Repository navigation
Conversation
…ow that og marks partial tests/test_write_manifest.sh now checks that INTERVALS set writes one run intervals row and unset or blank writes none, then reads that manifest with og summary collect beside a VCF on all 22 autosomes and expects the partial mark. Red on main: no step writes the row. Bead: D14
…s partial bin/write_manifest.sh writes run intervals from INTERVALS, and the Nextflow HTML_REPORT writes the same row from --intervals, so the scope check og summary collect already does (GeiserX#515) sees a run limited to intervals. No row for a whole-genome run. The CHANGELOG entry and the quality guide no longer say the row is never written. Bead: D14
|
Warning Review limit reachedYou've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Next included review available in 34 minutes. View limit details
No actionable comments were generated in the recent review. Walkthrough
Merge Risk:
|
|
@coderabbitai review |
Action not completed
|
… --intervals Section 5 of tests/test_write_manifest.sh runs the intervals lines of HTML_REPORT's task script in bash, with the value its def would hand them, and checks they write the row write_manifest.sh writes, and no row without --intervals. Red control: the module without the printf line does not write the row. Main's module fails 3 of these checks. Bead: D14
INTERVALS is shared with step 29, where genome means the whole genome and chip the CHIP genes. With either exported, a hand-run report marked a whole-genome run partial. write_manifest.sh now writes no row for them. HTML_REPORT puts --intervals on one line in its task script, as write_manifest.sh does, and its def only quotes the value. The test now evaluates that def and runs the task lines with a tab, a newline and a quote in the value, with red controls for each step. Bead: D14
A run limited with
--intervals(orINTERVALS="..."forrun-all.sh) gives callability, polygenic score, ancestry, telomere, ROH and archaic results that read as whole-genome results. #515 added a scope check toog summary collect: it marks those sections partial when the calls miss an autosome, or whenrun_manifest.tsvhas arun intervalsrow. But no step wrote that row. So a run limited to parts of all 22 autosomes (for examplechr1:1-1000000 ... chr22:1-1000000) was never marked partial.The fix
bin/write_manifest.shwritesrun intervals <regions>whenINTERVALSis set. The regions go on one line with single spaces, so a tab or newline in the value cannot break the row. A whole-genome run (unset or blank) gets no row.HTML_REPORTwrites the same row from--intervalsinto the manifest it givesog summary collect, beside thepipeline_versionrow it already writes.og-summaryneeds no change: it already readsrun intervals(feat(summary): record what part of the genome a run covered and mark genome-wide results partial #515,rust/crates/og-summary/src/collect/post.rs).docs/results/quality.mdsaid the row is never written. They now say who writes it.Output change:
run_manifest.tsvgets a newrun intervalsrow on a run limited with--intervals. No row, key or file is renamed or removed.Decisions:
docs/results/quality.mdis outside this package's file list. Its line said "No step writes this row yet", which this PR makes false, so I changed that one line. Leaving it would put a doc on main that contradicts the code.docs/run-state.md(in the file list) says nothing about the manifest rows or intervals, so it is unchanged.tests/test_write_manifest.shchecks it instead: it runs the intervals lines ofHTML_REPORT's task script in bash, with the value the module'sdefhands them, and checks they write the same row asbin/write_manifest.sh, and none without--intervals. The casenextflow-from-fastq-2-nextflowruns the Nextflow leg with--intervals, so it runs the new line end to end.How it was checked
tests/test_write_manifest.sh(it now carries the og-leg marker, so Guard's og-leg runs it with og built and fails if og is missing):INTERVALS=$' chr1:1-1000000\t chr2:1-1000000\n'writes exactly one row,run intervals chr1:1-1000000 chr2:1-1000000. Unset and blankINTERVALSwrite none.INTERVALSset.og summary collectreads the manifest the script wrote, beside a sample VCF with calls on all 22 autosomes:scope.autosomes_coveredis 22 and the mark ispartial: limited to the run's intervals (chr1:1-1000000 chr2:1-1000000)."mark": null.HTML_REPORT's intervals lines writerun\tintervals\tchr1:1-1000000 chr2:1-1000000for that value, and no row for an empty one.Red first: the new test run against main's
bin/write_manifest.sh(4ee60e6) with an og built from main fails twice:The Nextflow part, run against main's module (
HTML_REPORT_MODULEset to it), fails 3 checks: it has nodef intervalsand no intervals lines to run.On this branch the same run passes (
0 failure(s), 16 PASS lines), with og and without it (the og part is then skipped with a note). ShellCheck reports nothing on both shell files.CI on head 79c6fbc, the fast checks run on the fork's pool (the upstream runs were cancelled to free runners), all green:
rustjob runs the og-leg tests with og built;tests/test_write_manifest.shprints all 10 new PASS lines there, the three red controls included): https://github.com/geiresx/OwnGenome/actions/runs/44748821933E2E runs on the batch tip. The e2e scope picks every case that lists
modules/local/html_report/*orbin/write_manifest.sh. The ones that run the changed lines most directly:followups-nf-1-nf-wiring(the module and its manifest),nextflow-from-fastq-2-nextflow(the Nextflow leg with--intervals),reports-and-pgx-correctness-reports(the bash manifest) andrun-all-launcher.Bead: D14