mueller report sonification

generate_midi.py

from miditime.miditime import MIDITime
from statistics import mean, median
import csv

# settings
tempo_bpm = 400
low_pitch = 57
high_pitch = 92
low_velocity = 30
high_velocity = 127
longer_notes = True

# source data
num_pages = 0
chars_per_page = []
percent_redacted = []
with open('redactions.csv') as csv_file:
    csv_reader = csv.reader(csv_file, delimiter=',')

    percent_redacted_idx = -1
    ocr_chars_idx = -1
    redacted_chars_idx = -1
    index = 0

    for row in csv_reader:
        if num_pages == 0:
            ocr_chars_idx = row.index("characters")
            redacted_chars_idx = row.index("estimated_redacted_characters")
            percent_redact_idx = row.index("percent_redacted")
            num_pages += 1
        else:
            chars_per_page.append(float(row[ocr_chars_idx]) + float(row[redacted_chars_idx]))
            percent_redacted.append(float(row[percent_redact_idx]))
            num_pages += 1

print(f'Found {num_pages - 1} pages of data')
print(f'  chars-per-page: min={min(chars_per_page)} max={max(chars_per_page)} avg={sum(chars_per_page)/len(chars_per_page)}')
print(f'  percent_redacted: min={min(percent_redacted)} max={max(percent_redacted)} avg={sum(percent_redacted)/len(percent_redacted)}')

# notes: [time, pitch, velocity, duration]
midinotes = []
for i in range(len(chars_per_page)):
    page = i + 1
    chars = chars_per_page[i]
    percent_chars = (chars - min(chars_per_page)) / (max(chars_per_page) - min(chars_per_page))
    percent = percent_redacted[i]

    pitch = low_pitch + (percent * (high_pitch - low_pitch))
    velocity = low_velocity + (percent_chars * (high_velocity - low_velocity))

    # OPTIONAL extend previous note rather than adding a new one
    if longer_notes and len(midinotes) > 0 and midinotes[-1][1] == pitch:
        midinotes[-1][3] = midinotes[-1][3] + 1
    else:
        midinotes.append([i, round(pitch), round(velocity), 1])
    # print(f'p{page} - {i} {round(pitch)} {round(velocity)}')

# tempo and output destination
if longer_notes:
    mymidi = MIDITime(tempo_bpm, 'mueller_report_longer_notes.mid')
    mymidi.add_track(midinotes)
    mymidi.save_midi()
else:
    mymidi = MIDITime(tempo_bpm, 'mueller_report.mid')
    mymidi.add_track(midinotes)
    mymidi.save_midi()

mueller_report.mid

mueller_report_longer_notes.mid

redactions.csv