Feat/neural prophet

[BlazejNowicki]

Neural Prophet

Added basic wrapper for Neural Prophet.

Supported features

Univariate ✅
Multivariate ✅
Probabilistic ⬜️
Multiple-series training ⬜️
Past-observed covariates support ✅
Future-known covariates ✅

Other information

Usage examples in notebook:
neural_examples/examples.ipynb
This will be moved to another notebook or completely removed before merge.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[codecov-commenter]

Codecov Report

Base: 93.94% // Head: 93.21% // Decreases project coverage by -0.73% ⚠️

Coverage data is based on head (14f215a) compared to base (9c23beb).
Patch coverage: 0.00% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1436      +/-   ##
==========================================
- Coverage   93.94%   93.21%   -0.74%     
==========================================
  Files         122      123       +1     
  Lines       10723    10792      +69     
==========================================
- Hits        10074    10060      -14     
- Misses        649      732      +83     

Impacted Files	Coverage Δ
darts/models/forecasting/neural_prophet_model.py	0.00% <0.00%> (ø)
darts/ad/anomaly_model/filtering_am.py	91.93% <0.00%> (-0.13%)	⬇️
darts/timeseries.py	91.92% <0.00%> (-0.06%)	⬇️
...arts/models/forecasting/torch_forecasting_model.py	89.50% <0.00%> (-0.05%)	⬇️
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.

[hrzn]

A few generic comments about the PR:

• So far I think that the pinned requirement on Pytorch Lightning will be a no-go for the moment for us unfortunately.
• Could you import the model from darts/models/__init__.py ?
• Could you stash your changes to the two notebooks?
• Your model implementation is missing an abstract method implementation (_model_encoder_settings). But this can wait (don't spend more time on this until the dependency situation is figured out).

[ourownstory]

@BlazejNowicki Thank you for making NeuralProphet accessible to Darts! Let us know if you can use help with supporting more of the model's capabilities.

@hrzn Please excuse the delay in responding to the issue you raised. The dependencies were unnecessarily strict and have been relaxed. I hope this helps.
response to your issue
addressed in this PR

[hrzn]

@BlazejNowicki Thank you for making NeuralProphet accessible to Darts! Let us know if you can use help with supporting more of the model's capabilities.

@hrzn Please excuse the delay in responding to the issue you raised. The dependencies were unnecessarily strict and have been relaxed. I hope this helps. response to your issue addressed in this PR

Thanks a lot @ourownstory !

[AndrewJGroves]

Hi, is this something still be looked into?

[ourownstory]

Happy to support anyone wanting to complete this!

[cargecla1]

Happy to support anyone wanting to complete this!

Hello @ourownstory ,

I would like to do it, it is something I thought it will be good to do but didn't notice that someone else started it before.

Any chance you can also assist to finish this PR: #2092? The team has been busy.

Cheers.