upload_themes_to_archive.py

#!/usr/bin/env python3

import os
import sys
import csv
import requests
import datetime
from pprint import pprint
from internetarchive import upload

x = datetime.datetime.now()

theme_zip_files = {}

if sys.argv[1].endswith(".zip"):
    theme_zip_files[os.path.splitext(sys.argv[1].split("/")[-1])[0]] = sys.argv[1]
else:    
    for root, dirs, files in os.walk(sys.argv[1]):
        for file in files:
            if file.endswith(".zip"):
                theme_zip_files[os.path.splitext(file)[0]] = root +"/"+ file

preview_files = {}
for root, dirs, files in os.walk(sys.argv[2]):
    for file in files:
        if file.endswith(".png") or file.endswith(".txt"):
                theme_name = file.split("_")[0]
                item = os.path.splitext(file[file.find("_")+1:])[0]
                if theme_name not in preview_files:
                    preview_files[theme_name] = []
                preview_files[theme_name].append(root+"/"+file)


themes = {}
with open("../themes.csv") as infile:
    reader = csv.reader(infile, delimiter=';')
    for row in reader:
        themes[row[0]] = {'date': row[2], 'title': row[3]} 


for theme in theme_zip_files:
    md = {
        "collection" : "open_source_software", 
        "mediatype": "software",
        "subject": "Desktop Theme", 
        "creator" : "themeworld", 
        "date" : "",
        "description": '',
        "title": ""
        }
    if theme in themes:
        md['title'] = themes[theme]['title']
        md['description'] = "Windows 95/98/ME/XP theme: {}".format(themes[theme]['title'])
        md['date'] = themes[theme]['date']
        if theme in preview_files:
            upload_dict = {theme+".zip" : theme_zip_files[theme]}
            for file in preview_files[theme]:
                #print(file)
                if ".txt" in file:
                    if "author" in file:
                        author = ''
                        with open(file, 'r') as infile:
                            for line in infile:
                                author += line.strip().strip(":").strip().strip("*").strip("-")
                        if "simsanet theme-o-matic" not in author:
                            no_emails = author.split(" ")
                            author = ''
                            for word in no_emails:
                                if "@" not in word and "/" not in word and "-" not in word:
                                    author += word + " "
                        else:
                            author = ''
                        if author == '':
                            author = "themeworld"
                        md['creator'] = author
                    elif "text" in file:
                        text = "<div>Windows 95/98/ME/XP theme Read Me file:</div><div><br></div><pre>"
                        with open(file, 'r') as infile:
                            text += infile.read()
                        text +="</pre>"
                        md['description'] = text
                    continue
                if 'double' in file and theme+".1.png" not in upload_dict:
                    filename = theme+".1.png"
                else:
                    filename = file.split("/")[-1]
                upload_dict[filename] = file
            with open("log", "a") as logfile:
                try:
                    print("Uploading {}: {}".format(theme, md['title']))
                    logfile.write("Uploading {}: {}\n".format(theme, md['title']))
                    r = upload(theme, files=upload_dict, metadata=md, verbose=True, retries=5, retries_sleep=5)
                except requests.exceptions.HTTPError:
                    print("{} already exists, trying with {}_{}{:02}".format(theme, theme,x.year,x.month))
                    logfile.write("{} already exists, trying with {}_{}{:02}\n".format(theme, theme,x.year,x.month))
                    r = upload("{}_{}{:02}".format(theme,x.year,x.month), files=upload_dict, metadata=md, verbose=True, retries=5, retries_sleep=5)

                print("{} response code: {}".format(theme_zip_files[theme], r[0].status_code))
                logfile.write("{} response code: {}\n".format(theme_zip_files[theme], r[0].status_code))