00:00:00+00:00",
"window.clock_zero_type": "ignore",
"window.num_samples": 128,
"window.overlap": 32,
"window.type": "boxcar"
}
}

• [ ]

The FCs are being w