From 66e5236a4a8a8c7443b7a4016ada6f5ac0cfc685 Mon Sep 17 00:00:00 2001
From: MartinPippel <martin.pippel@nbis.se>
Date: Mon, 15 Jan 2024 10:46:39 +0100
Subject: [PATCH 1/3] update nf-core module blast/makeblastdb

---
 modules.json                                  | 95 ++++++++-----------
 .../blast/makeblastdb/blast-makeblastdb.diff  | 20 ----
 .../nf-core/blast/makeblastdb/environment.yml |  7 ++
 modules/nf-core/blast/makeblastdb/main.nf     | 52 ++++++++--
 modules/nf-core/blast/makeblastdb/meta.yml    | 16 +++-
 .../blast/makeblastdb/tests/main.nf.test      | 92 ++++++++++++++++++
 .../blast/makeblastdb/tests/main.nf.test.snap | 86 +++++++++++++++++
 .../blast/makeblastdb/tests/nextflow.config   |  5 +
 .../nf-core/blast/makeblastdb/tests/tags.yml  |  2 +
 9 files changed, 290 insertions(+), 85 deletions(-)
 delete mode 100644 modules/nf-core/blast/makeblastdb/blast-makeblastdb.diff
 create mode 100644 modules/nf-core/blast/makeblastdb/environment.yml
 create mode 100644 modules/nf-core/blast/makeblastdb/tests/main.nf.test
 create mode 100644 modules/nf-core/blast/makeblastdb/tests/main.nf.test.snap
 create mode 100644 modules/nf-core/blast/makeblastdb/tests/nextflow.config
 create mode 100644 modules/nf-core/blast/makeblastdb/tests/tags.yml

diff --git a/modules.json b/modules.json
index 4db56fb3..4caf7cd7 100644
--- a/modules.json
+++ b/modules.json
@@ -1,58 +1,45 @@
 {
-    "name": "NBIS Genome Annotation Workflow",
-    "homePage": "",
-    "repos": {
-        "https://github.com/nf-core/modules.git": {
-            "modules": {
-                "nf-core": {
-                    "blast/makeblastdb": {
-                        "branch": "master",
-                        "git_sha": "911696ea0b62df80e900ef244d7867d177971f73",
-                        "installed_by": [
-                            "modules"
-                        ],
-                        "patch": "modules/nf-core/blast/makeblastdb/blast-makeblastdb.diff"
-                    },
-                    "busco": {
-                        "branch": "master",
-                        "git_sha": "6d6552cb582f56b6101c452e16ee7c23073f91de",
-                        "installed_by": [
-                            "modules"
-                        ]
-                    },
-                    "fastp": {
-                        "branch": "master",
-                        "git_sha": "d497a4868ace3302016ea8ed4b395072d5e833cd",
-                        "installed_by": [
-                            "modules"
-                        ]
-                    },
-                    "fastqc": {
-                        "branch": "master",
-                        "git_sha": "9a4517e720bc812e95b56d23d15a1653b6db4f53",
-                        "installed_by": [
-                            "modules"
-                        ]
-                    },
-                    "interproscan": {
-                        "branch": "master",
-                        "git_sha": "3f5420aa22e00bd030a2556dfdffc9e164ec0ec5",
-                        "installed_by": [
-                            "modules"
-                        ]
-                    },
-                    "multiqc": {
-                        "branch": "master",
-                        "git_sha": "a6e11ac655e744f7ebc724be669dd568ffdc0e80",
-                        "installed_by": [
-                            "modules"
-                        ]
-                    }
-                }
-            },
-            "subworkflows": {
-                "nf-core": {}
-            }
+  "name": "NBIS Genome Annotation Workflow",
+  "homePage": "",
+  "repos": {
+    "https://github.com/nf-core/modules.git": {
+      "modules": {
+        "nf-core": {
+          "blast/makeblastdb": {
+            "branch": "master",
+            "git_sha": "3f5420aa22e00bd030a2556dfdffc9e164ec0ec5",
+            "installed_by": ["modules"]
+          },
+          "busco": {
+            "branch": "master",
+            "git_sha": "6d6552cb582f56b6101c452e16ee7c23073f91de",
+            "installed_by": ["modules"]
+          },
+          "fastp": {
+            "branch": "master",
+            "git_sha": "d497a4868ace3302016ea8ed4b395072d5e833cd",
+            "installed_by": ["modules"]
+          },
+          "fastqc": {
+            "branch": "master",
+            "git_sha": "9a4517e720bc812e95b56d23d15a1653b6db4f53",
+            "installed_by": ["modules"]
+          },
+          "interproscan": {
+            "branch": "master",
+            "git_sha": "3f5420aa22e00bd030a2556dfdffc9e164ec0ec5",
+            "installed_by": ["modules"]
+          },
+          "multiqc": {
+            "branch": "master",
+            "git_sha": "a6e11ac655e744f7ebc724be669dd568ffdc0e80",
+            "installed_by": ["modules"]
+          }
         }
+      },
+      "subworkflows": {
+        "nf-core": {}
+      }
     }
+  }
 }
diff --git a/modules/nf-core/blast/makeblastdb/blast-makeblastdb.diff b/modules/nf-core/blast/makeblastdb/blast-makeblastdb.diff
deleted file mode 100644
index daacf209..00000000
--- a/modules/nf-core/blast/makeblastdb/blast-makeblastdb.diff
+++ /dev/null
@@ -1,20 +0,0 @@
-Changes in module 'nf-core/blast/makeblastdb'
---- modules/nf-core/blast/makeblastdb/main.nf
-+++ modules/nf-core/blast/makeblastdb/main.nf
-@@ -8,11 +8,11 @@
-         'biocontainers/blast:2.13.0--hf3cf87c_0' }"
- 
-     input:
--    path fasta
-+    tuple val(meta), path(fasta)
- 
-     output:
--    path 'blast_db'     , emit: db
--    path "versions.yml" , emit: versions
-+    tuple val(meta), path('blast_db'), emit: db
-+    path "versions.yml"              , emit: versions
- 
-     when:
-     task.ext.when == null || task.ext.when
-
-************************************************************
diff --git a/modules/nf-core/blast/makeblastdb/environment.yml b/modules/nf-core/blast/makeblastdb/environment.yml
new file mode 100644
index 00000000..cf84a60e
--- /dev/null
+++ b/modules/nf-core/blast/makeblastdb/environment.yml
@@ -0,0 +1,7 @@
+name: blast_makeblastdb
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+dependencies:
+  - bioconda::blast=2.14.1
diff --git a/modules/nf-core/blast/makeblastdb/main.nf b/modules/nf-core/blast/makeblastdb/main.nf
index 93f88512..cd1083d0 100644
--- a/modules/nf-core/blast/makeblastdb/main.nf
+++ b/modules/nf-core/blast/makeblastdb/main.nf
@@ -1,30 +1,62 @@
 process BLAST_MAKEBLASTDB {
-    tag "$fasta"
+    tag "$meta.id"
     label 'process_medium'
 
-    conda "bioconda::blast=2.13.0"
+    conda "${moduleDir}/environment.yml"
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
-        'https://depot.galaxyproject.org/singularity/blast:2.13.0--hf3cf87c_0' :
-        'biocontainers/blast:2.13.0--hf3cf87c_0' }"
+        'https://depot.galaxyproject.org/singularity/blast:2.14.1--pl5321h6f7f691_0':
+        'biocontainers/blast:2.14.1--pl5321h6f7f691_0' }"
 
     input:
     tuple val(meta), path(fasta)
 
     output:
-    tuple val(meta), path('blast_db'), emit: db
-    path "versions.yml"              , emit: versions
+    tuple val(meta), path("${meta.id}"), emit: db
+    path "versions.yml"                , emit: versions
 
     when:
     task.ext.when == null || task.ext.when
 
     script:
     def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def is_compressed = fasta.getExtension() == "gz" ? true : false
+    def fasta_name = is_compressed ? fasta.getBaseName() : fasta
     """
+    if [ "${is_compressed}" == "true" ]; then
+        gzip -c -d ${fasta} > ${fasta_name}
+    fi
+
     makeblastdb \\
-        -in $fasta \\
-        $args
-    mkdir blast_db
-    mv ${fasta}* blast_db
+        -in ${fasta_name} \\
+        ${args}
+    mkdir ${prefix}
+    mv ${fasta_name}* ${prefix}
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        blast: \$(blastn -version 2>&1 | sed 's/^.*blastn: //; s/ .*\$//')
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def is_compressed = fasta.getExtension() == "gz" ? true : false
+    def fasta_name = is_compressed ? fasta.getBaseName() : fasta
+    """
+    touch ${fasta_name}.fasta
+    touch ${fasta_name}.fasta.ndb
+    touch ${fasta_name}.fasta.nhr
+    touch ${fasta_name}.fasta.nin
+    touch ${fasta_name}.fasta.njs
+    touch ${fasta_name}.fasta.not
+    touch ${fasta_name}.fasta.nsq
+    touch ${fasta_name}.fasta.ntf
+    touch ${fasta_name}.fasta.nto
+    mkdir ${prefix}
+    mv ${fasta_name}* ${prefix}
+
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
         blast: \$(blastn -version 2>&1 | sed 's/^.*blastn: //; s/ .*\$//')
diff --git a/modules/nf-core/blast/makeblastdb/meta.yml b/modules/nf-core/blast/makeblastdb/meta.yml
index 83c4b292..9ed63901 100644
--- a/modules/nf-core/blast/makeblastdb/meta.yml
+++ b/modules/nf-core/blast/makeblastdb/meta.yml
@@ -13,11 +13,21 @@ tools:
       doi: 10.1016/S0022-2836(05)80360-2
       licence: ["US-Government-Work"]
 input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
   - fasta:
       type: file
       description: Input fasta file
-      pattern: "*.{fa,fasta}"
+      pattern: "*.{fa,fasta,fa.gz,fasta.gz}"
 output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
   - db:
       type: directory
       description: Output directory containing blast database files
@@ -29,3 +39,7 @@ output:
 authors:
   - "@joseespinosa"
   - "@drpatelh"
+maintainers:
+  - "@joseespinosa"
+  - "@drpatelh"
+  - "@vagkaratzas"
diff --git a/modules/nf-core/blast/makeblastdb/tests/main.nf.test b/modules/nf-core/blast/makeblastdb/tests/main.nf.test
new file mode 100644
index 00000000..983b165f
--- /dev/null
+++ b/modules/nf-core/blast/makeblastdb/tests/main.nf.test
@@ -0,0 +1,92 @@
+nextflow_process {
+
+    name "Test Process BLAST_MAKEBLASTDB"
+    script "../main.nf"
+    process "BLAST_MAKEBLASTDB"
+    config "./nextflow.config"
+    tag "modules"
+    tag "modules_nfcore"
+    tag "blast"
+    tag "blast/makeblastdb"
+
+    test("Should build a blast db folder from a fasta file") {
+
+        when {
+            params {
+                outdir = "$outputDir"
+            }
+            process {
+                """
+                input[0] = [ [id:'test'], file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true) ]
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert process.out.db
+                    with(process.out.db) {
+                        assert size() == 1
+                        with(get(0).get(1)) {
+                            File folder = new File(get(0).get(1))
+                            File[] listOfFiles = folder.listFiles()
+                            listOfFiles = listOfFiles.sort { it.name }
+                            assert listOfFiles.length == 9
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[0].name}").match("genome.fasta")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[1].name}").match("genome.fasta.ndb")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[2].name}").match("genome.fasta.nhr")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[5].name}").match("genome.fasta.not")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[6].name}").match("genome.fasta.nsq")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[7].name}").match("genome.fasta.ntf")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[8].name}").match("genome.fasta.nto")
+                        }
+                    }
+                },
+                { assert process.out.versions }
+            )
+        }
+
+    }
+
+    test("Should build a blast db folder from a zipped fasta file") {
+
+        when {
+            params {
+                outdir = "$outputDir"
+            }
+            process {
+                """
+                input[0] = [ [id:'test'], file(params.test_data['sarscov2']['genome']['genome_fasta_gz'], checkIfExists: true) ]
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert process.out.db
+                    with(process.out.db) {
+                        assert size() == 1
