updated GCN circular regex matcher to expand matches

[Courey]

Description

This update expands the regex pattern to include:

• things that start with a parenthesis: (GCN 25)
• include an optional period after circ: GCN circ 25 AND GCN circ. 25
• include an optional # before the ID: GCN nasa-gcn/gcn.nasa.gov#25

Related Issue(s)

Resolves #51

Testing

I looked into the cases pointed out by Tyler in the Issue. I used an existing circular that I added to the seed file. In this circular I added many different patterns in many different positions to show how it behaves in different cases. These are the results:

[Courey]

NOTE: I did NOT include this pattern:
GCN note [#016] I don't think it occurs very often and the conversation in the Issue mentioned updating the circular to bring the format in line with others. Let me know if you also want me to add this pattern or if it's okay for us to update the circular to make it GCN nasa-gcn/gcn.nasa.gov#16 instead.

[lpsinger]

• Please add test cases.
• Why didn't it recognize text preceded by parentheses before?

[tylerbarna]

resolves #51

[Courey]

• Please add test cases.

  • Why didn't it recognize text preceded by parentheses before?

Because you have a lookback statement (?<=^|\\s) that specifically says that it must be preceded by the start of the string OR a space. So it won't match on a parenthesis because it is not a space and the start of the string is ( and not the specified pattern. Here is a side by side comparison. The one on the left is the updated version, the one on the right is the old version.

[lpsinger]

• Please add test cases.

  • Why didn't it recognize text preceded by parentheses before?

Because you have a lookback statement (?<=^|\\s) that specifically says that it must be preceded by the start of the string OR a space. So it won't match on a parenthesis because it is not a space and the start of the string is ( and not the specified pattern. Here is a side by side comparison. The one on the left is the updated version, the one on the right is the old version.

OK, so why don't we add open parentheses to that lookback statement?

[Courey]

resolves nasa-gcn/remark-rehype-astro#51

This won't work because that issue is in the gcn repo not this repo so it will not auto link. I'll transfer the issue and it will get a new number.

[lpsinger]

Looks good, thanks. I'll approve and merge once the CI finishes. It's running slowly now because there are a lot of queued pipelines from dependabot updates.

[Courey]

The linter automatically makes changes and makes my intentionally large whitespace gap normally spaced, which breaks two of the tests in each file.

It also looks like carriage returns are messing thigns up in the json version on CI.

[lpsinger]

The linter automatically makes changes and makes my intentionally large whitespace gap normally spaced, which breaks two of the tests in each file.

It also looks like carriage returns are messing thigns up in the json version on CI.

Yes, but #45 should fix the newline comparison issues on Windows. I just rebased it into your branch.

[Courey]

The linter automatically makes changes and makes my intentionally large whitespace gap normally spaced, which breaks two of the tests in each file.
It also looks like carriage returns are messing thigns up in the json version on CI.

Yes, but #45 should fix the newline comparison issues on Windows. I just rebased it into your branch.

It does not seem to have worked.

I also took the giant diff that it gave on the html file that it said was not matching and diffed them and could find no difference. I suspect that means it's an invisible whitespace character difference as well.