ALE should not be executed with long read data

[jfy133]

Description of the bug

The default ALE memory request appears to be too low, both as seen in the v5.4 AWS megatests (we failed repeatedly with exit code 139) and as reported during manual testing by @amizeranschi who said they (think) their minimum default was 64.GB

Update: actually it's because it fails on long-read assemblies.

Threfore the fisrt bit of this code for filtering out the long read samples is not working

mag/workflows/mag.nf

Lines 277 to 295 in b52fa53

	if (!params.skip_ale) {
	ch_shortread_assemblies_for_ale = ch_assemblies.filter { meta, _assembly ->
	meta.sr_platform != null && meta.sr_platform != []
	}
	ch_ale_input = BINNING_PREPARATION.out.grouped_mappings
	.join(ch_shortread_assemblies_for_ale, by: 0)
	.map { meta, _contigs, bams, _bais, assembly ->
	// Try to find the BAM where reads came from the same sample as the assembly (co-binning may include multiple BAMs)
	// If none matches (coassembly), take the first one after sorting for determinism
	def own_bam = bams.find { bam -> bam.name.endsWith("-${meta.id}.bam") }
	def bam = own_bam ?: bams.sort()[0]
	[meta, assembly, bam]
	}
	ALE(ch_ale_input)
	ch_versions = ch_versions.mix(ALE.out.versions.ifEmpty([]))
	}

Command used and terminal output

Relevant files

No response

System information

No response

[dialvarezs]

Fixed!