Script to correct file and exif dates in photos from Facebook dump

main.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

# This script will parse the JSON files from your Facebook dump and will copy the photos to a new folder
# with the correct date and time. It will also add the correct date and time to the EXIF data of the photo.
#
# Author: @tairosonloa at https://github.com/tairosonloa
# License: MIT
# Version: 1.0.0
# Date: 2023-12-23
#
# Usage: Execute this script with python3 in the root of the extracted zip file from the Facebook dump
# In that path, you should have a `your_activity_across_facebook/posts` folder with the JSON files and photos.
# Note: you might need to install third party libs (ftfy, piexif) with pip3

import os
import json
import codecs
import ftfy
import shutil
import piexif

from datetime import datetime

# This code was executed on a MacOS. Be sure to change `parse_path` to an existing folder in your computer.
# There is where photos with correct data and names will appear
parse_path = '/Users/example/Pictures/facebook_dump/' # note the trailing slash
albums = 'your_activity_across_facebook/posts/album'

for album in os.listdir(albums):
    with codecs.open(albums + '/' + album, 'r', encoding='ascii') as f:
        data = json.load(f)
        if len(data['photos']) > 0:
            # Correct bad encoding
            album_name = ftfy.ftfy(data['name'])
            print(f'Found {len(data["photos"])} photos for album \"{album_name}\"')

            # Create folder with the album name if it doesn't exist
            if not os.path.exists(parse_path + album_name):
               os.makedirs(parse_path + album_name)

            for photo in data['photos']:
                # Get the date from the photo. If exif data is not available, use the upload date of the photo
                taken_date = photo.get('media_metadata', {}).get('photo_metadata', {}).get('exif_data', [{}])[0].get('taken_timestamp', None)
                if taken_date is None:
                    taken_date = photo.get('creation_timestamp')
                
                # Copy the photo to the new folder
                destination_photo = parse_path + album_name + '/' + str(taken_date) + '.jpg'
                shutil.copy(photo['uri'], destination_photo)

                # Update EXIF data with the correct date
                exif_dict = piexif.load(destination_photo)
                new_date = datetime.fromtimestamp(taken_date).strftime("%Y:%m:%d %H:%M:%S")
                exif_dict['0th'][piexif.ImageIFD.DateTime] = new_date
                exif_dict['Exif'][piexif.ExifIFD.DateTimeOriginal] = new_date
                exif_dict['Exif'][piexif.ExifIFD.DateTimeDigitized] = new_date
                exif_bytes = piexif.dump(exif_dict)
                piexif.insert(exif_bytes, destination_photo)

                # Update the file creation and modification date
                os.utime(destination_photo, (taken_date, taken_date))