+                        with(get(0).get(1)) {
+                            File folder = new File(get(0).get(1))
+                            File[] listOfFiles = folder.listFiles()
+                            listOfFiles = listOfFiles.sort { it.name }
+                            assert listOfFiles.length == 10
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[0].name}").match("gz_genome.fasta")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[2].name}").match("gz_genome.fasta.ndb")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[3].name}").match("gz_genome.fasta.nhr")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[6].name}").match("gz_genome.fasta.not")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[7].name}").match("gz_genome.fasta.nsq")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[8].name}").match("gz_genome.fasta.ntf")
+                            assert snapshot("${get(0).get(1)}/${listOfFiles[9].name}").match("gz_genome.fasta.nto")
+                        }
+                    }
+                },
+                { assert process.out.versions }
+            )
+        }
+
+    }
+
+}
diff --git a/modules/nf-core/blast/makeblastdb/tests/main.nf.test.snap b/modules/nf-core/blast/makeblastdb/tests/main.nf.test.snap
new file mode 100644
index 00000000..b6f040ed
--- /dev/null
+++ b/modules/nf-core/blast/makeblastdb/tests/main.nf.test.snap
@@ -0,0 +1,86 @@
+{
+    "genome.fasta": {
+        "content": [
+            "genome.fasta:md5,6e9fe4042a72f2345f644f239272b7e6"
+        ],
+        "timestamp": "2023-11-07T12:52:38.457245596"
+    },
+    "gz_genome.fasta.ntf": {
+        "content": [
+            "genome.fasta.ntf:md5,de1250813f0c7affc6d12dac9d0fb6bb"
+        ],
+        "timestamp": "2023-11-07T12:58:02.121840034"
+    },
+    "genome.fasta.not": {
+        "content": [
+            "genome.fasta.not:md5,1e53e9d08f1d23af0299cfa87478a7bb"
+        ],
+        "timestamp": "2023-11-07T12:55:33.862012946"
+    },
+    "genome.fasta.nhr": {
+        "content": [
+            "genome.fasta.nhr:md5,f4b4ddb034fd3dd7b25c89e9d50c004e"
+        ],
+        "timestamp": "2023-11-07T12:55:33.857994517"
+    },
+    "gz_genome.fasta.nhr": {
+        "content": [
+            "genome.fasta.nhr:md5,f4b4ddb034fd3dd7b25c89e9d50c004e"
+        ],
+        "timestamp": "2023-11-07T12:58:02.102407993"
+    },
+    "genome.fasta.ntf": {
+        "content": [
+            "genome.fasta.ntf:md5,de1250813f0c7affc6d12dac9d0fb6bb"
+        ],
+        "timestamp": "2023-11-07T12:55:33.877288786"
+    },
+    "gz_genome.fasta.not": {
+        "content": [
+            "genome.fasta.not:md5,1e53e9d08f1d23af0299cfa87478a7bb"
+        ],
+        "timestamp": "2023-11-07T12:58:02.108135313"
+    },
+    "gz_genome.fasta.ndb": {
+        "content": [
+            "genome.fasta.ndb:md5,0d553c830656469211de113c5022f06d"
+        ],
+        "timestamp": "2023-11-07T12:58:02.094305556"
+    },
+    "gz_genome.fasta.nsq": {
+        "content": [
+            "genome.fasta.nsq:md5,982cbc7d9e38743b9b1037588862b9da"
+        ],
+        "timestamp": "2023-11-07T12:58:02.115010863"
+    },
+    "genome.fasta.nto": {
+        "content": [
+            "genome.fasta.nto:md5,33cdeccccebe80329f1fdbee7f5874cb"
+        ],
+        "timestamp": "2023-11-07T12:55:33.890761822"
+    },
+    "gz_genome.fasta.nto": {
+        "content": [
+            "genome.fasta.nto:md5,33cdeccccebe80329f1fdbee7f5874cb"
+        ],
+        "timestamp": "2023-11-07T12:58:02.12931429"
+    },
+    "genome.fasta.ndb": {
+        "content": [
+            "genome.fasta.ndb:md5,0d553c830656469211de113c5022f06d"
+        ],
+        "timestamp": "2023-11-07T12:55:33.853303997"
+    },
+    "genome.fasta.nsq": {
+        "content": [
+            "genome.fasta.nsq:md5,982cbc7d9e38743b9b1037588862b9da"
+        ],
+        "timestamp": "2023-11-07T12:55:33.866667927"
+    },
+    "gz_genome.fasta": {
+        "content": [
+            "genome.fasta:md5,6e9fe4042a72f2345f644f239272b7e6"
+        ],
+        "timestamp": "2023-11-07T12:58:02.081764854"
+    }
+}
\ No newline at end of file
diff --git a/modules/nf-core/blast/makeblastdb/tests/nextflow.config b/modules/nf-core/blast/makeblastdb/tests/nextflow.config
new file mode 100644
index 00000000..0899289b
--- /dev/null
+++ b/modules/nf-core/blast/makeblastdb/tests/nextflow.config
@@ -0,0 +1,5 @@
+process {
+    withName: BLAST_MAKEBLASTDB {
+        ext.args = '-dbtype nucl'
+    }
+}
diff --git a/modules/nf-core/blast/makeblastdb/tests/tags.yml b/modules/nf-core/blast/makeblastdb/tests/tags.yml
new file mode 100644
index 00000000..ab74c02a
--- /dev/null
+++ b/modules/nf-core/blast/makeblastdb/tests/tags.yml
@@ -0,0 +1,2 @@
+blast/makeblastdb:
+  - modules/nf-core/blast/makeblastdb/**

