fix: replacing lambda with named function to make model pickable

[madtoinou]

Fixes #1587.

Summary

save() was not working on FTT model if the trend argument was neither "poly" or "exp" because the function lambda x : 0 is not pickable. Just created a class method to have a named function that can be pickled and updated the docstring.

Other Information

I'll make sure all the other models save() method are working before converting this draft PR into a real one.

[codecov-commenter]

Codecov Report

Base: 94.08% // Head: 94.04% // Decreases project coverage by -0.04% ⚠️

Coverage data is based on head (dc1a2db) compared to base (c80049b).
Patch coverage: 100.00% of modified lines in pull request are covered.

📣 This organization is not using Codecov’s GitHub App Integration. We recommend you install it so Codecov can continue to function properly for your repositories. Learn more

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1594      +/-   ##
==========================================
- Coverage   94.08%   94.04%   -0.04%     
==========================================
  Files         125      125              
  Lines       11171    11164       -7     
==========================================
- Hits        10510    10499      -11     
- Misses        661      665       +4     

Impacted Files	Coverage Δ
darts/models/forecasting/fft.py	96.42% <100.00%> (+0.23%)	⬆️
darts/utils/data/tabularization.py	99.27% <0.00%> (-0.73%)	⬇️
darts/timeseries.py	92.28% <0.00%> (-0.22%)	⬇️
darts/ad/anomaly_model/filtering_am.py	91.93% <0.00%> (-0.13%)	⬇️
...arts/models/forecasting/torch_forecasting_model.py	89.88% <0.00%> (-0.04%)	⬇️
darts/models/forecasting/block_rnn_model.py	98.24% <0.00%> (-0.04%)	⬇️
darts/models/forecasting/nhits.py	99.27% <0.00%> (-0.01%)	⬇️
darts/datasets/__init__.py	100.00% <0.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[madtoinou]

The pickle issue was also occurring for trend=exp since it was also relying on a lambda function. Created helper function for all 3 cases. I had to store the trend_coefficients in the model itself in order to have a clean exponential detrending function. Let me know if there is a better way to implement such functionality.

Noticed that the save/load cycle was not tested on all models, I don't know if we want to change this? I manually checked all the local models, it seems like FFT was the only one causing pickling issues anyway.

[dennisbader]

Looks good, thanks a lot!
Just added the Callable for better interpretability