create_item_bank: Convert corpus to item bank

In sebsilas/itembankr: Create Item Banks From Musical Corpuses

Convert corpus to item bank

Description

Convert corpus to item bank

Usage

create_item_bank(
  name = "",
  input = c("files", "files_phrases", "item_df", "phrase_df"),
  output = c("all", "file", "item", "phrase", "ngram", "combined"),
  midi_file_dir = NULL,
  musicxml_file_dir = NULL,
  input_df = NULL,
  launch_app = FALSE,
  remove_redundancy = TRUE,
  remove_melodies_with_only_repeated_notes = TRUE,
  remove_melodies_with_any_repeated_notes = FALSE,
  scale_durations_to_have_min_abs_value_of_x_seconds = 0.25,
  slice_head = NULL,
  distinct_based_on_melody_only = TRUE,
  lower_ngram_bound = 3L,
  upper_ngram_bound = NULL,
  get_ngrukkon = TRUE,
  phrase_segment_outlier_threshold = 0.65,
  phrase_segment_ioi_threshold = 0.96,
  return_item_bank = FALSE,
  save_item_bank_to_file = TRUE
)

Arguments

name	A string of the item bank name.
input	A string denoting the input. Must be one of "files", "files_phrases", "item_df" or "phrases_df". files_phrases is when the e.g., MIDI files have already segmented phrases, and you don't want segmentation.
output	A character vector denoting the desired output type or types. You cannot create an output backwards in the hierarchy from the input.
midi_file_dir	If the input is files, a directory with MIDI files. Files should be in the format item_bank_name0.mid.
musicxml_file_dir	If the input is files, a directory with musicxml files. Files should be in the format item_bank_name0.musicxml.
input_df	If using input item_df or _phrases_df, an input dataframe.
launch_app	Should the app be launched at the end?
remove_redundancy	Should redundant relative melodies be removed? i.e., multiple representations of the same melody in relative form.
remove_melodies_with_only_repeated_notes	Remove any melodies which consist only of a single note repeated.
remove_melodies_with_any_repeated_notes	Remove any melodies which contain any consecutive repeated notes.
scale_durations_to_have_min_abs_value_of_x_seconds	Scale melody durations to have a minimum of x seconds.
slice_head	NULL by default. Can be an integer to slice up to a certain number of items - useful for testing.
distinct_based_on_melody_only	If TRUE, when removing redundancy, check for uniqueness only based on the melody column. Otherwise check based on melody and durations. Default is TRUE.
lower_ngram_bound	The lowest ngram size to use.
upper_ngram_bound	The highest ngram to use, default is the melody length -1.
get_ngrukkon	Whether to compute similarity between parent melodies and sub-melodies.
phrase_segment_outlier_threshold	A threshold for phrase segmenetation sensitivity.
phrase_segment_ioi_threshold	A threshold for phrase segmenetation sensitivity.
return_item_bank	If TRUE, return the item bank from the function
save_item_bank_to_file	Should the item_bank be saved?

sebsilas/itembankr documentation built on July 16, 2025, 10:18 p.m.