File size: 3,412 Bytes
4048d4f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 |
from pathlib import Path
import librosa
import csv
import json
import numpy as np
import music21 as m21
CSV_FILENAME = "./Edit_Data.csv"
SR = 44100
CROSS_FADE_SAMPLES = int(SR / 10)
def load_data():
# Load data from CSV file
csv_path = Path(__file__).parent / Path(CSV_FILENAME)
with open(csv_path) as csv_file:
csv_reader = csv.DictReader(csv_file, delimiter=',')
return list(csv_reader)
def load_syncpoints(score):
json_path = Path(__file__).parent / Path("../syncpoints/") / score.with_suffix(".json").name
return json.loads(json_path.read_text())
def get_nearest_measure(syncpoints, sample):
for measure_no, time_s in syncpoints:
if np.isclose(sample / SR, time_s, atol=0.6):
return measure_no
return None
def remove_measure_ranges(score, ranges):
for part in score.parts:
measures_to_remove = []
for measure in part.getElementsByClass(m21.stream.Measure):
if any(start <= measure.number <= end for start, end in ranges):
measures_to_remove.append(measure)
for measure in measures_to_remove:
part.remove(measure)
return score
musicxml_files = sorted((Path(__file__).parent / Path("../musicxml/")).glob("*.xml"))
for row in load_data():
edits = []
score = musicxml_files[int(row['Track ID']) - 1]
num_edits = int(row['Num Edits'])
adjust_samples = int(row['Adjust Samples'])
for n in range(num_edits):
edit_idx = str(n + 1)
edit_ref = int(row['Edit ' + edit_idx + ' Ref'])
edit_start = int(row['Edit ' + edit_idx + ' Start'])
edit_dur = int(row['Edit ' + edit_idx + ' Duration'])
# not needed for our purpose here
# if (n > 0):
# edit_start -= CROSS_FADE_SAMPLES
# edit_dur += CROSS_FADE_SAMPLES
source_start = edit_start - edit_ref - adjust_samples
source_end = source_start + edit_dur
edits.append([source_start, source_end, edit_start, (edit_start+edit_dur), row['Track ID']])
if len(edits) > 1:
seen_sections = []
repeated_measures = []
syncpoints = load_syncpoints(score)
for edit in edits:
source_section_times = (edit[0], edit[1])
if source_section_times in seen_sections:
target_section_times = (edit[2], edit[3])
# lookup measure number via syncpoints file
measure_from = get_nearest_measure(syncpoints, edit[2])
measure_to = get_nearest_measure(syncpoints, edit[3])
# print duration in minutes that were repeated
print(f"{score.name} {measure_from} to {measure_to} ({(edit[3] - edit[2]) / SR / 60:.2f} minutes)")
# collect measure ranges
repeated_measures.append((measure_from, measure_to))
seen_sections.append(source_section_times)
m21_score = m21.converter.parse(score)
remove_measure_ranges(m21_score, repeated_measures)
else:
m21_score = m21.converter.parse(score)
# make a new folder for the scores
new_folder = Path(__file__).parent / Path("../musicxml_no_repeats/")
new_folder.mkdir(parents=True, exist_ok=True)
# save new score
new_score_path = new_folder / Path(f"{row['Track ID']}.xml")
m21_score.write('musicxml', fp=new_score_path) |