Logan Proteins minor fix

[SaimMomin12]

FOR CONTRIBUTOR:

• I have read the CONTRIBUTING.md document and this tool is appropriate for the tools-iuc repo.
• License permits unrestricted use (educational + commercial)
• This PR adds a new tool or tool collection
• This PR updates an existing tool or tool collection
• This PR does something else (explain below)

There are two labels that allow to ignore specific (false positive) tool linter errors:

• skip-version-check: Use it if only a subset of the tools has been updated in a suite.
• skip-url-check: Use it if github CI sees 403 errors, but the URLs work.

[bgruening]

Thanks for the fix! The three changes here are correct and well-scoped:

• .shed.yml URL typo (search_proteins → logan_proteins) ✅
• @VERSION_SUFFIX@ bump ✅
• FASTA-extension symlink in search_database.xml ✅

While reviewing I went through the rest of the wrapper and found a few things worth folding in (or addressing in a quick follow-up). Requesting changes mostly because of #1 and #2 below — they're closely related to the fix in this PR and the same class of bug.

Most likely needs to land here

1. embed_query.xml very likely has the same FASTA-extension bug this PR fixes for search_database.xml.
   At embed_query.xml line 14–15, --query_sequences '$query_sequences' is passed the raw Galaxy .dat path. If embed_query.py infers FASTA from the suffix (which the sibling fix strongly implies), embedding will fail on real datasets the same way search_database.py did. Suggest applying the same ln -s '$query_sequences' query_sequences.fasta pattern here.

2. embed_query.xml help text describes a UI toggle that doesn't exist.
   Line 55 says "Force CPU usage: By default, CPU will be used. Disable this option to use GPU instead if available." — but the only mechanism is the GALAXY_LOGAN_PROTEIN_SEARCH_FORCE_CPU env var read at lines 8–12. There's no <param> exposing this. Either add a boolean param, or reword the help so users know this is an admin/job-destination setting.

Worth fixing while you're in here (cheap)

3. tool-data/faiss_database.loc.sample header is boilerplate from another tool — it claims the file is for "metagenomics files." Replace with text that actually describes the FAISS protein database loc format. Also align the example row with the canonical 4-column format used in test-data/faiss_database.loc (faiss-demo-db-20260203\tFAISS Test Database\t1.2.0\t/path); the current sample row uses a different value style.

4. test-data/queries.fasta contains a nucleotide sequence labeled test_protein_2. Line 4 starts ATGGCTAGCAAAGGAGAAGAA… (looks like GFP CDS). It's labeled as a protein but isn't one. Either rename the record to indicate it's a nucleotide test case, or replace it with a real protein sequence — otherwise it's a copy-paste footgun for users using this as a reference.

5. outfmt param in search_database.xml is marked optional="true" but always has a default. Lines 31–38 set value="0" selected="true", so the #if $outfmt: guard at line 17 is effectively dead code — 0 is always passed. Either drop optional="true" or remove the default if "unset" should mean "use upstream default."

6. -F flag in embed_query.xml line 17 is undocumented. A short trailing comment (<!-- force overwrite of existing output files --> or whatever it actually means) would help future maintainers.

Follow-up PR material (don't block this one)

7. Test coverage is thin. The search_database test at lines 47–57 only asserts "No results, exiting" because the bundled FAISS DB at test-data/test-db/faiss/db.txt is a 0-byte stub. So we currently verify the script bails gracefully on an empty index — not that the FAISS+MMseqs2 pipeline actually returns matches. A small real index that produces ≥1 hit would substantially raise the safety floor; today, a regression in the actual search logic could land green.

8. macros.xml citations are thin. Only the upstream search_protein GitHub repo is cited. Please add at minimum:

   • FAISS — Johnson et al., 10.1109/TBDATA.2019.2921572
   • MMseqs2 — Steinegger & Söding, 10.1038/nbt.3988
   • The GLM2 / gLM2 model citation (whichever paper the container actually uses)

9. Docker-only requirements. macros.xml lines 5–9 declare only quay.io/bgruening/logan-protein:1.2.0 — no <requirement type="package">. Galaxy instances without Docker support can't install this tool. Likely unavoidable given the GLM2 + torch + FAISS + MMseqs2 stack, but worth a one-liner in .shed.yml / help so admins know what to expect.

10. search_database.xml help mentions intermediate files that aren't exposed as outputs (lines 94–99: query_results.tsv, unique_centroids.fasta, matches.top_hit, …). Either expose the useful ones as <data> outputs, or trim the help so users aren't told about files they can't access.

11. Suite naming inconsistency — directory is logan_proteins (plural), suite is suite_logan_protein (singular), tool IDs use logan_protein_* (singular). Tool IDs are stable so don't change them; just flagging for new files.

Happy to approve once #1 and #2 are addressed (or with a quick justification that #1 isn't actually a problem). Everything else can be its own PR.