EGGNOGMAPPER: Handled optional inputs (#5477)

* EGGNOGMAPPER: Handled optional inputs * Included versions in snapshot * Added test for empty-params * Now using a single common setup
nf-core · Apr 15, 2024 · c070729 · c070729
1 parent 9c0d4eb
commit c070729
Show file tree

Hide file tree

Showing 4 changed files with 114 additions and 25 deletions.
diff --git a/modules/nf-core/eggnogmapper/main.nf b/modules/nf-core/eggnogmapper/main.nf
@@ -1,6 +1,6 @@
 process EGGNOGMAPPER {
     tag "$meta.id"
-    label 'process_long'
+    label 'process_high'
 
     conda "${moduleDir}/environment.yml"
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
@@ -23,11 +23,13 @@ process EGGNOGMAPPER {
     task.ext.when == null || task.ext.when
 
     script:
-    def args = task.ext.args ?: ''
-    def prefix = task.ext.prefix ?: "${meta.id}"
-    def is_compressed = fasta.name.endsWith(".gz")
-    def fasta_name = fasta.name.replace(".gz", "")
-    def dbmem = task.memory.toMega() > 40000 ? '--dbmem' : ''
+    def args            = task.ext.args                 ?: ''
+    def prefix          = task.ext.prefix               ?: "${meta.id}"
+    def is_compressed   = fasta.extension == '.gz'      ? true                              : false
+    def fasta_name      = is_compressed                 ? fasta.baseName                    : "$fasta"
+    def dbmem           = task.memory.toMega() > 40000  ? '--dbmem'                         : ''
+    def database_arg    = eggnog_db                     ? "--database $eggnog_db"           : ''
+    def dmnd_db_arg     = eggnog_diamond_db             ? "--dmnd_db $eggnog_diamond_db"    : ''
     """
     if [ "$is_compressed" == "true" ]; then
         gzip -c -d $fasta > $fasta_name
@@ -38,8 +40,8 @@ process EGGNOGMAPPER {
         -i ${fasta_name} \\
         --data_dir ${eggnog_data_dir} \\
         -m diamond \\
-        --dmnd_db ${eggnog_diamond_db} \\
-        --database ${eggnog_db} \\
+        $dmnd_db_arg \\
+        $database_arg \\
         --output ${prefix} \\
         ${dbmem} \\
         $args
@@ -51,8 +53,8 @@ process EGGNOGMAPPER {
     """
 
     stub:
-    def args = task.ext.args ?: ''
-    def prefix = task.ext.prefix ?: "${meta.id}"
+    def args    = task.ext.args ?: ''
+    def prefix  = task.ext.prefix ?: "${meta.id}"
     """
     touch ${prefix}.emapper.annotations
     touch ${prefix}.emapper.seed_orthologs

diff --git a/modules/nf-core/eggnogmapper/meta.yml b/modules/nf-core/eggnogmapper/meta.yml
@@ -60,3 +60,6 @@ output:
       pattern: "versions.yml"
 authors:
   - "@vagkaratzas"
+maintainers:
+  - "@vagkaratzas"
+  - "@gallvp"
diff --git a/modules/nf-core/eggnogmapper/tests/main.nf.test b/modules/nf-core/eggnogmapper/tests/main.nf.test
@@ -8,21 +8,21 @@ nextflow_process {
     tag "eggnogmapper"
     tag "diamond/makedb"
 
-    test("Should search for protein annotations against the eggnogmapper db") {
-
-        setup {
-            run("DIAMOND_MAKEDB") {
-                script "../../diamond/makedb/main.nf"
-                process {
-                    """
-                    input[0] = [ [id:'test2'], file(params.test_data['sarscov2']['genome']['proteome_fasta'], checkIfExists: true) ]
-                    input[1] = []
-                    input[2] = []
-                    input[3] = []
-                    """
-                }
+    setup {
+        run("DIAMOND_MAKEDB") {
+            script "../../diamond/makedb/main.nf"
+            process {
+                """
+                input[0] = [ [id:'test2'], file(params.test_data['sarscov2']['genome']['proteome_fasta'], checkIfExists: true) ]
+                input[1] = []
+                input[2] = []
+                input[3] = []
+                """
             }
         }
+    }
+
+    test("Should search for protein annotations against the eggnogmapper db") {
 
