Update kmcp modules (#3680)

* Update kmcp profile * Update kmcp modules * Remove one line * Update test.yml for kmcp/profile * test.yml * Fix linting * Update to kmcp:0.9.3 * kmcp=0.9.3 * Version 0.9.1 * Update test.yml for kmcp/search * Prettier * Fix linting * Update test.yml for kmcp/profile * Update test.yml * Test without stub * Prettier * Add stub to test.yml * Test gzip in search * Try to fix gzip * Try to fix gzip * Test gzip
nf-core · Aug 1, 2023 · e198734 · e198734
1 parent 5b116ec
commit e198734
Show file tree

Hide file tree

Showing 12 changed files with 34 additions and 42 deletions.
diff --git a/modules/nf-core/kmcp/profile/main.nf b/modules/nf-core/kmcp/profile/main.nf
@@ -9,8 +9,7 @@ process KMCP_PROFILE {
 
     input:
     tuple val(meta), path(search_results)
-    path taxdump
-    path taxid
+    path (db)
     val mode
 
     output:
@@ -24,11 +23,13 @@ process KMCP_PROFILE {
     def args = task.ext.args ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
     """
+    taxid=`find -L ${db} -name "*map"`
+    taxdump=`find -L ${db}/*/ -type d  -not -name "R001"`
     kmcp \\
         profile \\
         $args \\
-        -X $taxdump \\
-        -T $taxid \\
+        -X \$taxdump \\
+        -T \$taxid \\
         -m $mode \\
         -j $task.cpus \\
         -o ${prefix}.profile \\

diff --git a/modules/nf-core/kmcp/profile/meta.yml b/modules/nf-core/kmcp/profile/meta.yml
@@ -23,18 +23,13 @@ input:
       description: |
         Groovy Map containing sample information
         e.g. [ id:'test', single_end:false ]
+  - db:
+      type: directory
+      description: Database directory containing taxdump files and taxid file
   - search_results:
       type: file
       description: Gzipped file output from kmcp search module
       pattern: "*.gz"
-  - taxdump:
-      type: directory
-      description: Directory of NCBI taxonomy dump files - names.dmp, nodes.dmp
-      pattern: "*"
-  - taxid:
-      type: file
-      description: Tabular two-column file(s) mapping reference IDs to TaxIds.
-      pattern: "*.{csv,tsv}"
   - mode:
       type: integer
       description: Profiling mode.

diff --git a/modules/nf-core/kmcp/search/main.nf b/modules/nf-core/kmcp/search/main.nf
@@ -40,8 +40,7 @@ process KMCP_SEARCH {
     def args = task.ext.args ?: ''
     prefix = task.ext.prefix ?: "${meta.id}"
     """
-    touch ${prefix}
-    gzip ${prefix}
+    touch ${prefix}.gz
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":

diff --git a/modules/nf-core/kmcp/search/meta.yml b/modules/nf-core/kmcp/search/meta.yml
@@ -1,4 +1,4 @@
-name: "kmcp_compute"
+name: "kmcp_search"
 description: Search sequences against database
 keywords:
   - metagenomics

diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
@@ -2045,10 +2045,18 @@ kmcp/index:
   - modules/nf-core/kmcp/index/**
   - tests/modules/nf-core/kmcp/index/**
 
+kmcp/merge:
+  - modules/nf-core/kmcp/merge/**
+  - tests/modules/nf-core/kmcp/merge/**
+
 kmcp/profile:
   - modules/nf-core/kmcp/profile/**
   - tests/modules/nf-core/kmcp/profile/**
 
+kmcp/search:
+  - modules/nf-core/kmcp/search/**
+  - tests/modules/nf-core/kmcp/search/**
+
 kofamscan:
   - modules/nf-core/kofamscan/**
   - tests/modules/nf-core/kofamscan/**

diff --git a/tests/modules/nf-core/kmcp/compute/main.nf b/tests/modules/nf-core/kmcp/compute/main.nf
@@ -19,7 +19,7 @@ workflow test_kmcp_compute_directory {
 
 input = UNTAR ( [
             [ id:'test' ],
-            file("https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/kmcp.tar.gz", checkIfExists: true)
+            file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/kmcp/kmcp_profile.tar.gz", checkIfExists: true)
         ]).untar
 
