Fix tag indexing

[adevyish]

• Fix exception if tag has / in it
• Fix tag links not working

[bbolli]

That's great, thanks! Does slugify() also work with Windows?
And, seeing that you don't need the first tuple element, it should be removed completely.

[adevyish]

I think it should but I don’t have access to a testing environment. Can fix up the tuple thing tomorrow but I was trying to get it working quickly 😅

[mcscope]

use tags.values() to get just the values instead of tuples

[adevyish]

Yes, I’m aware of this.

[adevyish]

As a sidenote, I didn't want to add an additional dependency but unicode-slugify also handles if you want to slug with non-ascii characters (which I'm doing for my own archive, since I have plenty of CJK tags)

[aspensmonster]

It seems like tumblr tags are a wonderful example of diverse user input that's always trying to outsmart the slug code. CJK tags, tags with slashes, tags with all kinds of odd unicode, emoji, multiple tags condensing down to one slug (whose sets aren't identical, so the rendered HTML is incomplete)...

Assuming you don't mind too much what the folder name is --if your use of the backup is mostly using the rendered HTML-- this approach seems to work for me for various weird tags (haven't found a broken or empty link yet, though there are thousands of tags in my backup):

import hashlib
...
tag_index = [self.blog.header('Tag index', 'tag-index', self.blog.title, True), '<ul>']
for index in sorted(self.tags.values(), key=lambda v: v.name):
    tag = hashlib.sha256(index.name.encode('utf-8')).hexdigest()
    etc etc etc

I'm also pretty sure hashlib is part of the standard library, so no additional module install is needed.