#!/usr/bin/env python3
# encoding: utf-8

##
# This script reads the Artist, Copyright, UserComment, CreatDate, and ModifyDate Exif tags from the images and outputs them to a CSV
##

import argparse
import contextlib
import hashlib
import os
from pathlib import Path
from subprocess import check_output

##
# csv file layout:
# [0] = created date (taken from original CSV, isn't really the creation date, but should still be tracked)
# [1] = file path, relative to the repository root
# [2] = license name(s) -> Copyright Exif tag
# [3] = authorship information -> Artist Exif tag
# [4] = notes -> UserComment Exif tag
# [5] = modified date
# [6] = md5 hash
##

def read_exif(target,hash_data,error_file):
    return str(check_output(["exiftool", "-d", "%Y/%m/%d", "-efile1", error_file, "-f", "-api", "MissingTagValue=NO_DATA", "-p", "$CreateDate,$Directory/$FileName,$Copyright,$Artist,$UserComment,$ModifyDate,$hash", "-userparam", "hash=%s" % hash_data, target]), 'UTF-8')

def do_hash(file):
    md5 = hashlib.md5()
    with open(file, 'rb') as f:
        while True:
            data = f.read(65536)
            if not data:
                break
            md5.update(data)
    return str(md5.hexdigest())

##
# program logic start
##

args = argparse.ArgumentParser()
args.add_argument("--repo", default=".", help="The directory of the Wesnoth repository to run this script against.")
args.add_argument("--output", default="img-exif2csv_output.csv", help="The file to write the results of this script to.")
args.add_argument("--error", default="exiftool_errors.txt", help="The file to write the errors and warning from exiftool to.")
options = args.parse_args()
os.chdir(options.repo)

with contextlib.suppress(FileNotFoundError):
    os.remove(options.output)
    os.remove(options.error)

with open(options.output, 'w', encoding="utf-8") as f:
  f.write("Date, File, License, Author - Real Name(other name);Real Name(other name);etc, Notes, Modified Date, MD5 \n")
  for root, _, files in os.walk("."):
    for filename in files:
        filetype = Path(filename).suffix
        if filetype == ".png" or filetype == ".jpg" or filetype == ".webp":
            file = os.path.normpath(os.path.join(root, filename))
            hash = do_hash(file)
            f.write(read_exif(file,hash,options.error))