MTH5 Spectrogram Housekeeping

[kkappler]

Depends on mth5 PR 270

• Once mth5 PR 270 is merged to master, update aurora to import the Spectrogram class from mth5.
  • Also remove test_spectrogram
  • Also change all usages of extract_band to import from mth5
• Make all "stft_obj" methods in aurora/timeseries_helpers take and return Spectrogram objects

Depends on fixing mth5 issue 271 (PR to be created soon)

• Once STFT capabilities are in mth5, import prewhitening methods (apply_prewhitening, apply_recoloring) from mth5
• Also move run_ts_to_stft_scipy to mth5/timeseries/spectre/stft.py
• Give spectrogram the ability to call any of these functions, and it will pass in it's dataset and rewrite its dataset.

mt_metadata: Depends on PR 236

• rgreppy FrequencyBands and update all imports and references to be mt_metadata.transfer_functions.processing.aurora.FrequencyBands
• rgreppy get_fft_harmonics and update imports to come from mt_metadata.transfer_functions.processing.aurora.window
• Shoud FCChannel should have a sample_rate_decimation_level value, or just a sample_rate?

A future PR could address recursive grep sample_rate_decimation and fix these.

• Add better dtyping to aurora/pipelines/transfer_function_kernel.py. In particular, dtyping is causing circular imports. This is probably due to TransferFunctionKernel being imported by process_mth5.py. Note that TransferFunctionKernel is needed in three functions within process_mth5.py module: merge_stfts, get_spectrogams, and process_mth5_legacy. TODO: fix circular import dtyping.

If possible, import the following from mth5

• add_fcs_to_mth5 (now in timeseries/spectre/helpers.py)
• read_back_fcs (now in timeseries/spectre/helpers.py)
• run_ts_to_stft_scipy (now in timeseries/spectre/helpers.py)
• calibrate_stft_obj (now in timeseries/spectre/helpers.py)
• Set prewhitening_type default to be empty string if it is None ... this needs a second look (JP)
• [ ]