Copy the midi_df, add cumulative time column, keep only note_on messages
Arguments
- midi_df
A dataframe representing a midi file returned from
midi_to_object(). This dataframe must contain columns named 'i_track', 'type', and 'time', where 'i_track' is the track number, 'type' indicates the type of event (e.g., "note_on" or "note_off"), and 'time' is the time point of the event.- track_num
An integer indicating the track number to be processed. Default is 0.
Value
A modified dataframe with an additional 'total_time' column and filtered only for "note_on" type.
Details
This function filters and modifies an input midi dataframe, midi_df provided by midi_to_object. It first filters the data to keep rows from the specified track_num and both "note_on" and "note_off". Then adds a total_time column which accumulates the time variable. Lastly, it filters again to return only "note_on" rows.
The general workflow is to make a copy of midi_df, which comes from pyramidi::miditapyr$unnest_midi(message_list_df), and then modify the copy with additional columns useful transforming MIDI messages with R data-wranglind techniques. The modified copy is then stripped of the additional columns, and returned to a state where it can be imported back into the miditapyr_object, which can be updated with the new midi information using miditapyr_object$midi_frame_unnested$update_unnested_mf(), and then written to a file using miditapyr_object$write_file("rando_mario.mid").