Refactor magnet link copy function

[drew2a]

The function for copying the magnet link has been refactored. The code now is a bit safier.

Fixes #8016

[kozlovsky]

I have a mixed feeling about this PR.

First, it looks like the actual problem still needs to be fixed. If I understand correctly, the reason for the TypeError: 'NoneType' object is not subscriptable is that the field self.current_download is None.

At the beginning of other methods of the DownloadsDetailsTabWidget, we can see the following checks:

    def update_pages(self, new_download=False):
        if self.current_download is None:
            return
        ...

It looks to me that a similar check should be added to the beginning of the on_copy_magnet_clicked method.

Second, the resulting code after the PR refactoring does not look clearer than the original code. Originally, we had a 9-line function, and now we have two functions, 32 lines in total, that do the same, just in a more verbose form. Furthermore, now a future code developer should deal with two very similar functions, create_magnet and compose_magnetlink. It may be hard for a developer to understand, which one of these functions should be used in which case.

Can we instead add filtering valid tracker urls directly to the compose_magnetling function, completely avoiding the create_magnet static method?

[drew2a]

Yes, it is more verbose, but that is not necessarily a bad thing. More lines of code do not always equate to worse code.

It is more verbose because:

1. The logic was extracted into the function create_magnet, which makes it possible to properly test it. The previous version did not allow this.
2. The line:

        trackers = [
            tk['url'] for tk in self.current_download['trackers'] if 'url' in tk and tk['url'] not in ['[DHT]', '[PeX]']
        ]

was rewritten to:

        invalid_urls = {'[DHT]', '[PeX]'}
        urls = (t.get('url') for t in trackers)
        valid_urls = [u for u in urls if u and u not in invalid_urls]

This makes it more understandable by splitting a large and complex expression into two expressions with clear purposes:

1. Mapping the sequence
2. Filtering the sequence

[drew2a]

@kozlovsky, thank you for the review. I've addressed most of your comments except for the verbosity. For the explanation, please see above.

[kozlovsky]

@drew2a, thank you for addressing the comments; now the code looks clear and understandable!