Clean up hifiasm module + add bin file input (nf-core#7802)

* Tidy hifiasm model + update tests

* Update nf.test

* Fix linting

* Add fasta and PAF outputs to hifiasm; combine all bin files into a single output channel. Update documentation

* Improve descriptiveness of output channel names

Author: prototaxites

modules/nf-core/hifiasm/main.nf

@@ -11,114 +11,100 @@ process HIFIASM {
     tuple val(meta) , path(long_reads)        , path(ul_reads)
     tuple val(meta1), path(paternal_kmer_dump), path(maternal_kmer_dump)
     tuple val(meta2), path(hic_read1)         , path(hic_read2)
+    tuple val(meta3), path(bin_files)
 
     output:
-    tuple val(meta), path("*.r_utg.gfa")       , emit: raw_unitigs
-    tuple val(meta), path("*.ec.bin")          , emit: corrected_reads
-    tuple val(meta), path("*.ovlp.source.bin") , emit: source_overlaps
-    tuple val(meta), path("*.ovlp.reverse.bin"), emit: reverse_overlaps
-    tuple val(meta), path("*.bp.p_ctg.gfa")    , emit: processed_contigs, optional: true
-    tuple val(meta), path("*.p_utg.gfa")       , emit: processed_unitigs, optional: true
-    tuple val(meta), path("*.asm.p_ctg.gfa")   , emit: primary_contigs  , optional: true
-    tuple val(meta), path("*.asm.a_ctg.gfa")   , emit: alternate_contigs, optional: true
-    tuple val(meta), path("*.hap1.p_ctg.gfa")  , emit: paternal_contigs , optional: true
-    tuple val(meta), path("*.hap2.p_ctg.gfa")  , emit: maternal_contigs , optional: true
-    tuple val(meta), path("*.log")             , emit: log
-    path  "versions.yml"                       , emit: versions
+    tuple val(meta), path("*.r_utg.gfa")                             , emit: raw_unitigs
+    tuple val(meta), path("*.bin")                                   , emit: bin_files        , optional: true
+    tuple val(meta), path("*.p_utg.gfa")                             , emit: processed_unitigs, optional: true
+    tuple val(meta), path("${prefix}.{p_ctg,bp.p_ctg,hic.p_ctg}.gfa"), emit: primary_contigs  , optional: true
+    tuple val(meta), path("${prefix}.{a_ctg,hic.a_ctg}.gfa")         , emit: alternate_contigs, optional: true
+    tuple val(meta), path("${prefix}.*.hap1.p_ctg.gfa")              , emit: hap1_contigs     , optional: true
+    tuple val(meta), path("${prefix}.*.hap2.p_ctg.gfa")              , emit: hap2_contigs     , optional: true
+    tuple val(meta), path("*.ec.fa.gz")                              , emit: corrected_reads  , optional: true
+    tuple val(meta), path("*.ovlp.paf.gz")                           , emit: read_overlaps    , optional: true
+    tuple val(meta), path("${prefix}.stderr.log")                    , emit: log
+    path  "versions.yml"                                             , emit: versions
 
     when:
     task.ext.when == null || task.ext.when
 
     script:
     def args = task.ext.args ?: ''
-    def prefix = task.ext.prefix ?: "${meta.id}"
+    prefix = task.ext.prefix ?: "${meta.id}"
 
     def long_reads_sorted = long_reads instanceof List ? long_reads.sort{ it.name } : long_reads
     def ul_reads_sorted = ul_reads instanceof List ? ul_reads.sort{ it.name } : ul_reads
     def ultralong = ul_reads ? "--ul ${ul_reads_sorted}" : ""
 
-    if ((paternal_kmer_dump) && (maternal_kmer_dump) && (hic_read1) && (hic_read2)) {
-        error "Hifiasm Trio-binning and Hi-C integrated should not be used at the same time"
-    } else if ((paternal_kmer_dump) && !(maternal_kmer_dump)) {
-        error "Hifiasm Trio-binning requires maternal data"
-    } else if (!(paternal_kmer_dump) && (maternal_kmer_dump)) {
-        error "Hifiasm Trio-binning requires paternal data"
-    } else if ((paternal_kmer_dump) && (maternal_kmer_dump)) {
-        """
-        hifiasm \\
-            $args \\
-            -o ${prefix}.asm \\
-            -t $task.cpus \\
-            -1 $paternal_kmer_dump \\
-            -2 $maternal_kmer_dump \\
-            $ultralong \\
-            $long_reads_sorted \\
-            2> >( tee ${prefix}.stderr.log >&2 )
+    if([paternal_kmer_dump, maternal_kmer_dump].any() && [hic_read1, hic_read2].any()) {
+        log.error("ERROR: hifiasm trio binning mode and Hi-C phasing can not be used at the same time.")
+    }
 
