Merge pull request #268 from phac-nml/sistr

Sistr v1.1.3 new Galaxy wrapper release
phac-nml · Dec 20, 2024 · 4a9d0e7 · 4a9d0e7
2 parents f574c20 + c30cd28
commit 4a9d0e7
Show file tree

Hide file tree

Showing 17 changed files with 84 additions and 879 deletions.
diff --git a/.github/workflows/pr.yaml b/.github/workflows/pr.yaml
@@ -36,7 +36,7 @@ jobs:
   # which are needed in subsequent steps.
   setup:
     name: Setup cache and determine changed repositories
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     outputs:
       galaxy-head-sha: ${{ steps.get-galaxy-sha.outputs.galaxy-head-sha }}
       repository-list: ${{ steps.discover.outputs.repository-list }}
@@ -108,7 +108,7 @@ jobs:
     name: Lint tool-list
     needs: setup
     if: ${{ needs.setup.outputs.repository-list != '' || needs.setup.outputs.tool-list != '' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     strategy:
       fail-fast: false
       matrix:
@@ -154,7 +154,7 @@ jobs:
     name: Lint Python scripts
     needs: setup
     if: ${{ github.event_name == 'pull_request' && needs.setup.outputs.repository-list != '' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     strategy:
       fail-fast: false
       matrix:
@@ -245,7 +245,7 @@ jobs:
     name: Check file sizes
     needs: setup
     if: ${{ github.event_name == 'pull_request' && needs.setup.outputs.repository-list != '' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     steps:
     - uses: actions/checkout@v3
       with:
@@ -274,7 +274,7 @@ jobs:
     name: Test tools
     needs: setup
     if: ${{ needs.setup.outputs.repository-list != '' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     strategy:
       fail-fast: false
       matrix:
@@ -340,7 +340,7 @@ jobs:
     name: Combine chunked test results
     needs: [setup, test]
     if: ${{ always() && needs.setup.outputs.repository-list != '' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     strategy:
       matrix:
         python-version: ['3.7']
@@ -384,7 +384,7 @@ jobs:
     name: Deploy
     needs: [setup, lint, combine_outputs]
     if: ${{ (github.ref == 'refs/heads/master' || github.ref == 'refs/heads/main' ) && github.repository_owner == 'phac-nml' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     strategy:
       matrix:
         python-version: ['3.7']
@@ -421,7 +421,7 @@ jobs:
     name: Report deploy status
     needs: [deploy]
     if: ${{ always() && needs.deploy.result != 'success' && (github.ref == 'refs/heads/master' || github.ref == 'refs/heads/main' ) && github.repository_owner == 'phac-nml' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     steps:
     # report to the PR if deployment failed
     - name: Get PR object
@@ -443,7 +443,7 @@ jobs:
     name: Check workflow success
     needs: [setup, lint, flake8, lintr, file_sizes, combine_outputs]
     if: ${{ always() && github.ref != 'refs/heads/master' && github.ref != 'refs/heads/main' }}
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-20.04
     steps:
     - name: Check tool lint status
       if: ${{ needs.lint.result != 'success' && needs.flake8.result != 'skipped' }}

