feat(galaxy): add tool wrappers for all five subcommands

.github/labeler.yml

@@ -71,6 +71,11 @@ area:packaging:
           - "src-tauri/tauri.conf.json"
           - "CITATION.cff"
 
+area:galaxy:
+  - changed-files:
+      - any-glob-to-any-file:
+          - "galaxy/**"
+
 # Everything under .github, not a list of the files there. Naming them one by one
 # would leave this file itself uncovered, so the pull request adding a rule would
 # arrive with no label, which is a small thing that says exactly how such a list

galaxy/tools/pathotypr/.shed.yml

@@ -0,0 +1,21 @@
+name: pathotypr
+owner: iuc
+description: Lineage classification and marker-driven genotyping from assemblies or raw reads
+long_description: |
+  pathotypr classifies microbial genomes into lineages and genotypes them
+  against user-defined marker panels. It works from assembled genomes (FASTA)
+  and from raw reads (FASTQ) without aligning them, and it carries no assumption
+  about the organism: the marker panel supplied defines what is typed.
+categories:
+- Sequence Analysis
+- Variant Analysis
+remote_repository_url: https://github.com/galaxyproject/tools-iuc/tree/main/tools/pathotypr
+homepage_url: https://pathogenomics-lab.github.io/pathotypr/
+type: unrestricted
+auto_tool_repositories:
+  name_template: "{{ tool_id }}"
+  description_template: "Wrapper for the pathotypr tool {{ tool_name }}."
+suite:
+  name: "suite_pathotypr"
+  description: A suite of Galaxy tools for the five pathotypr commands.
+  type: repository_suite_definition

galaxy/tools/pathotypr/README.md

@@ -0,0 +1,74 @@
+# Galaxy wrappers for pathotypr
+
+Five tools, one per subcommand, sharing `macros.xml`.
+
+| Tool | Wraps | Outputs |
+|---|---|---|
+| `pathotypr_train` | `train` | model, and optionally the two feature-importance reports |
+| `pathotypr_predict` | `predict` | predictions TSV |
+| `pathotypr_classify` | `classify` | per-marker hits, summary, optionally masked FASTA |
+| `pathotypr_split_fastq` | `split-fastq` | per-marker calls, summary |
+| `pathotypr_match` | `match` | best-match report |
+
+## Running the tests
+
+```bash
+python3 -m venv .venv && .venv/bin/pip install planemo
+.venv/bin/planemo lint tools/pathotypr/*.xml
+.venv/bin/planemo test tools/pathotypr/
+```
+
+`planemo test` resolves `pathotypr` from Bioconda. On a machine where that
+package has no build for the local platform, put a `pathotypr` binary on `PATH`
+and add `--no_dependency_resolution`; that still exercises the command lines,
+the outputs and the assertions, and leaves only the dependency resolution to CI.
+
+## The test data
+
+Everything under `test-data/` is synthetic and generated deterministically:
+a 2 kb reference, eight markers, two samples, paired reads at about 18x, and a
+twelve-genome training set in two classes. 64 KB in total.
+
+It is not filler. The generator asserts that **every marker k-mer occurs exactly
+once in the reference**, in both orientations, because a marker that matched in
+two places would make the tests pass for the wrong reason. The expected calls
+follow from how the samples were built:
+
+- `sample1` carries the L2 to L2.1 to L2.1.1 path plus the RIF marker, so
+  `classify` calls L2 and reports `rpoB` / `S450L`.
+- `sample2` carries L3 to L3.2 plus the INH marker.
+- `split-fastq` on `sample1`'s reads produces the **same summary** as `classify`
+  on its assembly, which is the property the marker format promises.
+- `predict` assigns the two held-out queries to L2 and L3.
+- `match` picks `ref_A`, which is `sample1`'s own genome.
+
+## Three things the wrappers have to work around
+
+1. **Sample names come from filenames.** Galaxy datasets arrive as
+   `dataset_NNN.dat`, so `classify` and `split-fastq` would name their outputs
+   after a Galaxy id. Each input is symlinked to a stable name first.
+2. **`match` echoes input paths into its report.** Same fix, otherwise the first
+   column is full of absolute paths that differ between instances.
+3. **`--min-alt-percent` is an integer.** A Galaxy `float` parameter renders
+   `95.0`, which the tool rejects outright. It is declared as an integer, so
+   fractional thresholds such as 99.5% are not available.
+
+## Submitting to the IUC
+
+The directory layout matches `galaxyproject/tools-iuc`, so submission is a copy:
+
+```bash
+cp -r galaxy/tools/pathotypr <tools-iuc>/tools/pathotypr
+```
+
+then a pull request against that repository. Their CI runs the same planemo
+lint and tests on Linux, where the Bioconda package resolves.
+
+## Not done yet
+
+Marker panels and models come from the history. A **data manager** plus a
+`.loc` table would let an administrator install the published MTBC panels once
+so users pick them from a dropdown, which is how Galaxy normally handles
+reference data. That is a separate piece of work and was left out deliberately
+rather than half-built: shipping a data table with nothing to populate it moves
+the burden to administrators without helping anyone.

