Feat/Uber TLC Dataset

[gdevos010]

Fixes #.

Summary

New dataset based on the FiveThirtyEight dataset.
Supports both daily and hourly sampling.

Other Information

[codecov-commenter]

Codecov Report

Merging #1003 (ca332e1) into master (abf12da) will decrease coverage by 0.33%.
The diff coverage is 74.35%.

@@            Coverage Diff             @@
##           master    #1003      +/-   ##
==========================================
- Coverage   92.92%   92.58%   -0.34%     
==========================================
  Files          76       76              
  Lines        7628     7704      +76     
==========================================
+ Hits         7088     7133      +45     
- Misses        540      571      +31     

Impacted Files	Coverage Δ
darts/models/forecasting/block_rnn_model.py	98.14% <ø> (ø)
darts/models/forecasting/nbeats.py	98.06% <ø> (ø)
darts/models/forecasting/nhits.py	99.25% <ø> (ø)
darts/models/forecasting/rnn_model.py	97.46% <ø> (ø)
darts/models/forecasting/tcn_model.py	96.84% <ø> (ø)
darts/models/forecasting/tft_model.py	96.92% <ø> (ø)
darts/models/forecasting/transformer_model.py	100.00% <ø> (ø)
darts/datasets/dataset_loaders.py	91.48% <63.63%> (-3.92%)	⬇️
darts/datasets/__init__.py	78.26% <65.21%> (-21.74%)	⬇️
darts/models/forecasting/pl_forecasting_module.py	93.49% <100.00%> (+0.54%)	⬆️
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 14d671a...ca332e1. Read the comment docs.

[hrzn]

Once more, a very nice addition, thanks a lot!
I just have a comment concerning leaving it as an option to return a multivariate series, or multiple series. Let me know if you think that's a good idea.

[hrzn]

Cool, only one minor suggestion left then we can merge :)

[gdevos010]

I added the functionality to the electricity dataset too

[gdevos010]

Is something else needed, or can we merge this branch?

[hrzn]

Is something else needed, or can we merge this branch?

I think we can merge soon. The test coverage seems to have decreased. Are you able to have a look why?

[gdevos010]

@hrzn It's probably the _to_multi_series. I can add a test for them, but the Electricity's _to_multi_series is very very slow.

[hrzn]

@hrzn It's probably the _to_multi_series. I can add a test for them, but the Electricity's _to_multi_series is very very slow.

You could maybe add a test that calls _to_multi_series with a small DataFrame, and just have a sanity check that what the function returns has the right format.

[hrzn]

Could we maybe put this one in datasets/test/Electricity_test.csv ? i.e., we create a test subfolder.

[hrzn]

Thanks @gdevos010. I assume the tests should pass once it is merged, right? The only last tiny thing I'd ask is to put the test dataset in a subfolder datasets/test/.

[gdevos010]

@hrzn Yes, the tests passed locally. I moved the file to tests/datasets.

[gdevos010]

This will be my last PR for about a month. I will be traveling without a laptop.

[hrzn]

@hrzn Yes, the tests passed locally. I moved the file to tests/datasets.

Thanks! I was referring to datasets/Electricity_test.csv though, which I think is still living under datasets/. But it's fine, I will merge the PR and we can change it later on.

[hrzn]

This will be my last PR for about a month. I will be traveling without a laptop.

Enjoy :)