Export min / max intron length parameters

Author: pvanheus

galaxy/lukasa.xml

@@ -1,7 +1,7 @@
 <tool id="lukasa" name="Lukasa" version="@TOOL_VERSION@+galaxy0" python_template_version="3.5">
     <description>Align protein evidence against genomic contigs using metaeuk and spaln</description>
     <macros>
-        <token name="@TOOL_VERSION@">0.11.0</token>
+        <token name="@TOOL_VERSION@">0.12.0</token>
         <!-- note that this DOI is for lukasa 0.0.6 - update as needed -->
         <token name="@DOI@">10.5281/zenodo.4084862</token>
     </macros>

lukasa.py

@@ -9,21 +9,33 @@
 import tempfile
 from os.path import abspath
 
-template = """contigs_fasta:
-  class: File
-  format: edam:format_1929
-  path: {}
-proteins_fasta:
-  class: File
-  format: edam:format_1929
-  path: {}
-{}
 
-$namespaces:
-  edam: http://edamontology.org/
-$schemas:
-  - http://edamontology.org/EDAM_1.18.owl
-"""
+def build_template(
+    contig_path: str,
+    protein_path: str,
+    species_table: str = "",
+    max_intron: str = "",
+    min_intron: str = "",
+):
+    template = f"""
+    contigs_fasta:
+        class: File
+        format: edam:format_1929
+        path: {contig_path}
+    proteins_fasta:
+        class: File
+        format: edam:format_1929
+        path: {protein_path}
+    {max_intron}
+    {min_intron}
+    {species_table}
+
+    $namespaces:
+    edam: http://edamontology.org/
+    $schemas:
+    - http://edamontology.org/EDAM_1.18.owl
+    """
+    return template
 
 
 def is_fasta(input_filename):
@@ -45,6 +57,8 @@ def is_fasta(input_filename):
     )
     parser.add_argument("--output_filename", default="spaln_out.gff3")
     parser.add_argument("--workflow_dir", default=cwl_workflow_dir)
+    parser.add_argument("--max_intron", type=int, help="Maximum intron length")
+    parser.add_argument("--min_intron", type=int, help="Minimum intron length")
     parser.add_argument("contigs_filename", help="File with genomic contigs")
     parser.add_argument("proteins_filename", help="File with proteins to map")
     parser.add_argument("--species_table", help="spaln species table to use")
@@ -53,15 +67,21 @@ def is_fasta(input_filename):
         sys.exit("Error: Input files must be in FASTA format")
 
     cwl_input_file = tempfile.NamedTemporaryFile(delete=False, mode="w")
+
+    species_table = max_intron = min_intron = ""
     if args.species_table is not None:
-        species_table_string = "species_table: {}".format(args.species_table)
-    else:
-        species_table_string = ""
+        species_table = "species_table: {}".format(args.species_table)
+    if args.max_intron is not None:
+        max_intron = f"max_intron: {args.max_intron}\n"
+    if args.min_intron is not None:
+        min_intron = f"min_intron: {args.min_intron}\n"
     cwl_input_file.write(
-        template.format(
+        build_template(
             abspath(args.contigs_filename),
             abspath(args.proteins_filename),
-            species_table_string,
+            species_table=species_table,
+            max_intron=max_intron,
+            min_intron=min_intron,
         )
     )
     cwl_input_file.close()

protein_evidence_mapping.cwl

@@ -14,6 +14,14 @@ inputs:
   species_table:
     label: "Spaln species table to use (optional)"
     type: string?
+  max_intron:
+    label: "Maximum intron length"
+    type: int?
+    doc: "Maximum intron length, passed to metaeuk"
+  min_intron:
+    label: "Minimum intron length"
+    type: int?
+    doc: "Minimum intron length, passed to metaeuk and spaln"
 outputs:
   spaln_out:
     type: File
@@ -30,6 +38,8 @@ steps:
     in:
       contigs: contigs_fasta
       query: proteins_fasta
+      max_intron: max_intron
+      min_intron: min_intron
     out:
       - output_fasta
   samtools_index_contigs:
@@ -80,6 +90,7 @@ steps:
       genome_fasta: extract_region_pairs/contig_fasta
       query_fasta: extract_region_pairs/protein_fasta
       species: species_table
+      min_intron: min_intron
       output_format:
         default: 0
     out:

tools/metaeuk_easy_predict.cwl

@@ -19,7 +19,7 @@ inputs:
       prefix: --metaeuk-eval
       position: 1
   max_intron:
-    type: float?
+    type: int?
     inputBinding:
       prefix: --max-intron
       position: 1
@@ -29,7 +29,7 @@ inputs:
       prefix: --metaeuk-tcov
       position: 1
   min_intron:
-    type: float?
+    type: int?
     inputBinding:
       prefix: --min-intron
       position: 1

tools/spaln.cwl

@@ -10,6 +10,11 @@ inputs:
     format: edam:format_1929
     inputBinding:
       position: 10
+  min_intron:
+    type: int?
+    inputBinding:
+      prefix: -yL
+      position: 1
   output_format:
     type: int?
     inputBinding: