Add nf-test barebone configs by dialvarezs · Pull Request #855 · nf-core/mag

dialvarezs · 2025-08-25T15:27:25Z

According to #501 (comment), I added configs for the following profiles:

longreadonly
longreadonly_alternatives
hybrid
assembly_input

I also removed old configs.

PR checklist

jfy133

LGTM so far, except for that single comment below.

I will now try to run each test separately and check all the expected tools execute

jfy133 · 2025-09-05T11:02:05Z

Example of how I'm testing:

nextflow run ../main.nf -profile test_longreadonly,apptainer --outdir ./results
grep 'Submitted process' .nextflow.log | cut -d ' ' -f 13 | sort | uniq | rev | cut -d ':' -f 1 | rev | sort
## and
cat results/pipeline_info/nf_core_mag_software_mqc_versions.yml

Compare list to #501

(having to switch to run on a cluster now unfortuantely)

jfy133 · 2025-09-05T11:06:22Z

✅ longreadonly: has all expected processes executed, and versions represented
✅ longreadonly_alternatives: has all expected processes executed, and versions represented
✅ hybrid: has all expected processes executed, and versions represented, including the METASPADESHYBRID variant of SPAdes
❓ assembly_input:
- At testing, missing GUNC (but see latest commit)
- Missing MMSEQS: this runs if you specify a mmseqs db name via --metaeuk_mmseqs_db Kalamari, the smallest to my knowledge is kalamari, maybe we can try that?

…est-configs

jfy133 · 2025-09-08T11:13:07Z

Yeah, so adding the Kalamari then executiosn MMSEQS_DATABASES 👍

dialvarezs · 2025-09-12T13:06:36Z

Even Kalamari seems to be too big for the CI. It stopped after 1h40min.

prototaxites · 2025-09-12T13:11:39Z

Can you just pass a FASTA to metaeuk_db for testing? The MetaEuk test uses params.modules_testdata_base_path + 'proteomics/database/yeast_UPS.fasta

https://github.com/nf-core/modules/blob/e4760cac1fcad979e19c00da5f10573be91358a7/modules/nf-core/metaeuk/easypredict/tests/main.nf.test#L4

prototaxites · 2025-09-12T13:22:01Z

Possibly also tuning the metaeuk args:

-s: sensitivity, 1 is fastest
--max-seqs: Maximum results per query passing prefilter - default is 300, maybe try 50?

There's a bunch of other options that might be worth setting, too - see metaeuk easy-predict for the list.

dialvarezs · 2025-09-12T13:22:47Z

@prototaxites Yes, we were using that until now, but we decided to try a full db. I will try tunning the parameters, thanks for the suggestion!

jfy133 · 2025-09-13T10:05:53Z

Does giving the FASTA still execute the MMSEQS module though (my main concern)?

dialvarezs · 2025-09-13T13:42:14Z

@jfy133 No, it doesn't. The fasta is used directly by METAEUK_EASYPREDICT.
I tried several options to metauk to try to reduce the runtime, but didn't work.
I reverted the parameter to use the FASTA for now.

prototaxites · 2025-09-13T16:54:57Z

If you're open to a slight hack to test the process, we could modify the code so if metaeuk_mmseqs_db is set it always downloads a DB, even if skip_metaeuk is true. That way we can test downloading without running the test with the db...

edit: Or alternatively, as that might not be desirable - set task.ext.when = false in a test config for MetaEuk... Future-proof for as long as when: still supported.

jfy133 · 2025-09-15T10:34:03Z

OK so I think we can say:

GUNC is broken, so we have to skip for now.
MetaEuk runs too long if we use Kalamari (so we just have to skip testing MMSEqs). At least we have the main tool.

Otherwise we have to rely on user reports.

prototaxites · 2025-09-15T10:45:49Z

Kalamari is not actually the smallest MMSeqs2 database available - UniProtKB/Swiss-Prot is significantly smaller. The actual sequence files for SwissProt come to just 64Mb... the rest of the data is taxonomy, which I suspect will not significantly impact run time. Kalamari requires 633Mb of sequence files.

