get_corpus_features: Get corpus features
In pmcharrison/voicer: Analyzing and Generating Chord Voicings

Analyzes a musical corpus to return a tibble describing feature values for all possible voicings of a series of chord sequences.

1 2	get_corpus_features(x, revoice_from, min_octave, max_octave, min_notes, max_notes, features = voice_features(), verbose = TRUE)

`x`	The corpus to analyze, expressed as a list of chord sequences, where each chord sequence is expressed as a list of chords, and each chord is expressed as a numeric vector of non-duplicated MIDI note numbers in ascending order. The corpus may also be expressed as a `corpus` object from the `hrep` package, and chords may be expressed as `pi_chord` objects from the `hrep` package.
`revoice_from`	(Character scalar) Determines the representation to which each chord is reduced before generating candidate revoicings. Currently supported values are `"pc_chord"` and `"pc_set"`, corresponding to `pc_chord` and `pc_set` from the `hrep` package respectively. In the former case, bass pitch classes are preserved, but in the latter they are not.
`min_octave`	(Numeric scalar) The minimum octave from which the candidate voicings should be drawn, expressed relative to middle C. For example, `min_octave = -1` means that all voicings will be drawn from C3 (MIDI note number 48) or above (see also `voice_opt`). An error will be thrown if the corpus contains chords that span lower than permitted by this argument.
`max_octave`	(Numeric scalar) The maximum octave from which the candidate voicings should be drawn, expressed relative to middle C. For example, `min_octave = 0` means that all voicings will be drawn from the octave beginning on middle C or lower, corresponding to pitch heights strictly less than C5 (MIDI note number 72) (see also `voice_opt`). An error will be thrown if the corpus contains chords that span higher than permitted by this argument.
`min_notes`	(Integer scalar) The minimum number of unique notes permitted in each candidate voicing. An error will be thrown if the corpus contains chords with fewer notes than this number.
`max_notes`	(Integer scalar) The maximum number of unique notes permitted in each candidate voicing. An error will be thrown if the corpus contains chords with more notes than this number.
`features`	A list of features to compute for each chord voicing (see `voice_opt`).
`verbose`	(Logical scalar) Whether or not to print progress messages.

A tibble where each row describes a candidate voicing for a particular chord in the corpus, along with its computed features, with the following columns:

id - An integer ID variable that uniquely indexes each chord position in the corpus, beginning at 1 and counting upwards.
seq - An integer ID variable that uniquely indexes each sequence in the corpus, beginning at 1 and counting upwards.
pos - An integer ID variable that uniquely indexes each chord position in each sequence, restarting at 1 for the beginning of each sequence, and counting upwards.
chosen - A logical variable indicating whether this particular voicing was the one observed in the original corpus. Each level of id should have chosen = TRUE exactly once.
midi - A character variable describing the pitch content of the voicing, with each pitch represented as a MIDI note number, and with these MIDI note numbers pasted together and separated by spaces.
Subsequent columns provide features values as computed by the functions provided in features.