Review on #2035 Adding test files for new sourmash tests Reviewed and changes requested by pinin4fjords on erikrikarddaniel's pull request A few things to address: File sizes break the existing FASTQ convention on this branch The seven new files weigh ~71 MB, dominated by the FASTQs at 9-13 MB each. Every existing sourmash/magmap FASTQ on this branch (

s003_genomes.reads_1000_*

,

s009_genomes.reads_1000_*

,

e_coli.reads_1000_*

,

s0_*

,

s1_*

,

s2_*

) is 15-35 KB, sub-sampled to 1000 reads as the filename declares. The new reads are ~400x larger than that and would become the biggest files on the entire branch by ~2x (the current top is

e_coli.fna

at 5.1 MB). This also lines up with the broader nf-core test-data pushback I've had from @mashehu on other branches (e.g. #1979) - "<100kb if possible?", "smaller files means faster tests", and a request to rewrite history so the larger intermediate versions don't get baked in. Could you sub-sample in line with the

_reads_1000_

convention (or

_reads_10000_

if 1000 turns out to be too aggressive for the sourmash test you're targeting), and rewrite the branch history so the full-size blobs don't end up in the repo? The 10-genome SBT at ~550 KB is fine, it's just the FASTQs. Naming nits Two small inconsistencies vs the existing files on this branch: • Samplesheet:

sourmash_10genomes.samples.csv

uses a dot before

samples

, but every other samplesheet on

magmap

uses underscores (

sourmash_samples.csv

,

sourmash_mix_samples.csv

,

samplesheet_fulltest.csv

). Suggest

sourmash_10genomes_samples.csv

. • FASTQs:

.sample00.r1.fastq.gz

(dots, lowercase

r1

) vs the existing

_R1.fastq.gz

style on the paired sourmash data (

s003_genomes.reads_1000_R1.fastq.gz

etc.). Suggest matching that. The SBT name (

sourmash_10genomes.k31.index.sbt.zip

) is consistent with

sourmash_24genomes.k31.index.sbt.zip

already, no change needed there. URLs in the samplesheet are correctly nf-core hosted, also good. PR body Could you add a short description of what test in nf-core/magmap consumes these files and how the reads + SBT were generated? It'll save the next person re-deriving the provenance. nf-core/test-datasets

#2035 Adding test files for new sourmash tests Pull request closed by erikrikarddaniel nf-core/test-datasets