4
84%
main: 84%

Ran 13 May 2026 06:09PM UTC

Files 25

Run time 2s

Badge

Embed ▾

Committed 13 May 2026 06:07PM UTC coverage: 70.987% (+0.09%) from 70.894%

Job # 25817469984.4

Build Type

push

github

Committed by

web-flow

Commit Message

Fix #49: parse IEDB v3 hierarchical-header CSV in load_iedb (#51)

The IEDB receptor_full_v3.zip export tcrsift downloads ships as a CSV
with a two-row header — top row is the section ("Receptor", "Epitope",
"Assay", "Chain 1", "Chain 2"), second row is the field. The previous
loader used sep="\t" and looked for flat headers like "Chain 2 CDR3
Curated", silently producing a dataframe without cdr3_alpha/cdr3_beta
when handed the real cached file. That was the root cause of the
KeyError #46 papered over with a defensive guard.

Now:

- Sniff the first line: if it starts "Receptor,Receptor" treat as the
  v3 hierarchical CSV; otherwise fall back to the legacy flat TSV
  parser for backward compatibility.
- v3 path filters to Receptor Type == alphabeta, then extracts
  cdr3_alpha from Chain 1 and cdr3_beta from Chain 2, preferring
  "CDR3 Curated" over "CDR3 Calculated" per chain.
- Drops rows with no CDR3 on either chain (unmatchable).

Verified end-to-end against the cached tcr_full_v3.csv (99MB,
225,664 alphabeta rows extracted, 64K with alpha, 195K with beta).

Bumps to 0.8.10.

Closes #49.

Coverage Stats

4683 of 6597 relevant lines covered (70.99%)

0.71 hits per line

pirl-unc / tcrsift / 25817469984 / 4
84%
main: 84%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job python-3.11 - 25817469984.4

pirl-unc / tcrsift / 25817469984 / 4 84% main: 84%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job python-3.11 - 25817469984.4

pirl-unc / tcrsift / 25817469984 / 4
84%
main: 84%

README BADGES
x