So it might be worth trying one last time with Swissprot...!

❱ du -h --max-depth 1 .                                                                                                                                                                                                                                               
611M	./swissprot
964M	./kalamari

❱ ll swissprot/                                                                                                                                                                                                                                                       
total 611M
-rw-r--r-- 1 jd42 team311 200M Sep 15 11:41 swissprot
-rw-r--r-- 1 jd42 team311    4 Sep 15 11:41 swissprot.dbtype
-rw-r--r-- 1 jd42 team311  72M Sep 15 11:41 swissprot_h
-rw-r--r-- 1 jd42 team311    4 Sep 15 11:41 swissprot_h.dbtype
-rw-r--r-- 1 jd42 team311  11M Sep 15 11:41 swissprot_h.index
-rw-r--r-- 1 jd42 team311  12M Sep 15 11:41 swissprot.index
-rw-r--r-- 1 jd42 team311 8.7M Sep 15 11:41 swissprot.lookup
-rw-r--r-- 1 jd42 team311 7.3M Sep 15 11:41 swissprot_mapping
-rw-r--r-- 1 jd42 team311   25 Sep 15 11:41 swissprot.source
-rw-r--r-- 1 jd42 team311 710M Sep 15 11:41 swissprot_taxonomy
-rw-r--r-- 1 jd42 team311  151 Sep 15 11:41 swissprot.version

jfy133

OK, I pushed one minor fix (easy to edit in GitHub)

@dialvarezs if you could re-run the snapshots to fix that one null version in the snapshots for ADJUST_MAXBIN_EXT (codespaces wouldn't allow me to push 😢 ), we are good to go for this PR...

and we can open the release PR for 5.0.0!

(Which is PERFECT timing, as this is something I can report in a grant progress update in 2 weeks 😅)

jfy133 · 2025-09-15T10:48:58Z

Kalamari is not actually the smallest MMSeqs2 database available - UniProtKB/Swiss-Prot is significantly smaller. The actual sequence files for SwissProt come to just 64Mb... the rest of the data is taxonomy, which I suspect will not significantly impact run time. Kalamari requires 633Mb of sequence files.

So it might be worth trying one last time with Swissprot...!

❱ du -h --max-depth 1 .                                                                                                                                                                                                                                               
611M	./swissprot
964M	./kalamari

❱ ll swissprot/                                                                                                                                                                                                                                                       
total 611M
-rw-r--r-- 1 jd42 team311 200M Sep 15 11:41 swissprot
-rw-r--r-- 1 jd42 team311    4 Sep 15 11:41 swissprot.dbtype
-rw-r--r-- 1 jd42 team311  72M Sep 15 11:41 swissprot_h
-rw-r--r-- 1 jd42 team311    4 Sep 15 11:41 swissprot_h.dbtype
-rw-r--r-- 1 jd42 team311  11M Sep 15 11:41 swissprot_h.index
-rw-r--r-- 1 jd42 team311  12M Sep 15 11:41 swissprot.index
-rw-r--r-- 1 jd42 team311 8.7M Sep 15 11:41 swissprot.lookup
-rw-r--r-- 1 jd42 team311 7.3M Sep 15 11:41 swissprot_mapping
-rw-r--r-- 1 jd42 team311   25 Sep 15 11:41 swissprot.source
-rw-r--r-- 1 jd42 team311 710M Sep 15 11:41 swissprot_taxonomy
-rw-r--r-- 1 jd42 team311  151 Sep 15 11:41 swissprot.version

Oh wow! OK ! Yeah might be worht trying!

