Image Metadata Extraction

summarize.py

# -*- coding: utf-8 -*-
# This doesn't use any Python libraries for extracting image metadata. Instead,
# it relies on the very handy exiftool utility and parsing the results.
# exiftool: http://www.sno.phy.queensu.ca/~phil/exiftool/
from __future__ import unicode_literals
import unicodedata
import subprocess
import csv
import os
import re
from pymongo import MongoClient

def normalize_string(my_string):
    if my_string is None:
        return ''
    if type(my_string) != unicode:
        my_string = my_string.decode('utf-8', errors='ignore')
    return unicodedata.normalize('NFKD', my_string).encode('ASCII', 'ignore')

def extension(filename):
    return filename.split('.')[-1].lower()

def is_image(filename):
    return extension(filename) in ['jpg', 'jpeg', 'png']

def is_original_wp_image(filename):
    return not re.match('.*\-\d+x\d+\.jpg', filename)

def get_metadata_by_type(filename, metadata_type='EXIF'):
    metadata = {}
    p = subprocess.Popen(['exiftool', '-TAG', "-{0}:*".format(metadata_type), filename], stdout=subprocess.PIPE)
    out, err = map(normalize_string, p.communicate())
    output_lines = out.split('\n')
    output_lines = filter(lambda s: len(s) > 0, output_lines)
    for line in output_lines:
        k = metadata_type + ' ' + line.split(':')[0].strip()
        v = ''.join(line.split(':')[1:]).strip()
        metadata[k] = v
    return metadata

def get_all_metadata(filename):
    metadata = {}
    metadata.update(get_metadata_by_type(filename, 'EXIF'))
    metadata.update(get_metadata_by_type(filename, 'IPTC'))
    metadata.update(get_metadata_by_type(filename, 'XMP'))
    return metadata

if __name__ == "__main__":
    client = MongoClient()
    db = client.image_metadata
    wp_uploads = db.wp_uploads 

    for rootdir, dirs, filenames in os.walk('.'):
        for filename in filenames:
            if is_image(filename) and is_original_wp_image(filename):
                parent_folder = rootdir.split('/')[-1]
                full_path = os.path.abspath(os.path.join(rootdir, filename))
                print(full_path)
                this_metadata = {'path': full_path}
		this_metadata.update(get_all_metadata(full_path))
	        wp_uploads.insert(this_metadata)

Put summarize.py in the root folder where your images are and run it there. It will connect to a mongo instance on the localhost by default.