From 6911fea467d6098536f6c53c4b95a7c984fa0268 Mon Sep 17 00:00:00 2001
From: MartinPippel <martin.pippel@nbis.se>
Date: Mon, 15 Jan 2024 13:01:27 +0100
Subject: [PATCH 2/3] update local blast version

---
 modules/local/blast/blastp.nf | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/modules/local/blast/blastp.nf b/modules/local/blast/blastp.nf
index e4bf1127..de43bb44 100644
--- a/modules/local/blast/blastp.nf
+++ b/modules/local/blast/blastp.nf
@@ -2,10 +2,10 @@ process BLAST_BLASTP {
     tag "${fasta.baseName}"
     label 'process_medium'
 
-    conda 'bioconda::blast=2.12.0'
+    conda 'bioconda::blast=2.15.0'
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
-        'https://depot.galaxyproject.org/singularity/blast:2.12.0--pl5262h3289130_0' :
-        'biocontainers/blast:2.12.0--pl5262h3289130_0' }"
+        'https://depot.galaxyproject.org/singularity/blast:2.15.0--pl5321h6f7f691_1' :
+        'biocontainers/blast:2.15.0--pl5321h6f7f691_1' }"
 
     input:
     tuple val(meta), path(fasta)

From 48c4d9a3f5ecf66061ac510011d7a50dda9b6fbb Mon Sep 17 00:00:00 2001
From: MartinPippel <martin.pippel@nbis.se>
Date: Mon, 15 Jan 2024 13:03:36 +0100
Subject: [PATCH 3/3] update module nf-core/blast/makeblastdb

---
 modules.json                                      | 2 +-
 modules/nf-core/blast/makeblastdb/environment.yml | 2 +-
 modules/nf-core/blast/makeblastdb/main.nf         | 4 ++--
 3 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/modules.json b/modules.json
index 4caf7cd7..bd029df7 100644
--- a/modules.json
+++ b/modules.json
@@ -7,7 +7,7 @@
         "nf-core": {
           "blast/makeblastdb": {
             "branch": "master",
-            "git_sha": "3f5420aa22e00bd030a2556dfdffc9e164ec0ec5",
+            "git_sha": "a01c66c96e0bc610ad126e7adc4a94cd4acd1b48",
             "installed_by": ["modules"]
           },
           "busco": {
diff --git a/modules/nf-core/blast/makeblastdb/environment.yml b/modules/nf-core/blast/makeblastdb/environment.yml
index cf84a60e..a20783b0 100644
--- a/modules/nf-core/blast/makeblastdb/environment.yml
+++ b/modules/nf-core/blast/makeblastdb/environment.yml
@@ -4,4 +4,4 @@ channels:
   - bioconda
   - defaults
 dependencies:
-  - bioconda::blast=2.14.1
+  - bioconda::blast=2.15.0
diff --git a/modules/nf-core/blast/makeblastdb/main.nf b/modules/nf-core/blast/makeblastdb/main.nf
index cd1083d0..a2c73189 100644
--- a/modules/nf-core/blast/makeblastdb/main.nf
+++ b/modules/nf-core/blast/makeblastdb/main.nf
@@ -4,8 +4,8 @@ process BLAST_MAKEBLASTDB {
 
     conda "${moduleDir}/environment.yml"
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
-        'https://depot.galaxyproject.org/singularity/blast:2.14.1--pl5321h6f7f691_0':
-        'biocontainers/blast:2.14.1--pl5321h6f7f691_0' }"
+        'https://depot.galaxyproject.org/singularity/blast:2.15.0--pl5321h6f7f691_1':
+        'biocontainers/blast:2.15.0--pl5321h6f7f691_1' }"
 
     input:
     tuple val(meta), path(fasta)