From f9d25aad7fd6e9896b35552df3529e184c67ebc2 Mon Sep 17 00:00:00 2001
From: "James A. Fellows Yates" <jfy133@gmail.com>
Date: Wed, 23 Feb 2022 10:59:24 +0100
Subject: [PATCH] Switch to native mandatory gzipped reading for hmmalign
 (#1343)

* fix: remove left-over unnecessary code

* Update main.nf

* Update hmmer/hmmalign command to natively accept gzipped input

* Making @drpatelh happy

Co-authored-by: Moritz E. Beber <midnighter@posteo.net>

Co-authored-by: Moritz E. Beber <midnighter@posteo.net>
---
 modules/hmmer/hmmalign/main.nf       | 6 ++----
 modules/hmmer/hmmalign/meta.yml      | 9 +++++----
 tests/modules/hmmer/hmmalign/main.nf | 4 ++--
 3 files changed, 9 insertions(+), 10 deletions(-)

diff --git a/modules/hmmer/hmmalign/main.nf b/modules/hmmer/hmmalign/main.nf
index d3df3c89..f8156e04 100644
--- a/modules/hmmer/hmmalign/main.nf
+++ b/modules/hmmer/hmmalign/main.nf
@@ -21,13 +21,11 @@ process HMMER_HMMALIGN {
     script:
     def args = task.ext.args ?: ''
     def prefix = task.ext.prefix ?: "${meta.id}"
-    def fastacmd = fasta.getExtension() == 'gz' ? "gunzip -c $fasta" : "cat $fasta"
     """
-    $fastacmd | \\
-        hmmalign \\
+    hmmalign \\
         $args \\
         $hmm \\
-        - | gzip -c > ${meta.id}.sthlm.gz
+        $fasta | gzip -c > ${meta.id}.sthlm.gz
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
diff --git a/modules/hmmer/hmmalign/meta.yml b/modules/hmmer/hmmalign/meta.yml
index 296826d4..6c501120 100644
--- a/modules/hmmer/hmmalign/meta.yml
+++ b/modules/hmmer/hmmalign/meta.yml
@@ -19,12 +19,12 @@ input:
         e.g. [ id:'test' ]
   - fasta:
       type: file
-      description: Amino acid or nucleotide fasta file, gzipped or not
-      pattern: "*.{fna,fna.gz,faa,faa.gz,fasta,fasta.gz,fa,fa.gz}"
+      description: Amino acid or nucleotide gzipped compressed fasta file
+      pattern: "*.{fna.gz,faa.gz,fasta.gz,fa.gz}"
   - hmm:
       type: file
-      description: HMM file
-      pattern: "*.hmm"
+      description: A gzipped HMM file
+      pattern: "*.hmm.gz"
 
 output:
   - meta:
@@ -43,3 +43,4 @@ output:
 
 authors:
   - "@erikrikarddaniel"
+  - "@jfy133"
diff --git a/tests/modules/hmmer/hmmalign/main.nf b/tests/modules/hmmer/hmmalign/main.nf
index 8758b124..0a462a7a 100644
--- a/tests/modules/hmmer/hmmalign/main.nf
+++ b/tests/modules/hmmer/hmmalign/main.nf
@@ -8,10 +8,10 @@ workflow test_hmmer_hmmalign {
 
     input = [
         [ id:'test' ], // meta map
-        file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/e_coli_k12_16s.fna')      // Change to params.test_data syntax after the data is included in tests/config/test_data.config
+        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/bac.16S_rRNA.hmm')
+    hmm   = file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/hmmer/bac.16S_rRNA.hmm.gz')
 
     HMMER_HMMALIGN ( input, hmm )
 }