snapraid dup -l dups.log > /dev/null && snapraid-dup-log2csv.py dups.log dups.csv ';' && rm dups.log

snapraid-dup-log2csv.py

#!/usr/bin/env python3
import sys, errno, csv

if len(sys.argv) < 3:
    print('usage: snapraid', file=sys.stderr)
    sys.exit(errno.EINVAL)

def size_to_human(size, formatted_number=True):
    prefixes = ['B', 'KiB', 'MiB', 'GiB', 'TiB', 'PiB', 'EiB', 'ZiB', 'YiB']
    prefix_index = 0
    while size >= 1024 and prefix_index < len(prefixes)-1:
        size /= 1024
        prefix_index += 1
    # 3.00 -> 3, 3.141 => 3.14 - https://stackoverflow.com/questions/2440692/formatting-floats-in-python-without-superfluous-zeros
    if formatted_number:
        size = format(size, '.2f').rstrip('0').rstrip('.')
    return (size, prefixes[prefix_index])

with open(sys.argv[1], 'r', encoding='utf-8') as input:
    dups = []
    dups_size = 0
    for line in input:
        sline = line.split(':')
        if sline[0] != 'dup':
            continue
        dups.append({
          'source1': sline[1],
          'file1': sline[2],
          'source2': sline[3],
          'file2': sline[4],
          'size': sline[5]
        })
        dups_size += int(sline[5])

if not len(dups):
    print('No dups in the log file', file=sys.stderr)
    sys.exit(errno.ENOMSG)
else:
    print('{} duplicates found ({} {})'.format(len(dups), *size_to_human(dups_size)), file=sys.stderr)

dups.sort(key=(lambda x: int(x['size'])), reverse=True)

with open(sys.argv[2], 'w', encoding='utf-8-sig', newline='') as output:
    writer = csv.DictWriter(output, fieldnames=dups[0].keys(), delimiter=(sys.argv[3] if len(sys.argv) > 3 else None))
    writer.writeheader()
    for entry in dups:
        writer.writerow(entry)