feat(#11): consolidate UCR HPCC institutional profile

conf/modules.config

@@ -10,10 +10,6 @@
 
 process {
 
-    withName: 'ASM_STATS' {
-        // storeDir handles persistence; no publishDir needed.
-    }
-
     withName: 'ANTISMASH_RUN' {
         publishDir = [
             path:      { params.target },

conf/profile_annotate.config

@@ -118,130 +118,120 @@ params {
 workDir = "${launchDir}/work/annotate"
 cleanup = false
 
+// queue names, clusterOptions, and GPU options are UCR HPCC SLURM-specific —
+// they live in conf/provision_ucr_hpcc.config, not here. This block contains
+// only hardware resource requirements (cpus, memory, time) that are site-agnostic.
 process {
 
     withName: 'SETUP_TAXONDB' {
-        queue  = 'short'; cpus = 1; memory = '4 GB'; time = '1h'
+        cpus = 1; memory = '4 GB'; time = '1h'
     }
 
     withName: 'SETUP_FUNANNOTATE_DB' {
-        // Network-bound multi-GB download of BUSCO + all funannotate DBs; needs a
-        // long-running queue (short caps at 2h). Runs once, then storeDir-skipped.
-        queue  = 'epyc'; cpus = 2; memory = '8 GB'; time = '12h'
+        // Network-bound multi-GB download of BUSCO + all funannotate DBs.
+        // Runs once per project, then storeDir-skipped.
+        cpus = 2; memory = '8 GB'; time = '12h'
     }
 
     withName: 'SETUP_AUGUSTUS_CONFIG' {
-        // Local filesystem copy of the installed augustus config; quick and runs once,
-        // then storeDir-skipped.
-        queue  = 'short'; cpus = 1; memory = '4 GB'; time = '30m'
+        // Local filesystem copy of the installed augustus config; quick, runs once.
+        cpus = 1; memory = '4 GB'; time = '30m'
     }
 
     withName: '.*:GENOME_CLEAN' {
         // Full FCS-GX purge needs highmem (stages ~470 GB gx db into /dev/shm);
         // with --skip_fcs it's a light decompress + length-filter.
-        queue  = params.skip_fcs ? 'short'   : 'highmem'
-        cpus   = params.skip_fcs ? 2         : 16
-        memory = params.skip_fcs ? '8 GB'    : '500 GB'
-        time   = params.skip_fcs ? '1h'      : '6h'
+        cpus   = params.skip_fcs ? 2      : 16
+        memory = params.skip_fcs ? '8 GB' : '500 GB'
+        time   = params.skip_fcs ? '1h'   : '6h'
     }
 
     withName: '.*:GENOME_CLEAN_BATCH' {
-        // Stages the ~470 GB FCS-GX db into /dev/shm once, then cleans up to
-        // clean_batch_size genomes sequentially — needs highmem and a long walltime.
-        queue  = 'highmem'; cpus = 16; memory = '500 GB'; time = '7d'
+        // Stages the ~470 GB FCS-GX db into /dev/shm once, then cleans
+        // clean_batch_size genomes sequentially — needs highmem + long walltime.
+        cpus = 16; memory = '500 GB'; time = '7d'
     }
 
     withName: '.*:MASKREPEAT_TANTAN_RUN' {
-        queue  = 'short'; cpus = 8; memory = '16 GB'; time = '2h'
+        cpus = 8; memory = '16 GB'; time = '2h'
     }
 
     // EarlGrey TE-discovery path (run_earlgrey=true).
-    // Resource tuning inherits from profile_earlgrey.config's withLabel blocks
-    // (earlgrey / repeatmask / deliver); these withName overrides only set queue.
     withName: '.*:EARLGREY_BUILD_LIB' {
-        queue = 'epyc'; cpus = 32; memory = '64 GB'; time = '7d'
+        cpus = 32; memory = '64 GB'; time = '7d'
     }
     withName: '.*:REPEATMASK_STRAIN' {
-        queue = 'short'; cpus = 32; memory = '32 GB'; time = '8h'
+        cpus = 32; memory = '32 GB'; time = '8h'
     }
     withName: '.*:DELIVER_MASK' {
-        queue = 'short'; cpus = 1; memory = '2 GB'; time = '30m'
+        cpus = 1; memory = '2 GB'; time = '30m'
     }
 