+    def input_trio = ""
+    if([paternal_kmer_dump, maternal_kmer_dump].any()) {
+        if(![paternal_kmer_dump, maternal_kmer_dump].every()) {
+            log.error("ERROR: Either the maternal or paternal kmer dump is missing!")
+        } else {
+            input_trio = "-1 ${paternal_kmer_dump} -2 ${maternal_kmer_dump}"
+        }
+    }
 
-        cat <<-END_VERSIONS > versions.yml
-        "${task.process}":
-            hifiasm: \$(hifiasm --version 2>&1)
-        END_VERSIONS
-        """
-    } else if ((hic_read1) && !(hic_read2)) {
-        error "Hifiasm Hi-C integrated requires paired-end data (only R1 specified here)"
-    } else if (!(hic_read1) && (hic_read2)) {
-        error "Hifiasm Hi-C integrated requires paired-end data (only R2 specified here)"
-    } else if ((hic_read1) && (hic_read2)) {
-        """
-        hifiasm \\
-            $args \\
-            -o ${prefix}.asm \\
-            -t $task.cpus \\
-            --h1 $hic_read1 \\
-            --h2 $hic_read2 \\
-            $ultralong \\
-            $long_reads \\
-            2> >( tee ${prefix}.stderr.log >&2 )
+    def input_hic = ""
+    if([hic_read1, hic_read2].any()) {
+        if(![hic_read1, hic_read2].every()) {
+            log.error("ERROR: Either the forward or reverse Hi-C reads are missing!")
+        } else {
+            input_hic = "--h1 ${hic_read1} --h2 ${hic_read2}"
+        }
+    }
+    """
+    hifiasm \\
+        $args \\
+        -t ${task.cpus} \\
+        ${input_trio} \\
+        ${input_hic} \\
+        ${ultralong} \\
+        -o ${prefix} \\
+        ${long_reads_sorted} \\
+        2> >( tee ${prefix}.stderr.log >&2 )
 
+    if [ -f ${prefix}.ec.fa ]; then
+        gzip ${prefix}.ec.fa
+    fi
 
-        cat <<-END_VERSIONS > versions.yml
-        "${task.process}":
-            hifiasm: \$(hifiasm --version 2>&1)
-        END_VERSIONS
-        """
-    } else { // Phasing with Hi-C data is not supported yet
-        """
-        hifiasm \\
-            $args \\
-            -o ${prefix}.asm \\
-            -t $task.cpus \\
-            $ultralong \\
-            $long_reads \\
-            2> >( tee ${prefix}.stderr.log >&2 )
+    if [ -f ${prefix}.ovlp.paf ]; then
+        gzip ${prefix}.ovlp.paf
+    fi
 
-        cat <<-END_VERSIONS > versions.yml
-        "${task.process}":
-            hifiasm: \$(hifiasm --version 2>&1)
-        END_VERSIONS
-        """
-    }
-        stub:
-        def args = task.ext.args ?: ''
-        def prefix = task.ext.prefix ?: "${meta.id}"
-        """
-        touch ${prefix}.asm.r_utg.gfa
-        touch ${prefix}.asm.ec.bin
-        touch ${prefix}.asm.ovlp.source.bin
-        touch ${prefix}.asm.ovlp.reverse.bin
-        touch ${prefix}.asm.bp.p_ctg.gfa
-        touch ${prefix}.asm.p_utg.gfa
-        touch ${prefix}.asm.p_ctg.gfa
-        touch ${prefix}.asm.a_ctg.gfa
-        touch ${prefix}.asm.hap1.p_ctg.gfa
-        touch ${prefix}.asm.hap2.p_ctg.gfa
-        touch ${prefix}.stderr.log
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        hifiasm: \$(hifiasm --version 2>&1)
+    END_VERSIONS
+    """
 
-        cat <<-END_VERSIONS > versions.yml
-        "${task.process}":
-            hifiasm: \$(hifiasm --version 2>&1)
-        END_VERSIONS
-        """
+    stub:
+    prefix = task.ext.prefix ?: "${meta.id}"
+    """
+    touch ${prefix}.r_utg.gfa
+    touch ${prefix}.ec.bin
+    touch ${prefix}.ovlp.source.bin
+    touch ${prefix}.ovlp.reverse.bin
+    touch ${prefix}.hic.tlb.bin
+    touch ${prefix}.hic.lk.bin
+    touch ${prefix}.bp.p_ctg.gfa
+    touch ${prefix}.p_utg.gfa
+    touch ${prefix}.p_ctg.gfa
+    touch ${prefix}.a_ctg.gfa
+    touch ${prefix}.hap1.p_ctg.gfa
+    touch ${prefix}.hap2.p_ctg.gfa
+    echo "" | gzip > ${prefix}.ec.fa.gz
+    echo "" | gzip > ${prefix}.ovlp.paf.gz
+    touch ${prefix}.stderr.log
 
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        hifiasm: \$(hifiasm --version 2>&1)
+    END_VERSIONS
+    """
 }