image_classifier/image_classifier/image_classifier.py

313 lines
14 KiB
Python
Executable file

#!/usr/bin/env python3
# -*- encoding: utf-8 -*-
#
# This script is licensed under GNU GPL version 2.0 or above
# (c) 2021 Antonio J. Delgado
#
import sys
import os
import logging
import json
import shutil
import re
import click
import click_config_file
from logging.handlers import SysLogHandler
import face_recognition
import pyexiv2
import PIL
class CustomFormatter(logging.Formatter):
"""Logging colored formatter, adapted from https://stackoverflow.com/a/56944256/3638629"""
grey = '\x1b[38;21m'
blue = '\x1b[38;5;39m'
yellow = '\x1b[38;5;226m'
red = '\x1b[38;5;196m'
bold_red = '\x1b[31;1m'
reset = '\x1b[0m'
def __init__(self, fmt):
super().__init__()
self.fmt = fmt
self.FORMATS = {
logging.DEBUG: self.grey + self.fmt + self.reset,
logging.INFO: self.blue + self.fmt + self.reset,
logging.WARNING: self.yellow + self.fmt + self.reset,
logging.ERROR: self.red + self.fmt + self.reset,
logging.CRITICAL: self.bold_red + self.fmt + self.reset
}
def format(self, record):
log_fmt = self.FORMATS.get(record.levelno)
formatter = logging.Formatter(log_fmt)
return formatter.format(record)
class image_classifier:
def __init__(self, debug_level, log_file, faces_directory, directory, no_move,
people_folder, recursive):
''' Initial function called when object is created '''
self.debug_level = debug_level
if log_file is None:
home_path = os.environ.get('HOME', os.environ.get('USERPROFILE', os.getcwd()))
log_file = os.path.join(home_path, 'log', 'image_classifier.log')
self.log_file = log_file
self._init_log()
self.faces_directory = faces_directory
self.directory = directory
self.known_people = self.load_known_people()
self.no_move = no_move
self.people_folder = people_folder
self.recursive = recursive
if os.access(directory, os.R_OK):
if self.recursive:
entries = self.recursive_scandir(directory)
else:
entries = os.scandir(directory)
self._log.debug(f"Processing {len(entries)} files...")
for entry in entries:
if not entry.name.startswith('.') and entry.is_file():
self.process_file(entry.path)
def recursive_scandir(path, ignore_hidden_files=True):
files = []
try:
for file in os.scandir(path):
if not file.name.startswith('.'):
if file.is_file():
files.append(file)
elif file.is_dir(follow_symlinks=False):
more_files = self.recursive_scandir(
file.path,
ignore_hidden_files=ignore_hidden_files
)
if more_files:
files = files + more_files
except PermissionError as error:
self._log.warning(f"Permission denied accessing folder '{path}'")
return files
def process_metadata(self, file):
self.metadata = pyexiv2.ImageMetadata(file)
self.metadata.read()
if 'Xmp.iptcExt.PersonInImage' in self.metadata.xmp_keys:
self._log.debug(f"People (before): \
{self.metadata['Xmp.iptcExt.PersonInImage'].raw_value} \
(type: {type(self.metadata['Xmp.iptcExt.PersonInImage'].raw_value)})")
def process_file(self, file):
''' Process a file, find faces, add EXIF information and
move it to the folder of the day'''
self._log.debug(f"Looking for faces in file '{file}'...")
folder_date = 'unknown-time'
dirname = os.path.dirname(os.path.realpath(file))
filename = os.path.basename(file)
people = list()
if not os.access(file, os.R_OK):
self._log.error(f"The file '{file}' is not readable.")
else:
if self.is_image(file):
self.process_metadata(file)
people = self.find_faces(file)
if people:
self._log.debug(f"Found {len(people)} known people in the image.")