-    // SRA_QUERY (standalone, not in subworkflow) — matches bare name
+    // SRA_QUERY standalone (not in subworkflow) — matches bare name.
     withName: 'SRA_QUERY' {
-        queue = 'short'; cpus = 2; memory = '4 GB'; time = '60m'
+        cpus = 2; memory = '4 GB'; time = '60m'
         maxForks = 10; errorStrategy = 'retry'; maxRetries = 2
     }
     withName: '.*:SRA_QUERY_BATCH' {
-        queue = 'short'; cpus = 1; memory = '4 GB'; time = '2h'
+        cpus = 1; memory = '4 GB'; time = '2h'
         maxForks = 10; errorStrategy = 'retry'; maxRetries = 2
     }
     withName: '.*:COLLECT_SRA_QUERY' {
-        queue = 'short'; cpus = 1; memory = '2 GB'; time = '20m'
+        cpus = 1; memory = '2 GB'; time = '20m'
     }
     withName: '.*:WRITE_EMPTY_READS' {
-        queue = 'short'; cpus = 1; memory = '1 GB'; time = '5m'
+        cpus = 1; memory = '1 GB'; time = '5m'
     }
 
     withName: '.*:SRA_FETCH' {
         errorStrategy = 'retry'; maxRetries = 3
-        queue  = { task.attempt == 1 ? 'short' : 'epyc' }
-        memory = { task.attempt == 1 ? '48 GB' : '192 GB' }
+        memory = { task.attempt == 1 ? '48 GB'  : '192 GB' }
         time   = { task.attempt == 1 ? '2h' : task.attempt == 2 ? '8h' : task.attempt == 3 ? '24h' : '48h' }
         cpus   = { task.attempt == 1 ? 24 : 32 }
     }
     withName: '.*:SRA_FETCH_SE' {
         errorStrategy = 'retry'; maxRetries = 3
-        queue  = { task.attempt == 1 ? 'short' : 'epyc' }
-        memory = { task.attempt == 1 ? '48 GB' : '192 GB' }
+        memory = { task.attempt == 1 ? '48 GB'  : '192 GB' }
         time   = { task.attempt == 1 ? '2h' : task.attempt == 2 ? '8h' : '24h' }
         cpus   = { task.attempt == 1 ? 24 : 32 }
     }
 
     withName: '.*:RNASEQ_PREPARE' {
-        queue  = 'epyc'; cpus = 16
+        cpus = 16
         memory = { task.attempt == 1 ? '52 GB' : (48 + (task.attempt - 1) * 48) + ' GB' }
-        time   = '120h'; maxRetries = 3
+        time = '120h'; maxRetries = 3
     }
 
     withName: '.*:FUNANNOTATE_TRAIN' {
-        maxRetries     = 3
-        queue          = { task.attempt <= 2 ? 'epyc' : 'highmem' }
-        memory         = { task.attempt == 1 ? '48 GB' : task.attempt == 2 ? '192 GB' : '256 GB' }
-        time           = { task.attempt == 1 ? '72h' : '120h' }
-        cpus           = { task.attempt == 1 ? 8 : 24 }
-        clusterOptions = '-N 1 -n 1 --job-name=nf-TRAIN'
+        maxRetries = 3
+        memory     = { task.attempt == 1 ? '48 GB' : task.attempt == 2 ? '192 GB' : '256 GB' }
+        time       = { task.attempt == 1 ? '72h' : '120h' }
+        cpus       = { task.attempt == 1 ? 8 : 24 }
     }
 