         when {
             params {
@@ -47,7 +47,46 @@ nextflow_process {
                 { assert path(process.out.annotations.get(0).get(1)).readLines().contains("ENSSASP00005000002.1\tENSSASP00005000002.1\t0.0\t14179.0\tCOG0498@1|root,COG0498@2|Bacteria,1MUWQ@1224|Proteobacteria,2VHR6@28216|Betaproteobacteria,2KUMA@206389|Rhodocyclales\t1224|Proteobacteria\tE\tthreonine synthase\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-") },
                 { assert path(process.out.orthologs.get(0).get(1)).readLines().contains("ENSSASP00005000002.1\tENSSASP00005000002.1\t0.0\t14179.0\t1\t7096\t1\t7096\t100.0\t100.0\t100.0") },
                 { assert snapshot(process.out.hits).match("hits") },
-                { assert process.out.versions }
+                { assert snapshot(process.out.versions).match("versions") }
+            )
+        }
+
+    }
+
+    test("Should search for protein annotations against the eggnogmapper db -- empty-params") {
+
+        when {
+            params {
+                outdir = "$outputDir"
+            }
+            process {
+                """
+                eggnog_db = file("https://github.com/nf-core/test-datasets/raw/eddf5b0e3336e0f93c81d4b4843b07257f9efaec/data/delete_me/eggnogmapper/eggnog.db", checkIfExists: true)
+                eggnog_db.copyTo("${workDir}/tmp/eggnog.db")
+
+                ch_synced_inputs = DIAMOND_MAKEDB.out.db.map { meta, dmnd ->
+                        dmnd.copyTo("${workDir}/tmp/eggnog_proteins.dmnd")
+
+                        return true
+                    }
+                    | combine ( Channel.fromPath( "${workDir}/tmp/" ) )
+                eggnog_data_dir = ch_synced_inputs.map { sync_status, data_dir -> data_dir }
+
+                input[0] = [ [id:'test'], file(params.test_data['sarscov2']['genome']['proteome_fasta'], checkIfExists: true) ]
+                input[1] = []
+                input[2] = eggnog_data_dir
+                input[3] = [[], []]
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert path(process.out.annotations.get(0).get(1)).readLines().contains("ENSSASP00005000002.1\tENSSASP00005000002.1\t0.0\t14179.0\tCOG0498@1|root,COG0498@2|Bacteria,1MUWQ@1224|Proteobacteria,2VHR6@28216|Betaproteobacteria,2KUMA@206389|Rhodocyclales\t1224|Proteobacteria\tE\tthreonine synthase\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-\t-") },
+                { assert path(process.out.orthologs.get(0).get(1)).readLines().contains("ENSSASP00005000002.1\tENSSASP00005000002.1\t0.0\t14179.0\t1\t7096\t1\t7096\t100.0\t100.0\t100.0") },
+                { assert snapshot(process.out.hits).match("hits--empty-params") },
+                { assert snapshot(process.out.versions).match("versions--empty-params") }
             )
         }
 

diff --git a/modules/nf-core/eggnogmapper/tests/main.nf.test.snap b/modules/nf-core/eggnogmapper/tests/main.nf.test.snap
@@ -10,6 +10,51 @@
                 ]
             ]
         ],
-        "timestamp": "2023-11-08T20:43:50.173213923"
+        "meta": {
+            "nf-test": "0.8.4",
+            "nextflow": "23.10.1"
+        },
+        "timestamp": "2024-04-12T13:43:01.751295"
+    },
+    "versions--empty-params": {
+        "content": [
+            [
+                "versions.yml:md5,d3e4efad28b5a924585ea3dfcf72c32c"
+            ]
+        ],
+        "meta": {
+            "nf-test": "0.8.4",
+            "nextflow": "23.10.1"
+        },
+        "timestamp": "2024-04-15T11:42:23.737523"
+    },
+    "versions": {
+        "content": [
+            [
+                "versions.yml:md5,d3e4efad28b5a924585ea3dfcf72c32c"
+            ]
+        ],
+        "meta": {
+            "nf-test": "0.8.4",
+            "nextflow": "23.10.1"
+        },
+        "timestamp": "2024-04-12T13:45:26.555465"
+    },
+    "hits--empty-params": {
+        "content": [
+            [
+                [
+                    {
+                        "id": "test"
+                    },
+                    "test.emapper.hits:md5,864b7a1f902893d8aee6621baeab7be8"
+                ]
+            ]
+        ],
+        "meta": {
+            "nf-test": "0.8.4",
+            "nextflow": "23.10.1"
+        },
+        "timestamp": "2024-04-15T11:42:23.690105"
     }
 }