Repository navigation
Fix interproscan ignoring staged database - #13032
Conversation
interproscan.sh cd's into its install dir before reading INTERPROSCAN_CONF, so the relative path resolved to the container's own properties file and the bundled sample data was always used. Use an absolute INTERPROSCAN_CONF, point data.directory at the staged data/ folder and bin.directory at the InterProScan bin folder. Drop Hamap from the database test: the test data ships hamap/2023_05, whereas InterProScan 5.59-91 expects hamap/2021_04. Document the database requirements in meta.yml. Closes #13009 Generated by Claude Opus 5.5 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
erikrikarddaniel
left a comment
There was a problem hiding this comment.
Reviewed with Claude Code.
The fix is correct. I checked it against interproscan:5.59_91.0--hec16e2b_1:
interproscan.shdoescd "$(dirname "$INSTALL_DIR")"before it testsINTERPROSCAN_CONF, so the relative path resolved to the bundled properties file.- The bundled file has
data.directory=dataandbin.directory=bin, and the member binaries live under/usr/local/share/InterProScan/bin. - I ran the module tests on this branch with Docker. All four pass with unchanged snapshots, and the database test's
interproscan.propertiesnow pointsdata.directoryat the staged folder.
One suggestion inline, about making the database test catch a regression. It needs a test-datasets change, so it is fine as a follow-up.
| process { | ||
| withName: INTERPROSCAN { | ||
| ext.args = '--applications Hamap,TIGRFAM,sfld --disable-precalc' | ||
| ext.args = '--applications TIGRFAM,sfld --disable-precalc' |
There was a problem hiding this comment.
This test would still pass with the old relative INTERPROSCAN_CONF. The test database holds the same models as the data bundled in the container (TIGR01027, TIGR01241, TIGR01949 and SFLDS00057), so both code paths give the same output.
The full TIGRFAM 15.0 model TIGR00519 hits l_arginase.faa above its trusted cutoff (hmmsearch --cut_tc, E-value 1.1e-97), and the bundled data does not include it. If TIGR00519 is added to the test database's TIGRFAMs_HMM.LIB, a TIGRFAM line appears only when the staged database is used, and the test then fails if the staged database is ignored again.
There was a problem hiding this comment.
You have a point. Might add another test case like this in the future. For now I will merge and try it out in the pipeline to see if the fix works as expected. Thanks!
interproscan.sh cd's into its install dir before reading INTERPROSCAN_CONF, so the relative path resolved to the container's own properties file and the bundled sample data was always used. Use an absolute INTERPROSCAN_CONF, point data.directory at the staged data/ folder and bin.directory at the InterProScan bin folder. Drop Hamap from the database test: the test data ships hamap/2023_05, whereas InterProScan 5.59-91 expects hamap/2021_04. Document the database requirements in meta.yml. Closes #13009 Generated by Claude Opus 5.5 Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
interproscan.sh cd's into its install dir before reading INTERPROSCAN_CONF, so the relative path resolved to the container's own properties file and the bundled sample data was always used. Use an absolute INTERPROSCAN_CONF, point data.directory at the staged data/ folder and bin.directory at the InterProScan bin folder. Drop Hamap from the database test: the test data ships hamap/2023_05, whereas InterProScan 5.59-91 expects hamap/2021_04. Document the database requirements in meta.yml. Closes #13009 Generated by Claude Opus 5.5 Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Bump mgnifam/generatefamilies to 3.1.0 and add stats output mgnifam 3.1.0 writes <chunk>_stats.json, a MultiQC-ready run summary. Generated by Claude Opus 5.5 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * pigz/compress stub fix (nf-core#13015) .gz files should be properly created * Read eggnogmapper's version from package metadata (nf-core#13016) eggnog-mapper's get_version() runs `git describe --tags` with cwd set to its own installed package directory, and only falls back to __VERSION__ when that fails. Under conda the package is installed inside the pipeline's own checkout, so `emapper.py --version` prints that repository's tag rather than the tool's version: emapper-1.4.1-184-g1b3550e / Expected eggNOG DB version: 5.0.2 / ... The module's grep then reported 1.4.1 (a metatdenovo release tag) instead of 2.1.13. Container installs sit outside any repository, and the biocontainer has no git, so they were unaffected and the drift only showed up in a pipeline's conda CI. importlib.metadata reads the installed distribution's own metadata, which no surrounding repository can influence. The package name is passed through sys.argv to keep the expression free of nested quotes, which nf-core modules lint's main.nf parser does not handle. Claude-Session: https://claude.ai/code/session_01LydWahEnWZexTULEub5bfX Co-authored-by: Claude Opus 5 <noreply@anthropic.com> * checkm2/predict: pass database via CHECKM2DB and use small test DB (nf-core#13019) CheckM2 enforces a hardcoded checksum on databases passed with --database_path, but not on the one set through the CHECKM2DB environment variable. Using the env var allows reduced databases, so the test now uses the small DB from test-datasets instead of downloading the full CheckM2 database. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * gffread: accept gzipped gff and fasta input (nf-core#13024) * gffread: accept gzipped gff and fasta input gffread cannot read gzip: a gzipped gff silently yields no features and a gzipped fasta fails with "sequence lines in a FASTA record must have the same length". Stream a gzipped gff through stdin, and decompress a gzipped fasta to a temporary file, removed afterwards so it does not match the *.fasta output glob. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UvGYU6kuXJeVv9ffdSzdzH * Update modules/nf-core/gffread/main.nf Co-authored-by: Paolo Inglese <26252284+piplus2@users.noreply.github.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Paolo Inglese <26252284+piplus2@users.noreply.github.com> * Bump foldseek/createdb and foldseek/easysearch to 10.941cd33 (nf-core#13027) * Bump foldseek/createdb and foldseek/easysearch to 10.941cd33 - Update conda env and biocontainers to foldseek 10.941cd33 - easysearch: discover the main DB basename via its .lookup file instead of misusing ext.prefix2/meta2.id - Add maintainers and .m8 EDAM ontology to meta.yml - Fix stub test name and regenerate snapshots Generated by Claude Opus 5.5 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * foldseek/easysearch: select DB via meta2.id, fail fast on lookup fallback Use ${db}/${meta2.id} when it exists so pipelines can still pick a database in a multi-DB directory; otherwise fall back to the single .lookup file and exit with an error if zero or several are found. Add a test covering the fallback path. Generated by Claude Opus 5.5 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> * feat: update trna to add all optional outputs (nf-core#13029) * feat: update trna to add all optional outputs * fix: snapshots * Add taxonkit/lca module (nf-core#13023) * Add taxonkit/lca module * Restructure taxonkit/lca input error checking* *Add tests to confirm the module appropriately fails if neither taxids, nor a file containing taxids are passed and if both are passed * Fix precommit format issue in taxonkit/lca * Fix trailing whitespace issue for taxonkit/lca * Remove `MINIMAC4` sites input (nf-core#13022) * Fix miniconda version * Remove sites options from minimac4 * Update test * Add --sites argument as output * Set to optional * Add ontology --------- Co-authored-by: Kevin-Brockers <57921086+Kevin-Brockers@users.noreply.github.com> * New module: xengsort/classify (nf-core#12864) * initiated module template * added xengsort classify module * added xengsort classify tests * updated testing to only check for file existance * trimmed some whitespaces * minor formatting correction * updated testing * classify tests are now run on single thread * actual single thread adjustments * updated snapshot * defining unstable outputs * nf-test single-end graft output is empty and therefore unstable * removed graft from unstable keys * removed params.classify_args from nextflow.config * New module: samtools/trimheader (nf-core#13035) * New module: samtools/trimheader Removes @sq lines for references without reads, re-encoding every record against the smaller header. featureCounts (Subread 2.1.1) cannot read a BAM header over 2 GiB, which a fragmented assembly of about 75M contigs produces. samtools reheader cannot do this, since it leaves record reference ids as they are. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LydWahEnWZexTULEub5bfX * Update modules/nf-core/samtools/trimheader/tests/main.nf.test Co-authored-by: Louis Le Nézet <58640615+LouisLeNezet@users.noreply.github.com> * Key readsMD5 on the bam channel and regenerate the snapshot Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LydWahEnWZexTULEub5bfX --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Louis Le Nézet <58640615+LouisLeNezet@users.noreply.github.com> * Fix interproscan ignoring staged database (nf-core#13032) interproscan.sh cd's into its install dir before reading INTERPROSCAN_CONF, so the relative path resolved to the container's own properties file and the bundled sample data was always used. Use an absolute INTERPROSCAN_CONF, point data.directory at the staged data/ folder and bin.directory at the InterProScan bin folder. Drop Hamap from the database test: the test data ships hamap/2023_05, whereas InterProScan 5.59-91 expects hamap/2021_04. Document the database requirements in meta.yml. Closes nf-core#13009 Generated by Claude Opus 5.5 Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> * Update conda-incubator/setup-miniconda digest to be893c9 * fix(hmmer/eslsfetchindex): use GNU sort for large indexes (nf-core#13041) * fix(hmmer/eslsfetchindex): use GNU sort for large indexes Add Coreutils to the module environment and Wave containers so sort can spill to TMPDIR. Report the Coreutils version and document temporary disk use. Generated by Codex * remove empty config file * feat(rsem,dupradar): make temp-file scratch location overridable via ext.args (nf-core#13043) * feat(rsem,dupradar): make temp-file scratch location overridable via ext.args Both RSEM_CALCULATEEXPRESSION/SENTIEON_RSEMCALCULATEEXPRESSION's --temporary-folder and DUPRADAR's featureCounts tmpDir are currently hardcoded to a path relative to the task's working directory, with no way to point them elsewhere. On some network/object-storage-backed work directories (Fusion, NFS, Lustre), the tools' own scratch I/O can be a meaningful cost, and pipelines that have a genuine local scratch mount available have no way to use it without forking the module. This only adds an opt-in override, defaulting exactly as before: - RSEM: ext.args can now include --temporary-folder, guarded the same way many other nf-core modules guard a default CLI flag. - dupradar: a new tmp_dir option, parsed by the template's existing parse_args() mechanism (same pattern already used for feature_type), forwarded to analyzeDuprates()'s tmpDir, which dupRadar itself passes through to Rsubread::featureCounts. Context: nf-core/rnaseq#1957, nf-core/rnaseq#1962 * fix(dupradar): create tmp_dir if it doesn't already exist featureCounts doesn't create its own tmpDir - the default '.' always already exists so this was never visible before, but a custom tmp_dir override fails outright ("temporary directory is not writable") unless something creates it first. Verified locally: without this, overriding tmp_dir fails; with it, featureCounts picks it up and uses it correctly. * fix(dupradar): don't suppress dir.create warnings for tmp_dir showWarnings=FALSE hid the one diagnostic (dir.create's own warning naming the actual OS-level reason) that would explain why a custom tmp_dir couldn't be created, leaving only featureCounts' own less specific "not writable" error. Only call dir.create when the path doesn't already exist (so the default "." case never touches this at all), then explicitly verify existence and writability with a clear error naming the path. * fix(rsem/calculateexpression): remove unused publishDir causing setup failure The "homo_sapiens - bam" test's alignment.config set a publishDir referencing params.outdir, but the test's `then` block only ever asserts on process.out directly via snapshot() - nothing reads from a published path. That publishDir also applied to RSEM_PREPAREREFERENCE, run in this test's `setup` block before `when.params.outdir` is set, so every run failed with "Access to undefined parameter `outdir`" regardless of anything this PR changes. Checked the other modules sharing this same publishDir/outdir config pattern (bcl2fastq, lima, pycoqc, rseqc/bamstat, rseqc/inferexperiment, spring/decompress, subread/featurecounts) - none combine it with a setup dependency, so this exact conflict hasn't come up before. * fix(rsem/calculateexpression): remove second unused publishDir/outdir instance Same issue as the previous commit, in nextflow.config (used by the "fastq" and "stub" tests) rather than alignment.config. The "stub" test's own when block never set params.outdir at all - it only happened to pass when run together with the other tests in this file, because nf-test carries params state across sequential tests in the same run. CI shards tests individually, which exposed it. Removed the now-pointless params.outdir assignments from "bam" and "fastq" too, since nothing reads a published path in any of these tests' assertions. * Add octopusv submodules - filter + subset (nf-core#13040) * Add octopusv submodules - filter + subset * update tests * address review feedback * Update `bcftools_mpileup` (nf-core#13047) * Update bcftools mpileup * Update meta * Fix test * Fix test * Add parabricks/deepsomatic (nf-core#12797) * Add parabricks/deepsomatic * fix tests, use data with accepted readgroup values * remove hallucinated field --------- Co-authored-by: Friederike Hanssen <Friederike.hanssen@qbic.uni-tuebingen.de> Co-authored-by: Friederike Hanssen <friederike.hanssen@seqera.io> * Bump mgnifam/generatefamilies to 4.0.0 mgnifam 4.0.0 renames the MultiQC summary to <chunk>_mgnifam_stats.json and reorders/renames the metadata CSV columns. Point the documentation link to the new docs site. Generated by Claude Opus 5.5 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Matthieu Muffato <mm49@sanger.ac.uk> Co-authored-by: Daniel Lundin <erik.rikard.daniel@gmail.com> Co-authored-by: Diego Alvarez S. <dialvarezs@gmail.com> Co-authored-by: Paolo Inglese <26252284+piplus2@users.noreply.github.com> Co-authored-by: Jim Downie <19718667+prototaxites@users.noreply.github.com> Co-authored-by: jcbioinformatics <60235622+jcbioinformatics@users.noreply.github.com> Co-authored-by: Louis Le Nézet <58640615+LouisLeNezet@users.noreply.github.com> Co-authored-by: Kevin-Brockers <57921086+Kevin-Brockers@users.noreply.github.com> Co-authored-by: Leon Hornich <82643611+LeonHornich@users.noreply.github.com> Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> Co-authored-by: Jonathan Manning <jonathan.manning@seqera.io> Co-authored-by: Manas <manas.sehgal@dkfz.de> Co-authored-by: Simon Pearce <24893913+SPPearce@users.noreply.github.com> Co-authored-by: Friederike Hanssen <Friederike.hanssen@qbic.uni-tuebingen.de> Co-authored-by: Friederike Hanssen <friederike.hanssen@seqera.io>
* Bump mgnifam/updatefamilies to 3.1.0 and add stats output mgnifam 3.1.0 writes <chunk>_updated_stats.json, a MultiQC-ready run summary. Generated by Claude Opus 5.5 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * pigz/compress stub fix (nf-core#13015) .gz files should be properly created * Read eggnogmapper's version from package metadata (nf-core#13016) eggnog-mapper's get_version() runs `git describe --tags` with cwd set to its own installed package directory, and only falls back to __VERSION__ when that fails. Under conda the package is installed inside the pipeline's own checkout, so `emapper.py --version` prints that repository's tag rather than the tool's version: emapper-1.4.1-184-g1b3550e / Expected eggNOG DB version: 5.0.2 / ... The module's grep then reported 1.4.1 (a metatdenovo release tag) instead of 2.1.13. Container installs sit outside any repository, and the biocontainer has no