     withName: '.*:FUNANNOTATE_PREDICT' {
-        queue          = 'epyc'
-        memory         = { task.attempt == 1 ? '32 GB' : '64 GB' }
-        time           = '72h'
-        cpus           = { task.attempt == 1 ? 8 : 16 }
-        clusterOptions = '-N 1 -n 1 --job-name=nf-PREDICT'
+        memory = { task.attempt == 1 ? '32 GB' : '64 GB' }
+        time   = '72h'
+        cpus   = { task.attempt == 1 ? 8 : 16 }
     }
 
     withName: '.*:ANTISMASH_RUN' {
-        queue = 'epyc'; cpus = 8; memory = '24 GB'; time = '8h'
+        cpus = 8; memory = '24 GB'; time = '8h'
     }
 
     withName: '.*:INTERPROSCAN_RUN' {
-        queue = 'epyc'; cpus = 8; memory = '32 GB'; time = '60h'
+        cpus = 8; memory = '32 GB'; time = '60h'
     }
 
     withName: '.*:SIGNALP_RUN' {
-        // signalp6-gpu: GPU partition (gpu allows >2h; short_gpu if always ≤2h).
-        queue          = 'gpu'
-        cpus           = 16; memory = '64 GB'; time = '8h'
-        clusterOptions = '-N 1 -n 1 --gres=gpu:1'
+        cpus = 16; memory = '64 GB'; time = '8h'
     }
 
     withName: '.*:FUNANNOTATE_ANNOTATE' {
-        queue = 'epyc'; cpus = 16; memory = '64 GB'; time = '24h'
+        cpus = 16; memory = '64 GB'; time = '24h'
     }
     withName: '.*:FUNANNOTATE_UPDATE' {
-        queue = 'epyc'; cpus = 16; memory = '64 GB'; time = '48h'
+        cpus = 16; memory = '64 GB'; time = '48h'
     }
 }
 

conf/profile_earlgrey.config

@@ -12,46 +12,27 @@ workDir = "${launchDir}/work/earlgrey"
 
 process {
 
-    // De-novo TE discovery + curation on the representative genome — very heavy.
-    // Only long-queue partitions (epyc/stajichlab) can host a multi-day EarlGrey run.
-    withLabel: 'earlgrey' {
-        cpus           = 32
-        memory         = '64 GB'
-        time           = '7.day'
-        clusterOptions = '-N 1 -n 1'
-        beforeScript = """
-            source /etc/profile.d/modules.sh 2>/dev/null || true
-	    module load earlgrey
-	"""
+    // Resources only — tool provisioning (beforeScript) and SLURM queue/clusterOptions
+    // are UCR-specific and live in conf/provision_ucr_hpcc.config.
 
+    // De-novo TE discovery + curation — very heavy, multi-day run.
+    withLabel: 'earlgrey' {
+        cpus   = 32; memory = '64 GB'; time = '7.day'
     }
 
-    // Apply the curated library to a conspecific strain — moderate.
+    // Apply curated TE library to a conspecific strain with RepeatMasker — moderate.
     withLabel: 'repeatmask' {
-        cpus           = 32
-        memory         = '32 GB'
-        time           = '8.h'
-        clusterOptions = '-N 1 -n 1'
+        cpus   = 32; memory = '32 GB'; time = '8.h'
     }
 
     // Copy finished masks into input_clean_genomes — trivial I/O.
     withLabel: 'deliver' {
-        cpus   = 1
-        memory = '2 GB'
-        time   = '30 min'
-        queue  = 'short'
+        cpus   = 1; memory = '2 GB'; time = '30 min'
     }
 
