Merge pull request #286 from nf-core/multiqc-optimisations

MultiQC Optimisations

CHANGELOG.md

@@ -16,6 +16,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - [#270](https://github.com/nf-core/taxprofiler/pull/270/files) Fixed warning for host removal index parameter, and improved index checks (♥ to @prototaxites for reporting, fix by @jfy133)
 - [#274](https://github.com/nf-core/taxprofiler/pull/274/files) Substituted the samtools/bam2fq module with samtools/fastq module (fix by @sofstam)
 - [#275](https://github.com/nf-core/taxprofiler/pull/275/files) Replaced function used for error reporting to more Nextflow friendly method (fix by @jfy133)
+- [#286](https://github.com/nf-core/taxprofiler/pull/286/files) Runtime optimisation of MultiQC step via improved log file processing (fix by @Midnighter & @jfy133)
 
 ### `Dependencies`
 

assets/multiqc_config.yml

@@ -37,8 +37,7 @@ run_modules:
 
 sp:
   diamond:
-    contents: "diamond v"
-    num_lines: 10
+    fn_re: ".*.diamond.log$"
   fastqc/data:
     fn_re: ".*(fastqc|falco)_data.txt$"
   fastqc/zip:

workflows/taxprofiler.nf

@@ -273,7 +273,12 @@ workflow TAXPROFILER {
     ch_multiqc_files = ch_multiqc_files.mix(CUSTOM_DUMPSOFTWAREVERSIONS.out.mqc_yml.collect())
 
     if ( params.preprocessing_qc_tool == 'falco' ) {
-        ch_multiqc_files = ch_multiqc_files.mix(FALCO.out.txt.collect{it[1]}.ifEmpty([]))
+        // only mix in files acutally used by MultiQC
+        ch_multiqc_files = ch_multiqc_files.mix(FALCO.out.txt
+                            .map { meta, reports -> reports }
+                            .flatten()
+                            .filter { path -> path.name.endsWith('_data.txt')}
+                            .ifEmpty([]))
     } else {
         ch_multiqc_files = ch_multiqc_files.mix(FASTQC.out.zip.collect{it[1]}.ifEmpty([]))
     }