The warnings ought to have a class like BiopythonParserWarning to make it easy to filter on that. Also, are any of these triggered in the test suite? If so, the test should explicitly confirm that (and not just let it print to the terminal).

If the current tests do not hit these warnings, then ideally we'd need another test added...

Sure, I will check it out and hopefully finish next week. Will look into Biopython-specified warnings.

Also, are any of these triggered in the test suite? If so, the test should explicitly confirm that (and not just let it print to the terminal).

Need more elaboration, please.

If you run the test suite, do the new warnings appear in the terminal output? Note the online BLAST test is troublesome - I think there is an open issue on that.

Codecov Report

Attention: Patch coverage is 25.00000% with 6 lines in your changes missing coverage. Please review.

Project coverage is 86.31%. Comparing base (66ad0a0) to head (de6d507).

@@            Coverage Diff             @@
##           master    #5030      +/-   ##
==========================================
+ Coverage   85.45%   86.31%   +0.86%     
==========================================
  Files         286      286              
  Lines       59851    59853       +2     
==========================================
+ Hits        51144    51665     +521     
+ Misses       8707     8188     -519     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

the test suite
Just to clarify: where do I add test? Should i create new .xml in /Tests/Blast? Also is this correct that unittest is used rather than pytest?

Yes, unittest based - see https://biopython.org/docs/latest/Tutorial/chapter_testing.html

Nearly there - an indentation issue and a duplication:

Tests/test_NCBIXML.py:4319:9: SyntaxError: Expected an indented block after `with` statement
     |
4317 |         datafile = os.path.join("Blast", filename)
4318 |         with open(datafile) as handle:
4319 |         with self.assertWarns(BiopythonParserWarning) as cm:
     |         ^
4320 |             record = next(NCBIXML.parse(handle))
4321 |             self.assertRaises(StopIteration, next, NCBIXML.parse(handle))
     |

XML parser probably stops iteration when encounters unsupported tag, hence XML file appears to be empty for parser. Will look into it.

Unfortunately I dont have time recently for working on this issue, hopefully will return later.

Hey @head-ease08 and @peterjc, anything I can do to help?

@harisj739 Hey! You might look into the issue, afaik there is some underlying problem with XML parser

I guess that by adding this test, @head-ease08 found that the parser doesn't handle unepected tags nicely. It gives a warning (good), but then breaks?

If that is correct, then the simplest course of action is to raise an exception on the unexpected tag, and halt. And that should be an improvement on the current unclear failure.

@peterjc I can add that functionality building upon @head-ease08's code. If I do, should I create a sub-PR within this PR or an entirely separate PR?

I would suggest a new PR, using a new branch starting from and including (and thus crediting) Mark's initial work.