     // Representative selection — light, pure-Python (stdlib only).
     withLabel: 'select' {
-        cpus   = 1
-        memory = '4 GB'
-        time   = '30 min'
-        queue  = 'short'
-        beforeScript = """
-            source /etc/profile.d/modules.sh 2>/dev/null || true
-            module load biopython
-        """
+        cpus   = 1; memory = '4 GB'; time = '30 min'
     }
 }
 

conf/provision_ucr_hpcc.config

@@ -1,24 +1,29 @@
 /*
- * Provisioning: UCR HPCC Lmod modules — INSTITUTIONAL profile (-profile ucr_hpcc).
+ * Provisioning: UCR HPCC — INSTITUTIONAL profile (-profile ucr_hpcc).
  *
- * This is a site-specific profile: the module names (funannotate/dev-1.8.18,
- * AAFTF, taxonkit, ...) and the conda bootstrap below exist on the UCR HPCC and
- * will NOT resolve elsewhere. For a portable run use a container/env provisioning
- * profile instead (-profile <...>,singularity or ,pixi). New sites should copy
- * this file to conf/provision_<site>.config and register a matching profile in
- * nextflow.config — the nf-core/configs institutional-profile model.
+ * Encodes two UCR-specific concerns:
  *
- * Each process label gets a `beforeScript` that loads the tools it needs. The
- * beforeScript and the process `script:` are concatenated into one job script,
- * so the module environment persists into the script body. Process scripts
- * therefore contain NO `module load` lines themselves.
+ *   1. Tool provisioning (withLabel beforeScript) — Lmod module loads.
+ *      Module names and conda bootstrap exist on UCR HPCC; will not resolve
+ *      elsewhere. Process scripts contain NO `module load` lines; all tool
+ *      setup happens in beforeScript so the environment is set before the
+ *      script body runs.
  *
- * Every beforeScript first sources modules.sh so `module` is defined inside the
- * batch job (login-shell init does not always run on compute nodes).
+ *   2. SLURM partition routing (withName queue / clusterOptions) — UCR HPCC
+ *      partition names (short, epyc, highmem, gpu). These are separated from
+ *      profile_annotate.config so a non-UCR SLURM user can use
+ *      `-profile annotate,slurm` without a queue= override and just get the
+ *      default partition.
+ *
+ * To add a new site: copy this file to conf/provision_<site>.config, replace
+ * the module names with site equivalents, adjust queue names, and register a
+ * matching profile in nextflow.config. See docs/adding_a_site.md.
  */
 
 process {
 
+    // ── Tool provisioning (Lmod module loads) ─────────────────────────────────
+
     withLabel: 'funannotate' {
         beforeScript = '''
             source /etc/profile.d/modules.sh 2>/dev/null || true
@@ -37,9 +42,6 @@ process {
         beforeScript = '''
             source /etc/profile.d/modules.sh 2>/dev/null || true
             module load miniconda3
-            # Clear stale CONDA_* inherited from the login shell (bash -l), then
-            # source conda.sh directly — `conda shell.bash hook` crashes when the
-            # inherited state is inconsistent (CONDA_SHLVL set but CONDA_PREFIX empty).
             unset CONDA_SHLVL CONDA_PREFIX CONDA_PREFIX_1 CONDA_DEFAULT_ENV CONDA_PROMPT_MODIFIER
             source "${CONDA_ROOT:-$(conda info --base)}/etc/profile.d/conda.sh"
             module load taxonkit
@@ -71,9 +73,6 @@ process {
         beforeScript = '''
             source /etc/profile.d/modules.sh 2>/dev/null || true
             module load miniconda3
-            # Clear stale CONDA_* inherited from the login shell (bash -l), then
-            # source conda.sh directly — `conda shell.bash hook` crashes when the
-            # inherited state is inconsistent (CONDA_SHLVL set but CONDA_PREFIX empty).
             unset CONDA_SHLVL CONDA_PREFIX CONDA_PREFIX_1 CONDA_DEFAULT_ENV CONDA_PROMPT_MODIFIER
             source "${CONDA_ROOT:-$(conda info --base)}/etc/profile.d/conda.sh"
             module load antismash
@@ -94,9 +93,71 @@ process {
         '''
     }
 