diff --git a/tools/sistr_cmd/.shed.yml b/tools/sistr_cmd/.shed.yml
@@ -2,4 +2,6 @@ categories: ["Sequence Analysis"]
 description: SISTR in silico serotyping tool
 name: sistr_cmd
 owner: nml
+long_description: Salmonella In Silico Typing Resource (SISTR) command-line tool for serovar prediction
+remote_repository_url: https://github.com/phac-nml/sistr_cmd
 homepage_url: https://github.com/phac-nml/sistr_cmd
diff --git a/tools/sistr_cmd/sistr_cmd.xml b/tools/sistr_cmd/sistr_cmd.xml
@@ -1,17 +1,21 @@
-<tool id="sistr_cmd" name="sistr_cmd" version="@VERSION@+galaxy1">
+<tool id="sistr_cmd" name="sistr_cmd" version="@VERSION@+galaxy0">
   <description>
-    Salmonella In Silico Typing Resource commandline tool for serovar prediction 
+    Salmonella In Silico Typing Resource command-line tool for serovar prediction 
   </description>
   <macros>
-     <token name="@VERSION@">1.1.1</token>
+     <token name="@VERSION@">1.1.3</token>
   </macros>
+  <xrefs>
+      <xref type="bio.tools">SISTR</xref>
+  </xrefs>
   <requirements>
     <requirement type="package" version="@VERSION@">sistr_cmd</requirement>
   </requirements>
   <stdio>
     <exit_code range="1:" />
   </stdio>
   <command><![CDATA[
+
   sistr 
     #for $fasta in $input_fastas
       -i '$fasta' '${$fasta.element_identifier.replace("." + $fasta.ext, "")}'
@@ -25,6 +29,11 @@
     $no_cgmlst
     $run_mash
     $qc
+    #if $serovar_list.selection_mode == "default"
+      --list-of-serovars
+    #elif  $serovar_list.selection_mode == "custom"
+      --list-of-serovars  $serovar_list.custom_list_of_serovars_file
+    #end if
     --threads "\${GALAXY_SLOTS:-1}"
     -T "\${TMPDIR:-/tmp}"
     $keep_tmp
@@ -86,7 +95,21 @@
       truevalue="--qc"
       falsevalue=""
       label="Basic QC of results"
-      />
+    />
+    <conditional name="serovar_list">
+        <param label="Check predicted SISTR serovar against the list of serovars?" name="selection_mode" type="select">
+          <option value="none">Not selected</option>
+          <option selected="true" value="default">Use the default list</option>
+          <option value="custom">Use a custom list</option>
+        </param>
+        <when value="none"/>
+        <when value="default"/>
+        <when value="custom">
+          <param name="custom_list_of_serovars_file" type="data" format="text" optional="true" label="Custom list of serovars (optional)" 
+        help="if no file selected, the default one is used available at https://raw.githubusercontent.com/phac-nml/sistr_cmd/v1.1.3/sistr/data/serovar-list.txt"
+        />
+        </when>
+    </conditional>
     <param 
       name="more_output"
       type="select" 
@@ -131,63 +154,64 @@
     <data 
       name="output_prediction_csv" 
       format="csv" 
-      label="SISTR Results"
+      label="${tool.name} on ${input_fastas[0].element_identifier}: Main Report (csv)"
       from_work_dir="sistr-report.csv">
       <filter>output_format == "csv"</filter>
     </data>
     <data 
       name="output_prediction_json" 
       format="json" 
-      label="SISTR Results"
+      label="${tool.name} on ${input_fastas[0].element_identifier}: Main Report (json)"
       from_work_dir="sistr-report.json">
       <filter>output_format == "json"</filter>
     </data>
     <data 
       name="output_prediction_tab" 
       format="tabular" 
-      label="SISTR Results"
+      label="${tool.name} on ${input_fastas[0].element_identifier}: Main Report (tab)"
       from_work_dir="sistr-report.tab">
       <filter>output_format == "tab"</filter>
     </data>
     <data 
       name="cgmlst_profiles" 
       format="csv" 
-      label="cgMLST results" />
+      label="${tool.name} on ${input_fastas[0].element_identifier}: cgMLST allele match results (csv)" />
     <data
       name="novel_alleles"
       format="fasta" 
-      label="Novel cgMLST alleles" />
+      label="${tool.name} on ${input_fastas[0].element_identifier}: Novel cgMLST alleles (fasta)" />
     <data 
       name="alleles_output"
       format="json"
-      label="cgMLST allele match results" />
+      label="${tool.name} on ${input_fastas[0].element_identifier}: cgMLST allele match results (json)" />
   </outputs>
   <tests>
-    <test>
+    <test expect_num_outputs="4">
       <param name="input_fastas" value="AE014613-699860.fasta" ftype="fasta"/>
       <param name="output_format" value="tab"/>
+      <param name="selection_mode" value="none"/>
       <output 
         name="cgmlst_profiles" 
-        value="cgmlst-profiles.csv"
+        value="cgmlst-profiles-AE014613.csv"
         ftype="csv"
         compare="diff">
       </output>
       <output name="output_prediction_tab" 
         ftype="tabular"
-        value="sistr-results.tabular"
-        compare="sim_size">
+        value="sistr-results-AE014613-699860.tabular"
+        compare="sim_size"> 
         <assert_contents>
           <has_text text="AE014613-699860" />
           <has_text text="Typhi" />
           <has_text text="enterica" />
           <has_text_matching expression="FAIL\t-\t-:-:-\t-:-:-"/>
           <has_text text="2014-LET-0419"/>
-          <has_n_columns n="22" />
+          <has_n_columns n="23" />
         </assert_contents>
       </output>
       <output 
         name="alleles_output" 
-        value="alleles-output.json"
+        value="alleles-output-AE014613.json"
         ftype="json"
         compare="sim_size">
         <assert_contents>
@@ -198,8 +222,8 @@
         </assert_contents> 
       </output>
     </test>
-    <test>
-      <param name="input_fastas" value="13-1101-Paratyphi_B.fasta" ftype="fasta"/>
+    <test expect_num_outputs="4">
+      <param name="input_fastas" value="13-1101-Paratyphi_B_varJava.fasta" ftype="fasta"/>
       <param name="output_format" value="tab"/>
       <output 
         name="novel_alleles" 
@@ -222,8 +246,10 @@
           <has_text text="Paratyphi B var. Java"/>
           <has_text text="enterica"/>
           <has_text text="1,4,[5],12"/>
-          <has_text text="PASS"/>
-          <has_n_columns n="22"/>
+          <has_text text="1,4,[5],12:b:1,2"/>
+          <has_text text="FAIL"/>
+          <has_text text="Y"/>
+          <has_n_columns n="24"/>
         </assert_contents>
       </output>
       <output 
@@ -255,13 +281,13 @@ Usage::
     Serovar predictions from whole-genome sequence assemblies by determination of antigen gene and cgMLST gene alleles using BLAST.
 