 KMCP_COMPUTE ( input )

diff --git a/tests/modules/nf-core/kmcp/compute/test.yml b/tests/modules/nf-core/kmcp/compute/test.yml
@@ -1,8 +1,8 @@
 - name: kmcp compute test_kmcp_compute
   command: nextflow run ./tests/modules/nf-core/kmcp/compute -entry test_kmcp_compute -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/kmcp/compute/nextflow.config
   tags:
-    - kmcp
     - kmcp/compute
+    - kmcp
   files:
     - path: output/kmcp/test_/_info.txt
       contains: ["#path	name	chunkIdx	idxNum	genomeSize	kmers"]
@@ -27,5 +27,4 @@
     - path: output/kmcp/test_/_info.txt
       contains: ["#path	name	chunkIdx	idxNum	genomeSize	kmers"]
     - path: output/kmcp/test_/genome.fasta.unik
-    - path: output/kmcp/test_/transcriptome.fasta.unik
     - path: output/kmcp/versions.yml
diff --git a/tests/modules/nf-core/kmcp/profile/main.nf b/tests/modules/nf-core/kmcp/profile/main.nf
@@ -10,23 +10,21 @@ include { KMCP_PROFILE } from '../../../../../modules/nf-core/kmcp/profile/main.
 
 workflow test_kmcp_profile {
 
-    input_compute = [
-        [ id:'test', single_end:false ], // meta map
-        file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/kmcp/NC_045512.2.fasta', checkIfExists: true)
-    ]
+    input_compute= UNTAR ( [
+            [ id:'test' ],
+            file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/kmcp/kmcp_profile.tar.gz", checkIfExists: true)
+        ]).untar
+
 
     input = [
         [ id:'test', single_end:true ], // meta map
         file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
     ]
 
-    taxdump =  [ [],file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/kmcp/kmcp_profile.tar.gz', checkIfExists: true)]
-    ch_taxid = file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/kmcp/seqid2taxid.map', checkIfExists: true)
     mode = 3
 
-    UNTAR (taxdump)
-    KMCP_COMPUTE ( input_compute )
+    KMCP_COMPUTE (input_compute )
     KMCP_INDEX ( KMCP_COMPUTE.out.outdir )
     KMCP_SEARCH ( KMCP_INDEX.out.kmcp.map{it[1]}, input )
-    KMCP_PROFILE ( KMCP_SEARCH.out.result, UNTAR.out.untar.map{ it[1] }, ch_taxid, mode )
+    KMCP_PROFILE ( KMCP_SEARCH.out.result, input_compute.map{ it[1] }, mode )
 }
diff --git a/tests/modules/nf-core/kmcp/profile/nextflow.config b/tests/modules/nf-core/kmcp/profile/nextflow.config
@@ -2,7 +2,7 @@ process {
 
     publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
 
-    withName: KMCP_INDEX {
+    withName: KMCP_COMPUTE {
         ext.prefix = { "${meta.id}_" }
     }
 

diff --git a/tests/modules/nf-core/kmcp/profile/test.yml b/tests/modules/nf-core/kmcp/profile/test.yml
@@ -1,17 +1,17 @@
 - name: kmcp profile test_kmcp_profile
   command: nextflow run ./tests/modules/nf-core/kmcp/profile -entry test_kmcp_profile -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/kmcp/profile/nextflow.config
   tags:
-    - kmcp/profile
     - kmcp
+    - kmcp/profile
   files:
     - path: output/kmcp/test.profile
-    - path: output/untar/versions.yml
+    - path: output/kmcp/versions.yml
 
 - name: kmcp profile test_kmcp_profile stub_run
   command: nextflow run ./tests/modules/nf-core/kmcp/profile -entry test_kmcp_profile -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/kmcp/profile/nextflow.config -stub-run
   tags:
-    - kmcp/profile
     - kmcp
+    - kmcp/profile
   files:
     - path: output/kmcp/test.profile
-    - path: output/untar/versions.yml
+    - path: output/kmcp/versions.yml
diff --git a/tests/modules/nf-core/kmcp/search/nextflow.config b/tests/modules/nf-core/kmcp/search/nextflow.config
@@ -6,8 +6,4 @@ process {
         ext.prefix = { "${meta.id}_" }
     }
 
-    withName: KMCP_SEARCH {
-        ext.prefix = { "${meta.id}_" }
-    }
-
 }
diff --git a/tests/modules/nf-core/kmcp/search/test.yml b/tests/modules/nf-core/kmcp/search/test.yml
@@ -4,9 +4,7 @@
     - kmcp
     - kmcp/search
   files:
-    - path: output/kmcp/test_/R001/__db.yml
-    - path: output/kmcp/test_/R001/__name_mapping.tsv
-    - path: output/kmcp/test_/R001/_block001.uniki
+    - path: output/kmcp/test.gz
     - path: output/kmcp/versions.yml
 
 - name: kmcp search test_kmcp_search stub_run
@@ -15,7 +13,5 @@
     - kmcp
     - kmcp/search
   files:
-    - path: output/kmcp/test_/R001/__db.yml
-    - path: output/kmcp/test_/R001/__name_mapping.tsv
-    - path: output/kmcp/test_/R001/_block001.uniki
+    - path: output/kmcp/test.gz
     - path: output/kmcp/versions.yml