git, so they were unaffected and the drift only showed up in a pipeline's conda CI. importlib.metadata reads the installed distribution's own metadata, which no surrounding repository can influence. The package name is passed through sys.argv to keep the expression free of nested quotes, which nf-core modules lint's main.nf parser does not handle. Claude-Session: https://claude.ai/code/session_01LydWahEnWZexTULEub5bfX Co-authored-by: Claude Opus 5 <noreply@anthropic.com> * checkm2/predict: pass database via CHECKM2DB and use small test DB (nf-core#13019) CheckM2 enforces a hardcoded checksum on databases passed with --database_path, but not on the one set through the CHECKM2DB environment variable. Using the env var allows reduced databases, so the test now uses the small DB from test-datasets instead of downloading the full CheckM2 database. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * gffread: accept gzipped gff and fasta input (nf-core#13024) * gffread: accept gzipped gff and fasta input gffread cannot read gzip: a gzipped gff silently yields no features and a gzipped fasta fails with "sequence lines in a FASTA record must have the same length". Stream a gzipped gff through stdin, and decompress a gzipped fasta to a temporary file, removed afterwards so it does not match the *.fasta output glob. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UvGYU6kuXJeVv9ffdSzdzH * Update modules/nf-core/gffread/main.nf Co-authored-by: Paolo Inglese <26252284+piplus2@users.noreply.github.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Paolo Inglese <26252284+piplus2@users.noreply.github.com> * Bump foldseek/createdb and foldseek/easysearch to 10.941cd33 (nf-core#13027) * Bump foldseek/createdb and foldseek/easysearch to 10.941cd33 - Update conda env and biocontainers to foldseek 10.941cd33 - easysearch: discover the main DB basename via its .lookup file instead of misusing ext.prefix2/meta2.id - Add maintainers and .m8 EDAM ontology to meta.yml - Fix stub test name and regenerate snapshots Generated by Claude Opus 5.5 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * foldseek/easysearch: select DB via meta2.id, fail fast on lookup fallback Use ${db}/${meta2.id} when it exists so pipelines can still pick a database in a multi-DB directory; otherwise fall back to the single .lookup file and exit with an error if zero or several are found. Add a test covering the fallback path. Generated by Claude Opus 5.5 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> * feat: update trna to add all optional outputs (nf-core#13029) * feat: update trna to add all optional outputs * fix: snapshots * Add taxonkit/lca module (nf-core#13023) * Add taxonkit/lca module * Restructure taxonkit/lca input error checking* *Add tests to confirm the module appropriately fails if neither taxids, nor a file containing taxids are passed and if both are passed * Fix precommit format issue in taxonkit/lca * Fix trailing whitespace issue for taxonkit/lca * Remove `MINIMAC4` sites input (nf-core#13022) * Fix miniconda version * Remove sites options from minimac4 * Update test * Add --sites argument as output * Set to optional * Add ontology --------- Co-authored-by: Kevin-Brockers <57921086+Kevin-Brockers@users.noreply.github.com> * New module: xengsort/classify (nf-core#12864) * initiated module template * added xengsort classify module * added xengsort classify tests * updated testing to only check for file existance * trimmed some whitespaces * minor formatting correction * updated testing * classify tests are now run on single thread * actual single thread adjustments * updated snapshot * defining unstable outputs * nf-test single-end graft output is empty and therefore unstable * removed graft from unstable keys * removed params.classify_args from nextflow.config * New module: samtools/trimheader (nf-core#13035) * New module: samtools/trimheader Removes @sq lines for references without reads, re-encoding every record against the smaller header. featureCounts (Subread 2.1.1) cannot read a BAM header over 2 GiB, which a fragmented assembly of about 75M contigs produces. samtools reheader cannot do this, since it leaves record reference ids as they are. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LydWahEnWZexTULEub5bfX * Update modules/nf-core/samtools/trimheader/tests/main.nf.test Co-authored-by: Louis Le Nézet <58640615+LouisLeNezet@users.noreply.github.com> * Key readsMD5 on the bam channel and regenerate the snapshot Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LydWahEnWZexTULEub5bfX --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Louis Le Nézet <58640615+LouisLeNezet@users.noreply.github.com> * Fix interproscan ignoring staged database (nf-core#13032) interproscan.sh cd's into its install dir before reading INTERPROSCAN_CONF, so the relative path resolved to the container's own properties file and the bundled sample data was always used. Use an absolute INTERPROSCAN_CONF, point data.directory at the staged data/ folder and bin.directory at the InterProScan bin folder. Drop Hamap from the database test: the test data ships hamap/2023_05, whereas InterProScan 5.59-91 expects hamap/2021_04. Document the database requirements in meta.yml. Closes nf-core#13009 Generated by Claude Opus 5.5 Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> * Update conda-incubator/setup-miniconda digest to be893c9 * fix(hmmer/eslsfetchindex): use GNU sort for large indexes (nf-core#13041) * fix(hmmer/eslsfetchindex): use GNU sort for large indexes Add Coreutils to the module environment and Wave containers so sort can spill to TMPDIR. Report the Coreutils version and document temporary disk use. Generated by Codex * remove empty config file * feat(rsem,dupradar): make temp-file scratch location overridable via ext.args (nf-core#13043) * feat(rsem,dupradar): make temp-file scratch location overridable via ext.args Both RSEM_CALCULATEEXPRESSION/SENTIEON_RSEMCALCULATEEXPRESSION's --temporary-folder and DUPRADAR's featureCounts tmpDir are currently hardcoded to a path relative to the task's working directory, with no way to point them elsewhere. On some network/object-storage-backed work directories (Fusion, NFS, Lustre), the tools' own scratch I/O can be a meaningful cost, and pipelines that have a genuine local scratch mount available have no way to use it without forking the module. This only adds an opt-in override, defaulting exactly as before: - RSEM: ext.args can now include --temporary-folder, guarded the same way many other nf-core modules guard a default CLI flag. - dupradar: a new tmp_dir option, parsed by the template's existing parse_args() mechanism (same pattern already used for feature_type), forwarded to analyzeDuprates()'s tmpDir, which dupRadar itself passes through to Rsubread::featureCounts. Context: nf-core/rnaseq#1957, nf-core/rnaseq#1962 * fix(dupradar): create tmp_dir if it doesn't already exist featureCounts doesn't create its own tmpDir - the default '.' always already exists so this was never visible before, but a custom tmp_dir override fails outright ("temporary directory is not writable") unless something creates it first. Verified locally: without this, overriding tmp_dir fails; with it, featureCounts picks it up and uses it correctly. * fix(dupradar): don't suppress dir.create warnings for tmp_dir showWarnings=FALSE hid the one diagnostic (dir.create's own warning naming the actual OS-level reason) that would explain why a custom tmp_dir couldn't be created, leaving only featureCounts' own less specific "not writable" error. Only call dir.create when the path doesn't already exist (so the default "." case never touches this at all), then explicitly verify existence and writability with a clear error naming the path. * fix(rsem/calculateexpression): remove unused publishDir causing setup failure The "homo_sapiens - bam" test's alignment.config set a publishDir referencing params.outdir, but the test's `then` block only ever asserts on process.out directly via snapshot() - nothing reads from a published path. That publishDir also applied to RSEM_PREPAREREFERENCE, run in this test's `setup` block before `when.params.outdir` is set, so every run failed with "Access to undefined parameter `outdir`" regardless