Co-authored-by: James A. Fellows Yates <[email protected]>

dialvarezs · 2025-09-16T22:12:53Z

Ok, I managed to get it running. Using swissprot definitely helped a bit but not enough. What really did the trick was to reduce the number of bins ran with metaeuk setting postbinning_input to refined_bins_only.

jfy133

Whoop whoops! finally!

jfy133 · 2025-09-17T07:00:24Z

Thank you @dialvarezs ! Merging!

dialvarezs requested review from d4straub, jfy133, muabnezor and prototaxites as code owners August 25, 2025 15:27

dialvarezs added 3 commits August 25, 2025 12:31

Add new nf-test configs

b378fbf

Old config cleanup

2431e56

Update changelog

e1be7a7

dialvarezs changed the title ~~nf-test barebones configs~~ nf-test barebone configs Aug 25, 2025

dialvarezs changed the title ~~nf-test barebone configs~~ Add nf-test barebone configs Aug 25, 2025

dialvarezs and others added 11 commits August 26, 2025 02:09

Merge remote-tracking branch 'upstream/dev' into nf-test-configs

c335a41

Update snapshots

22214cd

Add multiqc snapshots

0c48e6e

Update hybrid

b86c379

Update snapshots

75055a4

Update snapshot

43d7d49

Standardise header sand tags

3172bf0

Use porechop_porechop in test_hbyrid profile

7c2cca1

update snapshot to include correct tool

94b2658

Add assembly input nf-test (WIP)

5eb6e09

Add metaeuk

799f3e7

jfy133 reviewed Sep 5, 2025

View reviewed changes

Comment thread conf/test_assembly_input.config Outdated

dialvarezs added 4 commits September 5, 2025 09:39

Use large disk to prevent storage issues

c060585

Merge branch 'nf-test-configs' of github.com:dialvarezs/mag into nf-t…

53d50bc

…est-configs

Merge remote-tracking branch 'upstream/dev' into nf-test-configs

b7415b9

Enable GUNC

137bb60

Update assembly_input test

16759ff

Update snapshots (nextflow version)

f014dbc

dialvarezs force-pushed the nf-test-configs branch from 2c1a6a7 to f014dbc Compare September 12, 2025 13:32

Update metaeuk config

e6e7b0d

Revert to fasta input for metaeuk

ed8af48

dialvarezs added 2 commits September 13, 2025 10:43

Update changelog

07862c7

Fix nextflow version

22e67aa

Fix version pattern capture

c5ab8bd

jfy133 approved these changes Sep 15, 2025

View reviewed changes

Comment thread CHANGELOG.md Outdated

Comment thread conf/test_assembly_input.config

dialvarezs and others added 3 commits September 16, 2025 17:00

Enable metauk using swissprot + refined bins only

9f8ef9a

Update CHANGELOG.md

439a02c

Co-authored-by: James A. Fellows Yates <[email protected]>

Update snapshots

81bc495

jfy133 approved these changes Sep 17, 2025

View reviewed changes

jfy133 merged commit 85c1a4b into nf-core:dev Sep 17, 2025
20 checks passed

jfy133 mentioned this pull request Sep 17, 2025

Config restructure #621

Closed

22 tasks

dialvarezs mentioned this pull request Apr 20, 2026

Update GUNC test db to use official one nf-core/test-datasets#1999

Merged

Conversation

dialvarezs commented Aug 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR checklist

Uh oh!

jfy133 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

jfy133 commented Sep 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jfy133 commented Sep 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jfy133 commented Sep 8, 2025

Uh oh!

dialvarezs commented Sep 12, 2025

Uh oh!

prototaxites commented Sep 12, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

prototaxites commented Sep 12, 2025

Uh oh!

dialvarezs commented Sep 12, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jfy133 commented Sep 13, 2025

Uh oh!

dialvarezs commented Sep 13, 2025

Uh oh!

prototaxites commented Sep 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jfy133 commented Sep 15, 2025

Uh oh!

prototaxites commented Sep 15, 2025

Uh oh!

jfy133 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

jfy133 commented Sep 15, 2025

Uh oh!

dialvarezs commented Sep 16, 2025

Uh oh!

jfy133 left a comment

Choose a reason for hiding this comment

Uh oh!

jfy133 commented Sep 17, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

dialvarezs commented Aug 25, 2025 •

edited

Loading

jfy133 commented Sep 5, 2025 •

edited

Loading

jfy133 commented Sep 5, 2025 •

edited

Loading

prototaxites commented Sep 12, 2025 •

edited

Loading

dialvarezs commented Sep 12, 2025 •

edited

Loading

prototaxites commented Sep 13, 2025 •

edited

Loading