galaxy/tools/pathotypr/macros.xml

@@ -0,0 +1,85 @@
+<macros>
+    <token name="@TOOL_VERSION@">1.0.2</token>
+    <token name="@VERSION_SUFFIX@">0</token>
+    <token name="@PROFILE@">23.0</token>
+
+    <xml name="requirements">
+        <requirements>
+            <requirement type="package" version="@TOOL_VERSION@">pathotypr</requirement>
+        </requirements>
+    </xml>
+
+    <xml name="xrefs">
+        <xrefs>
+            <xref type="bio.tools">pathotypr</xref>
+        </xrefs>
+    </xml>
+
+    <xml name="version_command">
+        <version_command>pathotypr --version</version_command>
+    </xml>
+
+    <xml name="citations">
+        <citations>
+            <citation type="doi">10.64898/2026.03.24.714002</citation>
+            <citation type="doi">10.5281/zenodo.19210044</citation>
+        </citations>
+    </xml>
+
+    <!-- Threads. Galaxy sets GALAXY_SLOTS; without it a tool helps itself to
+         every core on a shared node. -->
+    <token name="@THREADS@">-t \${GALAXY_SLOTS:-1}</token>
+
+    <xml name="kmer_size" token_default="31" token_help="">
+        <param argument="--kmer-size" type="integer" min="1" max="31" value="@DEFAULT@"
+               label="k-mer size"
+               help="Must be between 1 and 31. An odd value centres the variant inside the k-mer, so 31 and 21 are the usual choices. @HELP@"/>
+    </xml>
+
+    <!-- Marker panels and the reference they were called against. They travel
+         together on purpose: positions are 1-based coordinates on that exact
+         sequence, and a panel read against a different assembly produces
+         confident, wrong calls with no warning. -->
+    <xml name="markers_and_reference">
+        <param argument="--markers" type="data" format="tabular" label="Marker panel"
+               help="Tab-separated: position, REF, ALT, then one lineage level per column. The published MTBC panels are on Zenodo, doi:10.5281/zenodo.19210044."/>
+        <param argument="--reference" type="data" format="fasta" label="Reference genome"
+               help="Single-record FASTA. This must be the assembly the marker positions were called against."/>
+    </xml>
+
+    <xml name="nested_classification">
+        <param argument="--nested-classification" type="boolean" truevalue="--nested-classification"
+               falsevalue="" checked="false" label="Nested lineage classification"
+               help="Requires a panel whose lineage hierarchy is spread across one column per level."/>
+    </xml>
+
+    <token name="@MARKER_HELP@">
+**Marker file format**
+
+Tab-separated, no header required. Column order:
+
+::
+
+    position  REF  ALT  level1  level2  ...  (empty)  gene  mutation
+
+Positions are 1-based on the reference genome you supply. The lineage hierarchy
+runs from column 4 until the first empty cell; whatever follows that empty cell
+is read as gene and mutation. SNPs and equal-length MNVs work everywhere;
+indels are accepted by *classify* and skipped by *split-fastq*.
+
+The full rules are at
+https://pathogenomics-lab.github.io/pathotypr/marker_format/
+
+Ready-made panels for the *M. tuberculosis* complex, and a pre-trained model,
+are published on Zenodo: https://doi.org/10.5281/zenodo.19210044
+    </token>
+
+    <token name="@ORGANISM_NOTE@">
+**A note on other organisms**
+
+Nothing in pathotypr is hard-coded to one organism: the panel you supply defines
+what is typed. In practice it has only been validated on the *M. tuberculosis*
+complex, so treat other organisms as exploratory and check the calls against a
+truth set you trust.
+    </token>
+</macros>

galaxy/tools/pathotypr/pathotypr_classify.xml

@@ -0,0 +1,160 @@
+<tool id="pathotypr_classify" name="pathotypr classify" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
+    <description>genotypes assembled genomes against a marker panel</description>
+    <macros>
+        <import>macros.xml</import>
+    </macros>
+    <expand macro="xrefs"/>
+    <expand macro="requirements"/>
+    <expand macro="version_command"/>
+    <command detect_errors="exit_code"><![CDATA[
+## Galaxy datasets arrive as dataset_NNN.dat, and pathotypr takes the sample
+## name from the filename. Link each one to the element identifier so the
+## output names the sample rather than a Galaxy id.
+## The extension is stripped before it is re-added, or an identifier that
+## already ends in .fasta becomes sample.fasta.fasta, and that name is what the
+## summary reports back in the genome column.
+#import re
+#set $safe = lambda s: re.sub('[.](fasta|fa|fna|fas)$', '', re.sub('[^A-Za-z0-9_.-]', '_', str(s)))
+#for $i, $g in enumerate($genomes)
+    ln -s '$g' '${safe($g.element_identifier)}.fasta' &&
+#end for
+#if $annotation.use_gff == 'yes'
+    #for $i, $f in enumerate($annotation.gff_files)
+        ln -s '$f' '${safe($f.element_identifier)}.gff' &&
+    #end for
+#end if
+
+pathotypr classify
+    --markers '$markers'
+    --reference '$reference'
+    --input-files
+#for $g in $genomes
+    '${safe($g.element_identifier)}.fasta'
+#end for
+#if $annotation.use_gff == 'yes'
+    --gff-files
+    #for $f in $annotation.gff_files
+        '${safe($f.element_identifier)}.gff'
+    #end for
+#end if
+    --output-prefix 'result'
+    --kmer-size $kmer_size
+    --min-flank-bases $min_flank_bases
+    $nested_classification
+    $output_masked_fasta
+    @THREADS@
+    ]]></command>
+    <inputs>
+        <param name="genomes" type="data" format="fasta" multiple="true" label="Assembled genomes"
+               help="One or more FASTA assemblies to genotype."/>
+        <expand macro="markers_and_reference"/>
+        <conditional name="annotation">
+            <param name="use_gff" type="select" label="Annotate hits with a GFF">
+                <option value="no" selected="true">No</option>
+                <option value="yes">Yes</option>
+            </param>
+            <when value="no"/>
+            <when value="yes">
+                <param name="gff_files" type="data" format="gff,gff3" multiple="true" label="GFF annotation"
+                       help="Matched to the genomes by filename, so the names must correspond."/>
+            </when>
+        </conditional>
+        <expand macro="kmer_size" default="31"/>
+        <param argument="--min-flank-bases" type="integer" min="0" value="10"
+               label="Minimum flanking bases"
+               help="Bases required on each side of the allele inside the marker k-mer. At the defaults this caps usable alleles at 11 bp."/>
+        <expand macro="nested_classification"/>
+        <param argument="--output-masked-fasta" type="boolean" truevalue="--output-masked-fasta"
+               falsevalue="" checked="false" label="Also write masked FASTA"
+               help="Writes each genome with the marker positions replaced by N."/>
+    </inputs>
+    <outputs>
+        <data name="detailed" format="tabular" label="${tool.name} on ${on_string}: per-marker hits"
+              from_work_dir="result.tsv"/>
+        <data name="summary" format="tabular" label="${tool.name} on ${on_string}: summary"
+              from_work_dir="result_summary.tsv"/>
+        <collection name="masked" type="list" label="${tool.name} on ${on_string}: masked FASTA">
+            <filter>output_masked_fasta</filter>