-    // 'setup' label (SETUP_TAXONDB, COLLECT_SRA_QUERY, WRITE_EMPTY_READS) needs
-    // only base system tools (wget/tar/awk) — no module load required.
-    withLabel: 'setup' {
-        beforeScript = ':'
+    withLabel: 'earlgrey' {
+        beforeScript = '''
+            source /etc/profile.d/modules.sh 2>/dev/null || true
+            module load earlgrey
+        '''
+    }
+
+    withLabel: 'repeatmask' {
+        beforeScript = '''
+            source /etc/profile.d/modules.sh 2>/dev/null || true
+            module load RepeatMasker
+        '''
+    }
+
+    withLabel: 'select' {
+        beforeScript = '''
+            source /etc/profile.d/modules.sh 2>/dev/null || true
+            module load biopython
+        '''
+    }
+
+    // 'setup', 'deliver', 'process_single': base system tools only — no module load.
+    withLabel: 'setup'   { beforeScript = ':' }
+    withLabel: 'deliver' { beforeScript = ':' }
+
+    // ── UCR HPCC SLURM partition routing ──────────────────────────────────────
+    // Queue names are UCR-specific (short/epyc/highmem/gpu). Non-UCR SLURM users
+    // omit this profile and rely on their scheduler's default partition.
+
+    withName: 'SETUP_TAXONDB'         { queue = 'short' }
+    withName: 'SETUP_FUNANNOTATE_DB'  { queue = 'epyc'  }
+    withName: 'SETUP_AUGUSTUS_CONFIG' { queue = 'short' }
+
+    withName: '.*:GENOME_CLEAN'       { queue = params.skip_fcs ? 'short' : 'highmem' }
+    withName: '.*:GENOME_CLEAN_BATCH' { queue = 'highmem' }
+
+    withName: '.*:MASKREPEAT_TANTAN_RUN' { queue = 'short' }
+    withName: '.*:EARLGREY_BUILD_LIB'   { queue = 'epyc'  }
+    withName: '.*:REPEATMASK_STRAIN'    { queue = 'short' }
+    withName: '.*:DELIVER_MASK'         { queue = 'short' }
+
+    withName: 'SRA_QUERY'             { queue = 'short' }
+    withName: '.*:SRA_QUERY_BATCH'    { queue = 'short' }
+    withName: '.*:COLLECT_SRA_QUERY'  { queue = 'short' }
+    withName: '.*:WRITE_EMPTY_READS'  { queue = 'short' }
+    withName: '.*:SRA_FETCH'          { queue = { task.attempt == 1 ? 'short' : 'epyc' } }
+    withName: '.*:SRA_FETCH_SE'       { queue = { task.attempt == 1 ? 'short' : 'epyc' } }
+
+    withName: '.*:RNASEQ_PREPARE'     { queue = 'epyc' }
+
+    withName: '.*:FUNANNOTATE_TRAIN' {
+        queue          = { task.attempt <= 2 ? 'epyc' : 'highmem' }
+        clusterOptions = '-N 1 -n 1 --job-name=nf-TRAIN'
+    }
+    withName: '.*:FUNANNOTATE_PREDICT' {
+        queue          = 'epyc'
+        clusterOptions = '-N 1 -n 1 --job-name=nf-PREDICT'
+    }
+
+    withName: '.*:ANTISMASH_RUN'      { queue = 'epyc'  }
+    withName: '.*:INTERPROSCAN_RUN'   { queue = 'epyc'  }
+    withName: '.*:SIGNALP_RUN' {
+        queue          = 'gpu'
+        clusterOptions = '-N 1 -n 1 --gres=gpu:1'
     }
+    withName: '.*:FUNANNOTATE_ANNOTATE' { queue = 'epyc' }
+    withName: '.*:FUNANNOTATE_UPDATE'   { queue = 'epyc' }
 }