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

##
# This script reads from the existing Wesnoth copyrights CSV file, and writes it to the CreateDate, ModifyDate, Artist, Copyright, and UserComment Exif tags of the images
# requires exiftool and cwebp
##

import argparse
import contextlib
import csv
import os
from pathlib import Path
from subprocess import check_output
from datetime import datetime
import sys
import re

##
# csv file layout:
# [0] = current file modified date -> CreateDate tag (not totally accurate)
# [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] = new git commit date (not used) -> ModifyDate (today)
# [6] = md5 hash (not used, not updated)
##

##
# Add new licenses to this list:
# Avoid things like "GNU GPL v2+;CC BY-SA 4.0", unless you mean to dual license
# under either GNU GPL v2+ or CC BY-SA 4.0.  GNU GPL v2+ and CC BY-SA 4.0 (e.g.
# a GNU GPL v2+ file with CC BY-SA 4.0 modifications) isn't legally possible.
##
known_licenses = (
        "CC BY-SA 4.0",
        "CC0",
        "GNU GPL v2+",
        )
# to check that the date has a four-digit pattern (the year), meaning it really is a date
dv = re.compile('\d\d\d\d/')

# "-all=" to clear out existing metadata before writing the data from the CSV
def write_exif(target,date_value,copyright_value,artist_value,usercomment_value):
    if dv.match(date_value) != None:
        date_value = datetime.strptime(str(date_value),"%Y/%m/%d").strftime("%Y:%m:%d %H:%M:%S")
    else:
        date_value = datetime.today().strftime("%Y:%m:%d %H:%M:%S")
    return str(check_output(["exiftool", "-overwrite_original", "-all=", "-ModifyDate=now", "-CreateDate=%s" % date_value, "-Copyright=%s" % copyright_value, "-Artist=%s" % artist_value, "-UserComment=%s" % usercomment_value, target]), 'UTF-8')

# This cwebp call is to work around a current exiftool limitation, where RIFF alpha tags cannot be written, yet are needed by Extended (because we now include Exif tags) WEBP if there is to be any transparency.
def fix_webp(target):
    return str(check_output(["cwebp", "-quiet", "-mt", "-z", "9", "-metadata", "exif", target, "-o", target]), 'UTF-8')


##
# 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("--input", default="copyrights.csv", help="The file to read the existing copyright data from.")
args.add_argument("--output", default="csv2img-exif_output.csv", help="The file to write the results of this script to.")
options = args.parse_args()
os.chdir(options.repo)


csv_data = {}
# Too few fields
missing_fields = []
# Too many fields, possibly due to an unquoted comma
extra_fields = []
# previously untracked images
added = []
# lacking something in either the license or author fields
incomplete = []
unchanged = []
removed = []
unknown_licenses = []


with open(options.input, encoding="utf-8") as csvfile:
    reader = csv.reader(csvfile)
    previous_file = ""
    for row in reader:
        if row[0] == "Date":
            continue

        file = row[1]
        previous_file = file

        if not os.path.exists(file):
            removed.append(file)
            continue

        csv_data[file] = row

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

# only transfer the info from complete, properly formatted CSV rows (unchanged)
# but keep track of all the rest for summary output
for root, _, files in os.walk("."):
    for filename in files:
        filetype = Path(filename).suffix
        if filetype == ".png" or filetype == ".jpg":
            file = os.path.normpath(os.path.join(root, filename))

            if not file in csv_data:
                added.append(["NEW", file, "", "", "", "NO_DATA", "NO_DATA"])
            elif len(csv_data[file]) < 7:
                missing_fields.append(csv_data[file])
            elif len(csv_data[file]) > 7:
                extra_fields.append(csv_data[file])
            elif csv_data[file][2].strip() == "" or csv_data[file][3].strip() == "":
                incomplete.append(csv_data[file])
            elif not csv_data[file][2] in known_licenses:
                unknown_licenses.append(csv_data[file][2])
                incomplete.append(csv_data[file])
            else:
                unchanged.append(csv_data[file])
                write_exif(file,csv_data[file][0],csv_data[file][2],csv_data[file][3],csv_data[file][4])

        elif filetype == ".webp":
            file = os.path.normpath(os.path.join(root, filename))

            if not file in csv_data:
                added.append(["NEW", file, "", "", "", "NO_DATA", "NO_DATA"])
            elif len(csv_data[file]) < 7:
                missing_fields.append(csv_data[file])
            elif len(csv_data[file]) > 7:
                extra_fields.append(csv_data[file])
            elif csv_data[file][2].strip() == "" or csv_data[file][3].strip() == "":
                incomplete.append(csv_data[file])
            elif not csv_data[file][2] in known_licenses:
                unknown_licenses.append(csv_data[file][2])
                incomplete.append(csv_data[file])
            else:
                unchanged.append(csv_data[file])
                write_exif(file,csv_data[file][0],csv_data[file][2],csv_data[file][3],csv_data[file][4])
                fix_webp(file)

# output reports
final_output = added + missing_fields + extra_fields + incomplete + unchanged

if options.output != "":
    with open(options.output, 'w', encoding="utf-8") as f:
        writer = csv.writer(f, lineterminator="\n")
        writer.writerow(["Date", "File", "License", "Author - Real Name(other name);Real Name(other name);etc", "Notes", "Needs Update", "MD5"])
        writer.writerows(final_output)
else:
    writer = csv.writer(sys.stdout, lineterminator="\n")
    writer.writerows(final_output)

any_check_failed = False

count_missing_fields = len(missing_fields)
count_extra_fields = len(extra_fields)
count_added = len(added)
count_removed = len(removed)
count_incomplete = len(incomplete)

if count_missing_fields > 0 or count_extra_fields > 0 or count_added > 0 or count_incomplete > 0:
    any_check_failed = True
    print("\nThere are "+str(count_missing_fields)+" rows with too few fields\n"+"\n".join(",".join(a) for a in missing_fields))
    print("\nThere are "+str(count_extra_fields)+" rows with too many fields, possibly due to an unquoted comma\n"+"\n".join(",".join(a) for a in extra_fields))
    print("\nThere are "+str(count_added)+" new images:\n"+"\n".join(a[1] for a in added))
    print("\nThere are "+str(count_removed)+" missing images:\n"+"\n".join(removed))
    print("\nThere are "+str(count_incomplete)+" images that lack license or author information"+"\n".join(a[1] for a in incomplete))

if len(unknown_licenses) > 0:
    any_check_failed = True
    print("Unknown licenses:")
    print("    " + "\n    ".join(unknown_licenses))

if any_check_failed:
    sys.exit(1)