Investigate weird pgscatalog.core.lib.pgsexceptions.MatchRateError

[nebfield]

Getting these on UK Biobank when split per chromosome and I don't know why 🤔

[nebfield]

This didn't do what I thought it did:

pygscatalog/pgscatalog.match/src/pgscatalog/match/lib/matchresult.py

Lines 281 to 288 in e88b41f

	if not all(x[1] for x in self.filter_summary.iter_rows()):
	logger.warning(f"{self.filter_summary}")
	[
	x.unlink() for xs in outfs for x in xs
	] # don't provide dodgy scoring files
	raise MatchRateError(
	f"All scores fail to meet match threshold {min_overlap}"
	)

A single failed match (false value) would trigger a MatchRateError.

I wanted to trigger a ZeroMatchesError if all matches failed.

The new approach is a bit more verbose:

pygscatalog/pgscatalog.match/src/pgscatalog/match/lib/matchresult.py

Lines 300 to 315 in 0b59582

	for x in self.filter_summary.iter_rows():
	try:
	if not x[1]:
	raise MatchRateError("Fails matching")
	except MatchRateError:
	# we reported the exception, but it's fine
	logger.warning(f"Score {x[0]} matching failed with match rate {x[2]}")
	continue
	else:
	logger.info(f"Score {x[0]} matching passed with match rate {x[2]}")
	# unless everything is bad!
	if not any(x[1] for x in self.filter_summary.iter_rows()):
	raise ZeroMatchesError(
	f"All scores fail to meet match threshold {min_overlap}"
	)

But I'm trying to be a bit clearer about the expected behaviour