24688079889
73%

Ran 20 Apr 2026 08:15PM UTC

Jobs 1

Files 30

Run time 1min

Badge

Embed ▾

Committed 20 Apr 2026 08:14PM UTC coverage: 44.33% (+3.0%) from 41.353%

Build # 24688079889

Build Type

push

github

Committed by

web-flow

Commit Message

v0.10.1: cache k-mer enumeration in cta_exclusive + human_exclusive_viral (#18)

Both cta_exclusive_peptides and human_exclusive_viral_peptides walked
the Ensembl proteome on every call to build their subtraction set
(~10-60s each). Extract the expensive traversal into two @lru_cache'd
helpers keyed on (ensembl_release, lengths):

- peptides._non_cta_proteome_kmers -> frozenset of every non-CTA k-mer
- viral._human_proteome_kmers -> frozenset of every human k-mer

Both public functions now call the helper and apply pd.Series.isin for
the mask. No public API change, no output schema change. The second
call within a process skips the traversal entirely; batch / REPL /
notebook workflows that re-invoke personalize() or targets() now get
sub-millisecond enumeration instead of full proteome walks.

Cache is bounded at maxsize=4 combinations. Memory cost per slot is
~300MB-1GB depending on length mix. Tests use a stubbed EnsemblRelease
so they run in <1s without Ensembl data.

174 passed (was 166; +8 new caching tests).

Closes #12.

Coverage Stats

817 of 1843 relevant lines covered (44.33%)

0.44 hits per line

Coverage Regressions

Lines	Coverage	∆	File
64	39.24	11.14%	viral.py
40	36.14	25.89%	peptides.py

Jobs

ID	Job ID	Ran	Files	Coverage
1	24688079889.1	20 Apr 2026 08:15PM UTC	30	44.33	GitHub Action Run

pirl-unc / tsarina / 24688079889
73%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Coverage Regressions

Jobs

Source Files on build 24688079889

pirl-unc / tsarina / 24688079889 73%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Coverage Regressions

Jobs

Source Files on build 24688079889

pirl-unc / tsarina / 24688079889
73%

README BADGES
x