Delete all but one of jpgs taken at the same second

dedup.py

from PIL import Image, ExifTags
from collections import defaultdict
from datetime import datetime
import os
from tqdm import tqdm
from glob import glob
import sys

glob_string = sys.argv[1]
print(f'Searching in globstring: {glob_string}')
all_data = {}
for filename in tqdm(glob(glob_string, recursive=True)):
    if not os.path.isdir(filename):
        img = Image.open(filename)
        exif = { ExifTags.TAGS[k]: v for k, v in img.getexif().items() if k in ExifTags.TAGS }
        all_data[filename] = exif

def date_func(x): return datetime.strptime(x['DateTime'], '%Y:%m:%d %H:%M:%S') if x.get('DateTime', None) else None
overlapping_images = defaultdict(list)
for f, x in all_data.items():
    try:
        v = date_func(x)
        if v:
                overlapping_images[v].append(f)
    except:
        print(f'Could not parse date {x}')

keepers = []
all_files = set()
for k,v in overlapping_images.items():
    all_files.update(set(v))
    if len(v) > 1:
        m_ind = 0
        max_size = 0
        for i, vv in enumerate(v):
            if 'original' in vv.lower():
                keepers.append(vv)
                break
            if 'cov.' in vv.lower() and 'photolibrary' in vv.lower():
                keepers.extend([xx for xx in v if 'photolibrary' in xx.lower()])
                break
        else:
            largest_image = list(sorted(v, key=lambda x: os.stat(x).st_size))[-1]
            keepers.append(largest_image)



keepers = set(keepers)
files_to_delete = all_files.difference(keepers)
print(len(files_to_delete))
for x in tqdm(files_to_delete):
    os.remove(x)