You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
PR #89 ("ci: migrate parity tests from dwgsim/phiX174 to holodeck/chr22") established the canonical holodeck install path for this repo: test/holodeck/pixi.toml pinned to bioconda holodeck==0.2.1, invoked from CI via pixi run --frozen --manifest-path test/holodeck/pixi.toml -- holodeck ….
PR #90 ("feat(meth): emit Bismark-compatible XR/XG/XM auxiliary tags") added test/meth/test_bismark_tags.sh, a Layer-3 golden-truth diff against a holodeck-simulated bisulfite BAM. That test pins a different holodeck — SHA d240816c8c2dfbd72bdb3386679d0ad4d4f223e6 on branch nh/bisulfite-conversion, built via cargo build --release from a /tmp/holodeck-<sha> cache.
Why two install paths today
The nh/bisulfite-conversion branch (fg-labs/holodeck#11) adds --methylation-mode em-seq and the Bismark-compatible golden BAM emission that test_bismark_tags.sh exercises. fg-labs/holodeck#11 is open / not merged, so v0.2.1 (the bioconda release pinned by PR #89's pixi env) does not yet expose those flags. There is no released holodeck version that supports both the chr22-parity flow PR #89 uses and the bisulfite-chemistry flow PR #90's Layer-3 test needs, so PR #90 had to roll its own cargo-from-SHA install.
Context
PR #89 ("ci: migrate parity tests from dwgsim/phiX174 to holodeck/chr22") established the canonical holodeck install path for this repo:
test/holodeck/pixi.tomlpinned to biocondaholodeck==0.2.1, invoked from CI viapixi run --frozen --manifest-path test/holodeck/pixi.toml -- holodeck ….PR #90 ("feat(meth): emit Bismark-compatible XR/XG/XM auxiliary tags") added
test/meth/test_bismark_tags.sh, a Layer-3 golden-truth diff against a holodeck-simulated bisulfite BAM. That test pins a different holodeck — SHAd240816c8c2dfbd72bdb3386679d0ad4d4f223e6on branchnh/bisulfite-conversion, built viacargo build --releasefrom a/tmp/holodeck-<sha>cache.Why two install paths today
The
nh/bisulfite-conversionbranch (fg-labs/holodeck#11) adds--methylation-mode em-seqand the Bismark-compatible golden BAM emission thattest_bismark_tags.shexercises. fg-labs/holodeck#11 is open / not merged, so v0.2.1 (the bioconda release pinned by PR #89's pixi env) does not yet expose those flags. There is no released holodeck version that supports both the chr22-parity flow PR #89 uses and the bisulfite-chemistry flow PR #90's Layer-3 test needs, so PR #90 had to roll its own cargo-from-SHA install.Migration to do once upstream lands
main.test/holodeck/pixi.toml, bumpholodeck = "==0.2.1"→ the new version, refreshpixi.lock.test/meth/test_bismark_tags.sh:HOLODECK_SHA,HOLODECK_REPO,HOLODECK_DIR,HOLODECK_BIN, the cargo guard, and the clone/build block."$HOLODECK_BIN" simulate …withpixi run --frozen --manifest-path "$HERE/../holodeck/pixi.toml" -- holodeck simulate …(same pattern PR ci: migrate parity tests from dwgsim/phiX174 to holodeck/chr22 #89's CI uses).command -v cargoSKIP gate —pixibecomes the prerequisite instead.Setup holodeck (pixi)step PR ci: migrate parity tests from dwgsim/phiX174 to holodeck/chr22 #89 added is reusable; the bismark test just needs to run on a matrix row that has the pixi env installed.References