self._log.debug(json.dumps(people, indent=2))
self.append_people(file, people)
if 'Exif.Photo.DateTimeOriginal' in self.metadata.exif_keys:
original_date = self.metadata['Exif.Photo.DateTimeOriginal'].value
self._log.debug(f"File creation time: {original_date} \
(type: {type(original_date)})")
folder_date = original_date.strftime('%Y.%m.%d')
if folder_date == 'unknown-time':
match = re.search(r'(?P<year>20[0-9]{2})[\-/\._]?\
(?P<month>[0-1]?[0-9])[\-/\._]?(?P<day>[0-3]?[0-9])', filename)
if match:
folder_date = f"{match.group('year')}.{match.group('month')}.\
{match.group('day')}"
else:
match = re.search(r'(?P<day>[0-3]?[0-9])[\-/\._]?\
(?P<month>[0-1]?[0-9])[\-/\._]?(?P<year>20[0-9]{2})', filename)
if match:
folder_date = f"{match.group('year')}.{match.group('month')}.\
{match.group('day')}"
else:
self._log.debug(f"The file '{file}' doesn't seem to be an image for PIL.")
folder = os.path.join(dirname, folder_date, filename)
self._log.debug(f"Time based folder name section '{folder_date}'")
new_path = os.path.dirname(folder)
if not os.path.exists(new_path):
os.makedirs(new_path)
if self.people_folder:
for person in people:
person_path = os.path.join(self.people_folder,
person.replace(' ', '.'),
folder_date)
if not os.path.exists(person_path):
os.makedirs(person_path)
self._log.info(f"Copying file '{file}' to person '{person}' folder,\
'{person_path}'...")
try:
shutil.copy(file, person_path)
except FileNotFoundError as error:
self._log.error(f"Error copying. File not found. {error}")
if not self.no_move:
if os.path.exists(os.path.join(new_path, filename)):
self._log.debug(f"Destination '{new_path}/{filename}' exists, removing it...")
os.remove(os.path.join(new_path, filename))
self._log.info(f"Moving file '{file}' to '{new_path}'...")
try:
shutil.move(file, new_path)
except FileNotFoundError as error:
self._log.error(f"Error copying. File not found. {error}")
else:
self._log.info(f"NOT moving file '{file}' to '{new_path}' because of --no-move")
def print_metadata(self):
print("IPTC keys:")
for key in self.metadata.iptc_keys:
print(f" {key}: '{self.metadata[key].raw_value}'")
print("EXIF keys:")
for key in self.metadata.exif_keys:
print(f" {key}: '{self.metadata[key].raw_value}'")
print("XMP keys:")
for key in self.metadata.xmp_keys:
print(f" {key}: '{self.metadata[key].raw_value}'")
def append_people(self, file, people):
new_list = list()
if 'Xmp.iptcExt.PersonInImage' in self.metadata.xmp_keys:
for person in self.metadata['Xmp.iptcExt.PersonInImage'].raw_value:
new_list.append(person)
for person in people:
if person not in new_list:
self._log.debug(f"Adding person '{person}'...")
new_list.append(person)
if 'Xmp.iptcExt.PersonInImage' in self.metadata.xmp_keys:
self.metadata['Xmp.iptcExt.PersonInImage'].value = new_list
else:
self.metadata['Xmp.iptcExt.PersonInImage'] = pyexiv2.XmpTag('Xmp.iptcExt.PersonInImage',
new_list)
self._log.debug(f"People (after): \
{self.metadata['Xmp.iptcExt.PersonInImage'].raw_value} \
(type: {type(self.metadata['Xmp.iptcExt.PersonInImage'].raw_value)})")
try:
self.metadata.write()
self._log.debug(f"Updated file '{file}'.")
except OSError as error:
self._log.error(f"Error writing metadata to picture file. {error}")
def is_json(self, data):
try:
json.loads(data)
except TypeError:
return False
return True
def load_known_people(self):
known_people = list()
self._log.debug(f"Looking for known faces in directory '{self.faces_directory}'...")
