return None for lyrics if Tekstowo fails to extract lyrics

[matlads]

Description

I experienced a failure to parse Tekstowo for song lyrics.

This patch allowed the lyrics plugin to fetch the lyrics from another provider as opposed to failing.

No bug was reported before this patch was generated

[wisp3rwind]

Looks good to me; see the inline comment for a minor improvement to make. If you'd like you could also add your credits to the changelog at docs/changelog.rst.

No bug was reported before this patch was generated

That's fine! As mentioned in the other pull request, it would however be great if you could briefly explain in the pull request how and why the code used to be wrong, as opposed to only saying that it failed. That would help a lot with review.

[wisp3rwind]

Small nitpick: Better rename this variable to something meaningful, e.g. lyrics_div (which is a name also used by another lyrics source).

[wisp3rwind]

As a side note, the tekstowo provider appears to have more issues of this kind:

beets/beetsplug/lyrics.py

Lines 463 to 465 in efcc5b3

	song_rows = soup.find("div", class_="content"). \
	find("div", class_="card"). \
	find_all("div", class_="box-przeboje")

will crash as soon as the website is changed in such a way that this tag hierarchy is modified one of the find's returns None`.

[sampsyo]

Hmm, that's a good point, @wisp3rwind—maybe we were a little hasty with enabling this source by default and would be better off leaving it as optional until it's better tested?

[wisp3rwind]

Hmm, that's a good point, @wisp3rwind—maybe we were a little hasty with enabling this source by default and would be better off leaving it as optional until it's better tested?

Actually, the class Tekstowo seems pretty simple, and on a first glance, in other places this None checks are in place. So I'd be fine with keeping it enabled. I'll go ahead, merge this PR and follow up with another one for these missing checks.

[wisp3rwind]

Thanks @matlads for pointing out this issue 🚀