kbs: update report-nbumbers.kb

[fschwenn]

• Removes duplicate entries.

• Removes dashes from seek-terms.

• Adds new report numbers.

Signed-off-by: fschwenn florian.schwennsen@desy.de

[tsgit]

There are multiple patterns containing hyphens on the left side. I flagged only a few. Please revise. There is a duplicate pattern with ambiguous replacement.
I think we should have test cases along with this major update of patterns.

[tsgit]

so the same seek term now has 2 different replacement strings. I don't think that is well defined. probably the later one overwrites the earlier one

[tsgit]

hyphens should not be in the match pattern

[tsgit]

lots of unwanted hyphens

[tsgit]

hyphen

[tsgit]

not sure about the hyphen here, I don't think it'll match

[michamos]

I don't think either, as it should have been stripped

[tsgit]

I think it would be good to have a sample original reference line for each pattern, or at least each pattern group as a list of test cases. This would confirm the pattern does what is intended to do and no other pattern matches. At least would show when hyphens are incorrect.
e.g. similar to
https://github.com/inspirehep/refextract/blob/master/tests/test_tag.py#L132-L135
with tag_report (which you'd have to define) instead of tag_arxiv
or patterned after
https://github.com/inspirehep/refextract/blob/master/tests/test_engine.py#L102-L118 and elsewhere in that test file.
instead of multiple report_numbers I would focus on one at a time, or one pattern type at a time

[tsgit]

@fschwenn take a look at https://github.com/inspirehep/refextract/pull/67/files
there you see test-cases Melissa added for the Fermilab report number parsing

also, since that PR was merged, you need to rebase your PR

[fschwenn]

Dear Thorsten, I think I can add test routines following Melissas example. Thanks for pointing them out to me. But I will probably fail with the rebase stuff. Would it be ok for you to just reject my previous pull request and I implement my additions on top of the actual master (in a new branch)? I know that in principle git allows much more elegant ways to handle it. Best regards, Florian

…

-- Florian Schwennsen Deutsches Elektronen-Synchrotron DESY Building 01 Room O1.446 phone: +49-40-8998-6190

From: "Thorsten Schwander" ***@***.***> To: "inspirehep/refextract" ***@***.***> Cc: "Florian Schwennsen" ***@***.***>, "Mention" ***@***.***> Sent: Tuesday, 11 June, 2019 19:15:13 Subject: Re: [inspirehep/refextract] kbs: update report-nbumbers.kb (#66) [ https://github.com/fschwenn | @fschwenn ] take a look at [ https://github.com/inspirehep/refextract/pull/67/files | https://github.com/inspirehep/refextract/pull/67/files ] there you see test-cases Melissa added for the Fermilab report number parsing also, since that PR was merged, you need to rebase your PR — You are receiving this because you were mentioned. Reply to this email directly, [ #66?email_source=notifications&email_token=ABZ5NE6F6G4HACBJHRCZLKLPZ7MSDA5CNFSM4HO3KUV2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXN3TEQ#issuecomment-500939154 | view it on GitHub ] , or [ https://github.com/notifications/unsubscribe-auth/ABZ5NE4WMMWBML7RHAKJTZDPZ7MSDANCNFSM4HO3KUVQ | mute the thread ] .