pirl-unc / hitlist / 25032528769

Build # 25032528769

Build Type

push

github

Committed by web-flow

Commit Message

v1.28.0: hitlist report reads observations.parquet by default (CLI refactor PR 4) (#170)

The legacy ``hitlist report`` re-scanned the raw IEDB/CEDAR CSVs every
invocation (~3 min on a full corpus, with the visible
"Scanning mhc_ligand_full.csv:  7%|███" progress bar).  All 11 columns
the report uses are already in observations.parquet, so the rewrite is
just ``load_observations(columns=[...])`` instead.

Default path is now near-instant.  The legacy live-scan path is kept
behind ``--from-csv`` for the rare case where someone wants a sanity
check on a fresh download before ``hitlist build observations`` has run.

When observations.parquet is missing and ``--from-csv`` wasn't passed,
``run_report`` now prints a one-line hint pointing at
``hitlist build observations`` (the canonical build path from PR 2)
instead of silently scanning.

Tests: tests/test_report.py — 3 tests
- default path reads parquet and never calls scanner.scan
- --from-csv routes to scanner.scan with classify_source=True
- missing parquet without --from-csv prints the hint and returns ""

Coverage Stats

2946 of 4465 relevant lines covered (65.98%)

0.66 hits per line

Source File

Press 'n' to go to next uncovered line, 'b' for previous

Source Not Available