From c7f022008c561e70e2ed4a875c17ea1ece109f43 Mon Sep 17 00:00:00 2001
From: James Fellows Yates <jfy133@gmail.com>
Date: Mon, 21 Mar 2022 15:07:59 +0100
Subject: [PATCH] Start getting database prep ready

---
 modules.json                           |  3 +++
 modules/nf-core/modules/untar/main.nf  | 36 ++++++++++++++++++++++++++
 modules/nf-core/modules/untar/meta.yml | 28 ++++++++++++++++++++
 subworkflows/local/db_check.nf         | 14 +++++++++-
 subworkflows/local/input_check.nf      |  2 +-
 5 files changed, 81 insertions(+), 2 deletions(-)
 create mode 100644 modules/nf-core/modules/untar/main.nf
 create mode 100644 modules/nf-core/modules/untar/meta.yml

diff --git a/modules.json b/modules.json
index 6a785b8..9a7f1df 100644
--- a/modules.json
+++ b/modules.json
@@ -23,6 +23,9 @@
             },
             "multiqc": {
                 "git_sha": "20d8250d9f39ddb05dfb437603aaf99b5c0b2b41"
+            },
+            "untar": {
+                "git_sha": "7ec09d0ef4df89617baacc9b2dafcddb7cd4b05a"
             }
         }
     }
diff --git a/modules/nf-core/modules/untar/main.nf b/modules/nf-core/modules/untar/main.nf
new file mode 100644
index 0000000..bbae948
--- /dev/null
+++ b/modules/nf-core/modules/untar/main.nf
@@ -0,0 +1,36 @@
+process UNTAR {
+    tag "$archive"
+    label 'process_low'
+
+    conda (params.enable_conda ? "conda-forge::tar=1.32" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://containers.biocontainers.pro/s3/SingImgsRepo/biocontainers/v1.2.0_cv1/biocontainers_v1.2.0_cv1.img' :
+        'biocontainers/biocontainers:v1.2.0_cv1' }"
+
+    input:
+    path archive
+
+    output:
+    path "$untar"      , emit: untar
+    path "versions.yml", emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def args2 = task.ext.args2 ?: ''
+    untar        = archive.toString() - '.tar.gz'
+    """
+    tar \\
+        -xzvf \\
+        $args \\
+        $archive \\
+        $args2 \\
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        untar: \$(echo \$(tar --version 2>&1) | sed 's/^.*(GNU tar) //; s/ Copyright.*\$//')
+    END_VERSIONS
+    """
+}
diff --git a/modules/nf-core/modules/untar/meta.yml b/modules/nf-core/modules/untar/meta.yml
new file mode 100644
index 0000000..e877a97
--- /dev/null
+++ b/modules/nf-core/modules/untar/meta.yml
@@ -0,0 +1,28 @@
+name: untar
+description: Extract files.
+keywords:
+  - untar
+  - uncompress
+tools:
+  - untar:
+      description: |
+        Extract tar.gz files.
+      documentation: https://www.gnu.org/software/tar/manual/
+      licence: ["GPL-3.0-or-later"]
+input:
+  - archive:
+      type: file
+      description: File to be untar
+      pattern: "*.{tar}.{gz}"
+output:
+  - untar:
+      type: file
+      description:
+      pattern: "*.*"
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+authors:
+  - "@joseespinosa"
+  - "@drpatelh"
diff --git a/subworkflows/local/db_check.nf b/subworkflows/local/db_check.nf
index 909d98f..6f061d7 100644
--- a/subworkflows/local/db_check.nf
+++ b/subworkflows/local/db_check.nf
@@ -19,8 +19,20 @@ workflow DB_CHECK {
         .dump(tag: "db_channel_prepped")
         .set{ dbs }
 
+
+    parsed_samplesheet
+        .branch {
+            untar: it[0]['db_path'].toString().endsWith(".tar.gz")
+            skip: true
+        }
+        .set{ ch_dbs_for_untar }
+
+    UNTAR ( ch_dbs_for_untar.untar )
+
+    ch_final_dbs = ch_dbs_for_untar.skip.mix( ch_dbs_untarred )
+
     emit:
-    dbs                                       // channel: [ val(meta), [ db ] ]
+    dbs = ch_final_dbs                        // channel: [ val(meta), [ db ] ]
     versions = DATABASE_CHECK.out.versions // channel: [ versions.yml ]
 }
 
diff --git a/subworkflows/local/input_check.nf b/subworkflows/local/input_check.nf
index d66fb3a..938c87f 100644
--- a/subworkflows/local/input_check.nf
+++ b/subworkflows/local/input_check.nf
@@ -35,7 +35,7 @@ workflow INPUT_CHECK {
 }
 
 // Function to get list of [ meta, [ fastq_1, fastq_2 ] ]
-def create_fastq_channel(LinkedHashMap row) {
+def create_fastq_channels(LinkedHashMap row) {
     // create meta map
     def meta = [:]
     meta.id                     = row.sample