Parse hashtags for Bluesky #515
Conversation
|
|
| link: str | ||
| card: SocialCard | None |
There was a problem hiding this comment.
Let's assign a default value to these new variables. This will help us avoid the UnboundLocalError which can occur if the variable is referenced before it's explicitly assigned a value within the code's scope. Providing a default value ensures the variable is always initialized and has a defined state.
| link: str | |
| card: SocialCard | None | |
| link: str | None = None | |
| card: SocialCard | None = None |
There was a problem hiding this comment.
I assumed that checking if the value isn't None would narrow the type down, which Pylance seems to agree, but Mypy seems to be arguing otherwise, is there something I'm missing on?
There was a problem hiding this comment.
Ah, dropping the != None works.
| # given that our needs of a hashtag is very simple, we can do away with | ||
| # only parsing alphanumeric characters | ||
| tag_regex = rb"(?:^|\s)#(?P<tag>[0-9]*[a-zA-Z][a-zA-Z0-9]*)" | ||
| text_bytes = text.encode("UTF-8") |
There was a problem hiding this comment.
If the text is a string, which it appears to be, then there's no need to cast everything to bytes objects.
The regex doesn't need the b prefix.
There was a problem hiding this comment.
@johnludwigm Thanks for taking the time to review the PR! I apologize for the late reply.
I'd like to explain the reason behind casting everything to bytes: Bluesky strings are internally UTF-8 encoded, and the documentation suggests handling strings in UTF-8 to compute "valid indexes" for elements like Links, Mentions, and Hashtags (you can learn more about it in this link). While UTF-8 encoding wouldn't be a problem for plain text posts, we often use emojis in our templates. Emojis can be encoded as multiple characters, leading to incorrect indexes if not handled properly.
I'll create a PR to tweak this helper.
There was a problem hiding this comment.
oooh right yeah I forgot about it needing to be in UTF-8 specifically
|
Hi @mary-ext and @johnludwigm! Thank you both! I'm the director here at Free Law Project, and I'm amazed that this just sort of got done. Open source, man — every once in a while it blows you away. I'll also introduce @ERosendo, who is the maintainer for this area of the code, so that's why he's chiming in with the review. @mary-ext, I know you just went ahead and did this work, but if you want to get in touch with me at [email protected], I'll be happy to PayPal you (or whatever) some amount of money for your time on this one (probably should have agreed on that before you jumped in, but I imagine we'll come up with a good number!). Thank you all! |
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
mary-ext
force-pushed
the
feat/bluesky-tags
branch
from
7dbaf14 to
a0393a4
Compare
|
It works! The embed doesn't have functional hashtags, but here's the first post with them:
Thank you @mary-ext!!! |
Fixes #514
A little rusty with Python, and wasn't sure how I'd get the Bluesky bot to work, thought I'd chip in just a little.