Commits for braney
switch to files view, user index
v498_preview2 to v498_base (2026-05-04 to 2026-05-11) v498
Show details
e214ffa17d2231d1a7413af72c539c1d3bfa0af2 Wed May 6 13:36:15 2026 -0700
- hgConvert quickLift: still lift hidden composite siblings
c4e77d5 stopped validating hidden subtracks to keep them out of the
"failed to lift" warning, but that also dropped them from the generated
quickLift hub, so a composite came across with only its visible
subtracks. Validate every sibling again so all liftable ones make it
into the hub; pass a NULL badList for hidden siblings so non-liftable
ones still don't show up in the user-facing complaint.
Verified with refSeqComposite (default vis: only ncbiRefSeqCurated
visible): pre-fix hub contained only ncbiRefSeqCurated; post-fix hub
contains all 10 RefSeq subtracks with hidden ones marked
"parent refSeqComposite off". cons100way badList unchanged: hidden
phastCons100way (wig, non-liftable) is silenced on both.
refs #36125
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
01badfbaa97a6547c7b938f79d48e454f1f09dce Wed May 6 14:23:26 2026 -0700
- hgc bigBed: only emit <BR> and set found after passing all filters, refs #36335
- src/hg/hgc/bigBedClick.c - lines changed 10, context: html, text, full: html, text
2654a3d85847492877e5d2b48d921a292e991771 Thu May 7 09:40:26 2026 -0700
- hgConvert quickLift: fix subtrack selection, hgc Predicted Protein /
CDS FASTA links
Three remaining issues on this ticket:
(vii) After "+ Select all subtracks" on hg38 refSeqComposite then
quickLifting to hs1, only RefSeq Curated rendered. dumpTdbAndChildren
built buffer="<track>_sel" but then read cart[tdb->track] (the visibility,
not the checkbox state); the visibility is never "1", so every subtrack
emitted "parent <comp> off" and only trackDb-default-on subtracks
rendered. Read cart[buffer] and accept any cartUsualBoolean-true value
("on" or positive int).
Predicted Protein blank page on quickLifted refGene:
getGenePredForPositionSql passed the hub-prefixed tdb->table to
quickLiftSql, which then queried the source DB (hg38) for a
"hub_NNN_refGene" table that doesn't exist. Strip the prefix with
trackHubSkipHubName before the source-DB query. Also, doRefGene was
overriding pepName to rl->mrnaAcc for every quickLifted item, so for
non-coding NR_* the mrnaAcc matched gbSeq and we offered a Predicted
Protein link that htcTranslatedPredMRna can only abort on. Only
override pepName for coding mRNAs; non-coding falls through to the
existing "Non-protein coding gene..." message.
CDS FASTA "Unknown database hub_NNN_<db>": addPalLink generates an hgPal
URL with no db=, so hgPal tries to open the destination hub-virtual db
and dies. palInfo coords are in destination space anyway while the
multiZ alignment lives in source coords, so skip addPalLink for
quickLifted tracks.
refs #36125
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
e58fa7f310af32cf0b4e3c6037be9baf268c1ef8 Fri May 8 08:47:46 2026 -0700
- Avoid SQL connect to hub assembly when looking up table info in hgTables, refs #37519
- src/hg/hgTables/hgTables.c - lines changed 8, context: html, text, full: html, text
- src/hg/hgTables/mainPage.c - lines changed 7, context: html, text, full: html, text
0daab9c07e1fb7b75223d61b0faa52e34d94302e Fri May 8 10:31:34 2026 -0700
- Block hgTables output for QuickLift tracks and hide QuickLift groups, refs #37519
QuickLift remaps tracks from another assembly on the fly for display,
but the underlying bigBed data is in the source assembly's coordinates,
so output queries against the destination assembly's region return no
results (or, with a whole-genome query, dump source-assembly data with
source coordinates). Detect quickLiftUrl on the selected track and
errAbort from doTopSubmit and doSummaryStats with a clear message.
Also filter groups whose label starts with "QuickLift" out of the
Group dropdown in hgTables, and clear the cart's group var if it was
pointing at a QuickLift group.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- src/hg/hgTables/hgTables.c - lines changed 15, context: html, text, full: html, text
- src/hg/hgTables/hgTables.h - lines changed 6, context: html, text, full: html, text
- src/hg/hgTables/mainPage.c - lines changed 9, context: html, text, full: html, text
- src/hg/hgTables/sumStats.c - lines changed 1, context: html, text, full: html, text
782bdb2e04aa2724f33c328bb080bedf52c976e5 Fri May 8 12:03:28 2026 -0700
- Hide QuickLift groups from hgIntegrator track list, refs #37519
QuickLift remaps tracks from another assembly on the fly for display,
but the underlying data is in the source assembly's coordinates, so
output queries against the destination assembly don't make sense. In
hgIntegrator this surfaced two ways for QuickLifted tracks: the
"Choose fields" pop-up hung, and Get output errored with
"Unrecognized type 'genePred' for hub track". Match the hgTables fix
by skipping groups whose label starts with "QuickLift" in
cartJsonGetGroupedTrackDb (the only consumer is hgIntegrator's React
UI).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- src/hg/cgilib/cartJson.c - lines changed 5, context: html, text, full: html, text
switch to files view, user index