of anything this PR changes. Checked the other modules sharing this same publishDir/outdir config pattern (bcl2fastq, lima, pycoqc, rseqc/bamstat, rseqc/inferexperiment, spring/decompress, subread/featurecounts) - none combine it with a setup dependency, so this exact conflict hasn't come up before. * fix(rsem/calculateexpression): remove second unused publishDir/outdir instance Same issue as the previous commit, in nextflow.config (used by the "fastq" and "stub" tests) rather than alignment.config. The "stub" test's own when block never set params.outdir at all - it only happened to pass when run together with the other tests in this file, because nf-test carries params state across sequential tests in the same run. CI shards tests individually, which exposed it. Removed the now-pointless params.outdir assignments from "bam" and "fastq" too, since nothing reads a published path in any of these tests' assertions. * Add octopusv submodules - filter + subset (nf-core#13040) * Add octopusv submodules - filter + subset * update tests * address review feedback * Update `bcftools_mpileup` (nf-core#13047) * Update bcftools mpileup * Update meta * Fix test * Fix test * Add parabricks/deepsomatic (nf-core#12797) * Add parabricks/deepsomatic * fix tests, use data with accepted readgroup values * remove hallucinated field --------- Co-authored-by: Friederike Hanssen <Friederike.hanssen@qbic.uni-tuebingen.de> Co-authored-by: Friederike Hanssen <friederike.hanssen@seqera.io> * Bump mgnifam/updatefamilies to 4.0.0 mgnifam 4.0.0 renames the MultiQC summary to <chunk>_updated_mgnifam_stats.json, reorders/renames the metadata CSV columns and leaves `converged` empty under --skip_refine. Point the documentation link to the new docs site. Generated by Claude Opus 5.5 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Matthieu Muffato <mm49@sanger.ac.uk> Co-authored-by: Daniel Lundin <erik.rikard.daniel@gmail.com> Co-authored-by: Diego Alvarez S. <dialvarezs@gmail.com> Co-authored-by: Paolo Inglese <26252284+piplus2@users.noreply.github.com> Co-authored-by: Jim Downie <19718667+prototaxites@users.noreply.github.com> Co-authored-by: jcbioinformatics <60235622+jcbioinformatics@users.noreply.github.com> Co-authored-by: Louis Le Nézet <58640615+LouisLeNezet@users.noreply.github.com> Co-authored-by: Kevin-Brockers <57921086+Kevin-Brockers@users.noreply.github.com> Co-authored-by: Leon Hornich <82643611+LeonHornich@users.noreply.github.com> Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com> Co-authored-by: Jonathan Manning <jonathan.manning@seqera.io> Co-authored-by: Manas <manas.sehgal@dkfz.de> Co-authored-by: Simon Pearce <24893913+SPPearce@users.noreply.github.com> Co-authored-by: Friederike Hanssen <Friederike.hanssen@qbic.uni-tuebingen.de> Co-authored-by: Friederike Hanssen <friederike.hanssen@seqera.io>
interproscanignored the database passed asinterproscan_databaseand always used the sample data bundled in the container (see #13009).interproscan.shchanges into its install directory before it readsINTERPROSCAN_CONF. The module exported that path as relative, so InterProScan read the container's owninterproscan.properties, with relativedata/bindirectories. Thesedalso setbin.directoryto/usr/local/binrather than/usr/local/share/InterProScan/bin, and it never rewrotedata.directory.Changes:
INTERPROSCAN_CONFis now an absolute path,data.directorypoints to the stageddata/folder, andbin.directorypoints to the InterProScanbin/folder.hamap/2023_05, but 5.59-91 expectshamap/2021_04. The old test passed only because it silently used the bundled sample data. With the fix, Hamap fails withFile existence/permissions problem in trying to open HMM file .../data/hamap/2021_04/hamap.hmm.lib.meta.ymldocuments thatinterproscan_databasemust be thedata/folder of a matching release, with HMMs pressed beforehand.Snapshots are unchanged.
Generated by Claude Opus 5.5
PR checklist
Closes #13009
topic: versions- See version_topicslabelnf-core modules test <MODULE> --profile dockernf-core modules test <MODULE> --profile singularitynf-core modules test <MODULE> --profile conda🤖 Generated with Claude Code