Add postprocessing option to hmmer/eslreformat (#2061)

* Add postprocessing option to hmmer/eslreformat

* Update tests/modules/hmmer/eslreformat/nextflow.config

Co-authored-by: Daniel Straub <42973691+d4straub@users.noreply.github.com>

* Update modules/hmmer/eslreformat/main.nf

Co-authored-by: Daniel Straub <42973691+d4straub@users.noreply.github.com>

Co-authored-by: Daniel Straub <42973691+d4straub@users.noreply.github.com>

modules/hmmer/eslreformat/main.nf

@@ -20,13 +20,14 @@ process HMMER_ESLREFORMAT {
     script:
     def args     = task.ext.args ?: ''
     def prefix   = task.ext.prefix ?: "${meta.id}"
+    // Use for any postprocessing of the sequence file, e.g. removal of gap characters
+    def postproc = task.ext.postprocessing ?: ""
     """
     esl-reformat \\
-        -o ${prefix}.sequences \\
         $args \\
-        $seqfile
+        $seqfile \\
-
+        $postproc \\
-    gzip ${prefix}.sequences
+        | gzip -c > ${prefix}.sequences.gz
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":

tests/modules/hmmer/eslreformat/main.nf

@@ -4,6 +4,7 @@ nextflow.enable.dsl = 2
 
 include { HMMER_HMMALIGN                                 } from '../../../../modules/hmmer/hmmalign/main.nf'
 include { HMMER_ESLREFORMAT as HMMER_ESLREFORMAT_AFA     } from '../../../../modules/hmmer/eslreformat/main.nf'
+include { HMMER_ESLREFORMAT as HMMER_ESLREFORMAT_UNALIGN } from '../../../../modules/hmmer/eslreformat/main.nf'
 
 workflow test_hmmer_eslreformat_afa {
     
@@ -18,3 +19,17 @@ workflow test_hmmer_eslreformat_afa {
 
     HMMER_ESLREFORMAT_AFA ( HMMER_HMMALIGN.out.sthlm )
 }
+
+workflow test_hmmer_eslreformat_unalign {
+    
+    input = [
+        [ id:'test' ], // meta map
+        file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/hmmer/e_coli_k12_16s.fna.gz')      // Change to params.test_data syntax after the data is included in tests/config/test_data.config
+    ]
+
+    hmm   = file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/hmmer/bac.16S_rRNA.hmm.gz')
+
+    HMMER_HMMALIGN ( input, hmm )
+
+    HMMER_ESLREFORMAT_UNALIGN ( HMMER_HMMALIGN.out.sthlm )
+}

tests/modules/hmmer/eslreformat/nextflow.config

@@ -6,4 +6,9 @@ process {
         ext.args   = 'afa'
     }
     
+    withName: HMMER_ESLREFORMAT_UNALIGN {
+        ext.args   = '--gapsym=- afa'
+        ext.postprocessing = '| sed "/^>/!s/-//g"'
+    }
+    
 }

tests/modules/hmmer/eslreformat/test.yml

@@ -7,6 +7,20 @@
     - path: output/hmmer/test.sequences.gz
       contains:
         - ">CP025268.1"
+        - "aaa---UUGAAGAGUUUGAUCAUGGCUCAGAUUGAACGCUGGCGGCAGGCCUAACACAUG"
+    - path: output/hmmer/versions.yml
+      contains:
+        - "easel:"
+- name: hmmer eslreformat test_hmmer_eslreformat_unalign
+  command: nextflow run ./tests/modules/hmmer/eslreformat -entry test_hmmer_eslreformat_unalign -c ./tests/config/nextflow.config  -c ./tests/modules/hmmer/eslreformat/nextflow.config
+  tags:
+    - hmmer/eslreformat
+    - hmmer
+  files:
+    - path: output/hmmer/test.sequences.gz
+      contains:
+        - ">CP025268.1"
+        - "aaaUUGAAGAGUUUGAUCAUGGCUCAGAUUGAACGCUGGCGGCAGGCCUAACACAUG"
     - path: output/hmmer/versions.yml
       contains:
         - "easel:"