+            <discover_datasets pattern="(?P&lt;designation&gt;.+)_masked\.fasta" format="fasta" directory="."/>
+        </collection>
+    </outputs>
+    <tests>
+        <!-- One genome carrying the L2 -> L2.1 -> L2.1.1 path and a RIF marker. -->
+        <test expect_num_outputs="2">
+            <param name="genomes" value="sample1.fasta" ftype="fasta"/>
+            <param name="markers" value="markers.tsv" ftype="tabular"/>
+            <param name="reference" value="reference.fasta" ftype="fasta"/>
+            <output name="summary">
+                <assert_contents>
+                    <has_text text="major_lineage"/>
+                    <has_text text="L2;L2.1;L2.1.1"/>
+                </assert_contents>
+            </output>
+            <output name="detailed">
+                <assert_contents>
+                    <has_text text="rpoB"/>
+                    <has_text text="S450L"/>
+                    <has_n_lines n="5"/>
+                </assert_contents>
+            </output>
+        </test>
+        <!-- Two genomes at once, and the second one is a different lineage. -->
+        <test expect_num_outputs="2">
+            <param name="genomes" value="sample1.fasta,sample2.fasta" ftype="fasta"/>
+            <param name="markers" value="markers.tsv" ftype="tabular"/>
+            <param name="reference" value="reference.fasta" ftype="fasta"/>
+            <param name="kmer_size" value="21"/>
+            <output name="summary">
+                <assert_contents>
+                    <has_text text="L2"/>
+                    <has_text text="L3"/>
+                    <has_n_lines n="3"/>
+                </assert_contents>
+            </output>
+        </test>
+        <!-- The masked FASTA is a discovered collection, so its naming deserves
+             a test of its own. The input genomes contain no N at all, so an N in
+             the output can only have come from masking. -->
+        <test expect_num_outputs="3">
+            <param name="genomes" value="sample1.fasta" ftype="fasta"/>
+            <param name="markers" value="markers.tsv" ftype="tabular"/>
+            <param name="reference" value="reference.fasta" ftype="fasta"/>
+            <param name="output_masked_fasta" value="true"/>
+            <output_collection name="masked" type="list" count="1">
+                <element name="sample1" ftype="fasta">
+                    <assert_contents>
+                        <has_text text="&gt;sample1"/>
+                        <has_text text="N"/>
+                    </assert_contents>
+                </element>
+            </output_collection>
+        </test>
+    </tests>
+    <help><![CDATA[
+**What it does**
+
+`pathotypr classify` calls a panel of diagnostic variants directly on assembled
+genomes. For each marker it builds a k-mer centred on the variant, looks for it
+in the assembly, and reports which allele was found. Hits are then collapsed
+into a lineage call per genome.
+
+It produces two tables:
+
+- **per-marker hits**: one row per marker found, with the k-mer, its position,
+  the reference and alternate alleles, the lineage the marker belongs to and,
+  where the panel carries them, the gene and amino-acid change.
+- **summary**: one row per genome, with the count of hits per lineage and the
+  major lineage call.
+
+@MARKER_HELP@
+
+@ORGANISM_NOTE@
+
+**Choosing between classify and split-fastq**
+
+Use this tool when you have assemblies. Use *pathotypr split-fastq* when you
+have raw reads and would rather not assemble them. Both read the same panel and
+agree on the call; only the input differs.
+    ]]></help>
+    <expand macro="citations"/>
+</tool>