Magdalena Fuentes

This PR fixes #540 and #541. ccing @bmcfee While looking at the `to_jams` functions I realized that they are unnecessarily complicated to accommodate `Salami` multi-annotator schema. I'm going to simplify...

Otherwise the first track of the index is loaded. Contributors right now realize this because they submit their PR and it fails. An small explanation should be added to the...

Add MultiAnnotator class

2

This PR adds the MultiAnnotator class to group annotations of same track by different annotators in a smooth and clear way. **IMPORTANT**: this PR deprecates the API for the `salami`...

Right now is not clear from the dataset table which datasets have downbeat, we should add that info. Also add a to_downbeat function for those datasets that make sense (at...

Add subsets in Zenodo for educational purposes

2

We discussed that we could select a couple of datasets and upload small subsets of them in Zenodo to have a few quickly downloadable datasets to use in MIR courses.

So users can choose the `sr` if they want to use `load_audio` directly. Right now `sr` is hard-coded, but it could be just a default parameter instead.

Docs inherited from Dataset showing fields they shouldn't

1

@PRamoneda good catch! `Multitrack` functionalities are showing up in the docs of **all datasets** but they shouldn't. Some of them: `choice_multitrack`, `choice_multitrack`, `load_multitracks`, etc

When we print a track, we get something like this (example from `cante100`): ``` Track( artist="Juanito Valderrama", audio_path="/Users/mf3734/mir_datasets/cante100/cante100audio/077_JunaitoCalderrama_CantesAmericanos.mp3", duration=185.0, f0_path="/Users/mf3734/mir_datasets/cante100/cante100midi_f0/077_JunaitoCalderrama_CantesAmericanos.f0.csv", identifier="48173727-7f4d-496e-b0ff-ee1d082b24cf", notes_path="...r_datasets/cante100/cante100_automaticTranscription/077_JunaitoCalderrama_CantesAmericanos.notes.csv", release="Magna antología del cante flamenco", spectrogram_path="...s/mf3734/mir_datasets/cante100/cante100_spectrum/077_JunaitoCalderrama_CantesAmericanos.spectrum.csv", title="Mi mulata",...

There's a nice example @nkundiushuti wrote in a notebook, we should add it to the docs

Increase test coverage

1

There are some lines of code here and there that are not covered, we should at some point include them in the tests as well to be closer to 100%....