     Note about using the "--use-full-cgmlst-db" flag:
-        The "centroid" allele database is ~10% the size of the full set so analysis is much quicker with the "centroid" vs "full" set of alleles. Results between 2 cgMLST allele sets should not differ.
+      The "centroid" allele database is ~10% the size of the full set so analysis is much quicker with the "centroid" vs "full" set of alleles. 
+      Results between 2 cgMLST allele sets should not differ.
 
     If you find this program useful in your research, please cite as:
-
-    The Salmonella In Silico Typing Resource (SISTR): an open web-accessible tool for rapidly typing and subtyping draft Salmonella genome assemblies.
-    Catherine Yoshida, Peter Kruczkiewicz, Chad R. Laing, Erika J. Lingohr, Victor P.J. Gannon, John H.E. Nash, Eduardo N. Taboada.
-    PLoS ONE 11(1): e0147101. doi: 10.1371/journal.pone.0147101
+      The Salmonella In Silico Typing Resource (SISTR): an open web-accessible tool for rapidly typing and subtyping draft Salmonella genome assemblies.
+      Catherine Yoshida, Peter Kruczkiewicz, Chad R. Laing, Erika J. Lingohr, Victor P.J. Gannon, John H.E. Nash, Eduardo N. Taboada.
+      PLoS ONE 11(1): e0147101. doi: 10.1371/journal.pone.0147101
 
     positional arguments:
       F                     Input genome FASTA file
@@ -300,6 +326,11 @@ Usage::
                             serovar prediction results.
       -t THREADS, --threads THREADS
                             Number of parallel threads to run sistr_cmd analysis.
+      -l [LIST_OF_SEROVARS], --list-of-serovars [LIST_OF_SEROVARS]
+                            A path to a single column text file containing list of
+                            serovars to check SISTR serovar prediction against.
+                            Result reported in the "predicted_serovar_in_list"
+                            field as Y (present) or N (absent) value.                      
       -v, --verbose         Logging verbosity level (-v == show warnings; -vvv ==
                             show debug info)
       -V, --version         show program's version number and exit