pirl-unc / mhcgnomes / 33475426152
90%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 41
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 89.417% (+0.02%) from 89.393%
33475426152

push

github

web-flow 
Carry the caller's spelling to the gene lookup (#160) (#164)

parse_gene_without_species ranks candidate species by
declares_gene_with_same_case first, and the comment above it explains why:
"Ia1" belongs to Paralichthys olivaceus and "IA1" to Chrysolophus pictus, and
gene lookup normalizes case so only the spelling tells them apart.

That key had never fired. The tokenizer lower-cases every token, so the
function only ever saw "ia1", which neither species spells that way. It was
dead code the comment described as working, and the names it was written for
returned None.

Token.raw_string kept the spelling the whole time. Threading it to the gene and
allele lookups makes 19 gene forms resolve that returned None before, and none
stop resolving: Ia1/IA1, Ia2/IA2, AB1 (mouse) versus Ab1 (Roborovski hamster),
DAA-1, DAA-2, DAA2, DMB-1. Each has exactly one same-case declarer.

A second bug fell out of the first. The preference for a species the caller
named compared default_species -- usually a latin-name string -- against a list
of Species objects, so it only fired for callers passing an object. Nothing
noticed while the case key was dead. Once spellings worked, a UniProt line
reading "OS=Mus musculus OX=10090 GN=Mr1" lost its own species to Rattus sp.,
the only entry spelling MR1 as "Mr1". Resolving default_species before the
membership test puts a named species back above a spelling.

0 of 36,752 corpus names change.

Claude-Session: https://claude.ai/code/session_012s4siLj2Vm33eNMGjNSazH

4250 of 4753 relevant lines covered (89.42%)

0.89 hits per line

Coverage Regressions

Lines Coverage ∆ File
62
94.12
 0.06% parser.py
Jobs
ID Job ID Ran Files Coverage
1 33475426152.1 41
89.42
 GitHub Action Run
Source Files on build 33475426152