if os.access(self.faces_directory, os.R_OK):
with os.scandir(self.faces_directory) as faces_items:
for entry in faces_items:
if (not entry.name.startswith('.') and entry.is_file() and
self.is_image(self.faces_directory + os.sep + entry.name)):
self._log.debug(f"Identifying face in file '{entry.name}'...")
person = dict()
person['filename'] = face_recognition.load_image_file(self.faces_directory +
os.sep + entry.name)
person['name'] = os.path.basename(os.path.splitext(self.faces_directory +
os.sep + entry.name)[0])
encodings = face_recognition.face_encodings(person['filename'])
if len(encodings) > 0:
person['encoding'] = encodings[0]
known_people.append(person)
else:
self._log.info(f"No faces found in file '{entry.name}'.")
return known_people
def find_faces(self, file):
''' Find faces in an image/video file '''
people = list()
if self.is_image(file):
image = face_recognition.load_image_file(file)
encodings = face_recognition.face_encodings(image)
self._log.debug(f"Found {len(encodings)} faces.")
for known_person in self.known_people:
for encoding in encodings:
if face_recognition.compare_faces([known_person['encoding']], encoding)[0]:
if known_person['name'] not in people:
people.append(known_person['name'])
else:
return False
return people
def is_image(self, file):
try:
PIL.Image.open(file)
except OSError as error:
self._log.debug(f"File '{file}' is not readable by PIL. {error}")
return False
except PIL.UnidentifiedImageError as error:
self._log.debug(f"File '{file}' is not an image recognizable by PIL. {error}")
return False
return True
def _init_log(self):
''' Initialize log object '''
self._log = logging.getLogger("image_classifier")
self._log.setLevel(logging.DEBUG)
sysloghandler = SysLogHandler()
sysloghandler.setLevel(logging.DEBUG)
self._log.addHandler(sysloghandler)
streamhandler = logging.StreamHandler(sys.stdout)
streamhandler.setLevel(logging.getLevelName(self.debug_level))
# formatter = '%(asctime)s | %(levelname)8s | %(message)s'
formatter = '[%(levelname)s] %(message)s'
streamhandler.setFormatter(CustomFormatter(formatter))
self._log.addHandler(streamhandler)
if not os.path.exists(os.path.dirname(self.log_file)):
os.mkdir(os.path.dirname(self.log_file))
filehandler = logging.handlers.RotatingFileHandler(self.log_file, maxBytes=102400000)
# create formatter
formatter = logging.Formatter('%(asctime)s %(name)-12s %(levelname)-8s %(message)s')
filehandler.setFormatter(formatter)
filehandler.setLevel(logging.DEBUG)
self._log.addHandler(filehandler)
return True
@click.command()
@click.option("--debug-level", "-d", default="INFO",
type=click.Choice(
["CRITICAL", "ERROR", "WARNING", "INFO", "DEBUG", "NOTSET"],
case_sensitive=False,
),
help='Set the debug level for the standard output.')
@click.option('--log-file', '-l', help="File to store all debug messages.")
@click.option("--faces-directory", "-f", required=True, help="Folder containing the pictures that \
identify people. Filename would be used as the name for the person. Just one person per picture.")
@click.option("--directory", "-d", required=True, help="Folder with the pictures to classify.")
@click.option("--no-move", "-n", is_flag=True, help="Don't move files, just add people's tag.")
@click.option('--people-folder', '-p', help="Define a folder for people's folders and copy \
pictures to each person's folder. Be sure to have deduplication in the filesystem to avoid using \
too much storage.")
@click.option('--recursive', '-r', is_flag=True, help='Recursively search for files in the provided --directory')
@click_config_file.configuration_option()
def __main__(debug_level, log_file, faces_directory, directory, no_move,
people_folder, recursive):
return image_classifier(debug_level, log_file, faces_directory, directory, no_move,
people_folder, recursive)
if __name__ == "__main__":
__main__()