Simplify operator logic with static-typed maps

Signed-off-by: Ben Sherman <[email protected]>

Author: bentsherman

main.nf

@@ -22,7 +22,6 @@ include { PIPELINE_INITIALISATION } from './subworkflows/local/utils_nfcore_fetc
 include { PIPELINE_COMPLETION     } from './subworkflows/local/utils_nfcore_fetchngs_pipeline'
 include { SOFTWARE_VERSIONS       } from './subworkflows/nf-core/utils_nfcore_pipeline'
 include { DownloadMethod          } from './workflows/sra'
-include { SraParams               } from './workflows/sra'
 include { Sample                  } from './workflows/sra'
 
 /*
@@ -45,7 +44,7 @@ params {
     download_method: DownloadMethod = 'ftp'
 
     // Only download metadata for public data database ids and don't download the FastQ files.
-    skip_fastq_download: boolean = false
+    skip_fastq_download: Boolean = false
 
     // dbGaP repository key.
     dbgap_key: Path? = null
@@ -78,12 +77,12 @@ workflow {
     //
     sra = SRA (
         Channel.fromList(params.input),
-        SraParams(
-            params.ena_metadata_fields,
-            params.download_method,
-            params.skip_fastq_download,
-            params.dbgap_key
-        )
+        [
+            ena_metadata_fields: params.ena_metadata_fields,
+            download_method: params.download_method,
+            skip_fastq_download: params.skip_fastq_download,
+            dbgap_key: params.dbgap_key
+        ]
     )
 
     //
@@ -128,7 +127,7 @@ output {
         }
     }
 
-    metadata {
+    metadata: Channel<Path> {
         path 'metadata'
     }
 

modules/local/aspera_cli/main.nf

@@ -1,5 +1,5 @@
 process ASPERA_CLI {
-    tag "$meta.id"
+    tag "$id"
     label 'process_medium'
 
     conda "${moduleDir}/environment.yml"
@@ -8,34 +8,39 @@ process ASPERA_CLI {
         'biocontainers/aspera-cli:4.14.0--hdfd78af_1' }"
 
     input:
-    meta    : Map<String,String>
-    user    : String
+    id              : String
+    single_end      : Boolean
+    fastq_aspera    : String
+    md5_1           : String
+    md5_2           : String?
+    user            : String
 
     output:
+    id      : String = id
     fastq_1 : Path  = file('*_1.fastq.gz')
     fastq_2 : Path? = file('*_2.fastq.gz')
     md5_1   : Path  = file('*_1.fastq.gz.md5')
     md5_2   : Path? = file('*_2.fastq.gz.md5')
 
     topic:
-    ( task.process, 'aspera_cli', eval('ascli --version') ) >> 'versions'
+    [process: task.process, name: 'aspera_cli', version: eval('ascli --version')] >> 'versions'
 
     script:
     def args = task.ext.args ?: ''
     def conda_prefix = ['singularity', 'apptainer'].contains(workflow.containerEngine) ? "export CONDA_PREFIX=/usr/local" : ""
-    def fastq = meta.fastq_aspera.tokenize(';')
-    if (meta.single_end.toBoolean()) {
+    def fastq = fastq_aspera.tokenize(';')
+    if (single_end) {
         """
         $conda_prefix
 
         ascp \\
             $args \\
             -i \$CONDA_PREFIX/etc/aspera/aspera_bypass_dsa.pem \\
             ${user}@${fastq[0]} \\
-            ${meta.id}.fastq.gz
+            ${id}.fastq.gz
 
-        echo "${meta.md5_1}  ${meta.id}.fastq.gz" > ${meta.id}.fastq.gz.md5
-        md5sum -c ${meta.id}.fastq.gz.md5
+        echo "${md5_1}  ${id}.fastq.gz" > ${id}.fastq.gz.md5
+        md5sum -c ${id}.fastq.gz.md5
         """
     } else {
         """
@@ -45,19 +50,19 @@ process ASPERA_CLI {
             $args \\
             -i \$CONDA_PREFIX/etc/aspera/aspera_bypass_dsa.pem \\
             ${user}@${fastq[0]} \\
-            ${meta.id}_1.fastq.gz
+            ${id}_1.fastq.gz
 
-        echo "${meta.md5_1}  ${meta.id}_1.fastq.gz" > ${meta.id}_1.fastq.gz.md5
-        md5sum -c ${meta.id}_1.fastq.gz.md5
+        echo "${md5_1}  ${id}_1.fastq.gz" > ${id}_1.fastq.gz.md5
+        md5sum -c ${id}_1.fastq.gz.md5
 
         ascp \\
             $args \\
             -i \$CONDA_PREFIX/etc/aspera/aspera_bypass_dsa.pem \\
             ${user}@${fastq[1]} \\
-            ${meta.id}_2.fastq.gz
+            ${id}_2.fastq.gz
 
-        echo "${meta.md5_2}  ${meta.id}_2.fastq.gz" > ${meta.id}_2.fastq.gz.md5
-        md5sum -c ${meta.id}_2.fastq.gz.md5
+        echo "${md5_2}  ${id}_2.fastq.gz" > ${id}_2.fastq.gz.md5
+        md5sum -c ${id}_2.fastq.gz.md5
         """
     }
 }

modules/local/sra_fastq_ftp/main.nf

@@ -1,6 +1,6 @@
 
 process SRA_FASTQ_FTP {
-    tag "$meta.id"
+    tag "$id"
     label 'process_low'
     label 'error_retry'
 
@@ -10,46 +10,52 @@ process SRA_FASTQ_FTP {
         'biocontainers/wget:1.20.1' }"
 
     input:
-    meta    : Map<String,String>
+    id              : String
+    single_end      : Boolean
+    fastq_1         : String
+    fastq_2         : String?
+    md5_1           : String
+    md5_2           : String?
 
     output:
+    id      : String = id
     fastq_1 : Path  = file('*_1.fastq.gz')
     fastq_2 : Path? = file('*_2.fastq.gz')
     md5_1   : Path  = file('*_1.fastq.gz.md5')
     md5_2   : Path? = file('*_2.fastq.gz.md5')
 
     topic:
-    ( task.process, 'wget', eval("echo \$(wget --version | head -n 1 | sed 's/^GNU Wget //; s/ .*\$//')") ) >> 'versions'
+    [process: task.process, name: 'wget', version: eval("echo \$(wget --version | head -n 1 | sed 's/^GNU Wget //; s/ .*\$//')")] >> 'versions'
 
     script:
     def args = task.ext.args ?: ''
-    if (meta.single_end.toBoolean()) {
+    if (single_end) {
         """
         wget \\
             $args \\
-            -O ${meta.id}.fastq.gz \\
-            ${meta.fastq_1}
+            -O ${id}.fastq.gz \\
+            ${fastq_1}
 
-        echo "${meta.md5_1}  ${meta.id}.fastq.gz" > ${meta.id}.fastq.gz.md5
-        md5sum -c ${meta.id}.fastq.gz.md5
+        echo "${md5_1}  ${id}.fastq.gz" > ${id}.fastq.gz.md5
+        md5sum -c ${id}.fastq.gz.md5
         """
     } else {
         """
         wget \\
             $args \\
-            -O ${meta.id}_1.fastq.gz \\
-            ${meta.fastq_1}
+            -O ${id}_1.fastq.gz \\
+            ${fastq_1}
 
-        echo "${meta.md5_1}  ${meta.id}_1.fastq.gz" > ${meta.id}_1.fastq.gz.md5
-        md5sum -c ${meta.id}_1.fastq.gz.md5
+        echo "${md5_1}  ${id}_1.fastq.gz" > ${id}_1.fastq.gz.md5
+        md5sum -c ${id}_1.fastq.gz.md5
 
         wget \\
             $args \\
-            -O ${meta.id}_2.fastq.gz \\
-            ${meta.fastq_2}
+            -O ${id}_2.fastq.gz \\
+            ${fastq_2}
 
-        echo "${meta.md5_2}  ${meta.id}_2.fastq.gz" > ${meta.id}_2.fastq.gz.md5
-        md5sum -c ${meta.id}_2.fastq.gz.md5
+        echo "${md5_2}  ${id}_2.fastq.gz" > ${id}_2.fastq.gz.md5
+        md5sum -c ${id}_2.fastq.gz.md5
         """
     }
 }

modules/local/sra_ids_to_runinfo/main.nf

@@ -16,7 +16,7 @@ process SRA_IDS_TO_RUNINFO {
     file('*.runinfo.tsv')
 
     topic:
-    ( task.process, 'python', eval("python --version | sed 's/Python //g'") ) >> 'versions'
+    [process: task.process, name: 'python', version: eval("python --version | sed 's/Python //g'")] >> 'versions'
 
     script:
     def metadata_fields = fields ? "--ena_metadata_fields ${fields}" : ''

modules/local/sra_runinfo_to_ftp/main.nf

@@ -13,7 +13,7 @@ process SRA_RUNINFO_TO_FTP {
     file('*.runinfo_ftp.tsv')
 
     topic:
-    ( task.process, 'python', eval("python --version | sed 's/Python //g'") ) >> 'versions'
+    [process: task.process, name: 'python', version: eval("python --version | sed 's/Python //g'")] >> 'versions'
 
     script:
     """