Handle custom terms in datalink.bysemantics again.

[msdemlei]

This is done by extracting non-core terms -- really, anything with a URI
that starts with something that is not the core URI -- and excepting
it from our normalisation.

This would fix #298.

[codecov]

Codecov Report

Merging #299 (cf3c458) into main (b5b33e8) will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #299      +/-   ##
==========================================
+ Coverage   75.50%   75.53%   +0.02%     
==========================================
  Files          44       44              
  Lines        5128     5134       +6     
==========================================
+ Hits         3872     3878       +6     
  Misses       1256     1256              

Impacted Files	Coverage Δ
pyvo/dal/adhoc.py	66.58% <100.00%> (+0.48%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b5b33e8...cf3c458. Read the comment docs.

[andamian]

You need to append the PR number ([#299]) to make the validation happy. If this is an issue that requires quick release, we should do a minor release - 1.2.2. @tomdonaldson - do you agree?

[andamian]

Is there a reason to use lists instead of sets?

[msdemlei]

On Thu, Feb 10, 2022 at 05:52:46PM -0800, Adrian wrote: > @@ -1,6 +1,8 @@ 1.3 (unreleased) ================ +- pyvo deals with non-core terms in datalink.bysemantics again. + This would fix issue #298. You need to append the PR number (`[#299]`) to make the validation

Put that in, forced-pushed.

happy. If this is an issue that requires quick release, we should do a minor release - 1.2.2. @tomdonaldson - do you agree?

A minor release would be lovely, as it'd put lift quite a load from my consciousness...

> if isinstance(semantics, str): semantics = [semantics] - semantics = [term.lstrip("#") for term in semantics] + + core_terms, other_terms = [], [] Is there a reason to use lists instead of sets?

No -- one *could* work with sets all along and not just at the end. There's not much benefit to performance or clarity either way, though, so unless there's an aspect to this I'm missing I'd leave it this way.

[msdemlei]

I've asked the bug reporter about a verifier, and he offered

{{{
import pyvo
datalink_url = 'http://archive.eso.org/datalink/links?ID=ivo://eso.org/ID?SPHER.2016-09-26T03:04:09.308'
datalink = pyvo.dal.adhoc.DatalinkResults.from_result_url(datalink_url)
semantics = 'http://archive.eso.org/rdf/datalink/eso#calSelector_raw2master'
raw2master_url = next(datalink.bysemantics( semantics )).access_url
print(raw2master_url)
}}}

This is looking good. Hence I'd say this is fine to merge.

Is there anything I can do to get this into a release as fast as
possible such that ESO's notebooks work again?

[andamian]

@msdemlei - thanks for the fix. We'll try to do the release in the next day or so.