Mercurial > repos > richard-burhans > ncbi_egapx

changeset 40:b095c69eab5b draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/richard-burhans/galaxytools/tree/main/tools/ncbi_egapx commit 69aca8290c162e4bce52ff788553a3676b0a7be6
author richard-burhans
date Fri, 18 Jul 2025 00:58:39 +0000
parents 4f85da43b40c
children d129c3d0e920
files macros.xml ncbi_egapx.xml test-data/input.yaml
diffstat 3 files changed, 59 insertions(+), 47 deletions(-) [+]
line wrap: on
line diff
--- a/macros.xml	Mon Jun 30 23:21:01 2025 +0000
+++ b/macros.xml	Fri Jul 18 00:58:39 2025 +0000
@@ -4,7 +4,7 @@
             <container type="docker">quay.io/galaxy/egapx:@TOOL_VERSION@-alpha</container>
         </requirements>
     </xml>
-    <token name="@TOOL_VERSION@">0.4.0</token>
+    <token name="@TOOL_VERSION@">0.4.1</token>
     <token name="@VERSION_SUFFIX@">0</token>
     <token name="@PROFILE@">23.0</token>
     <xml name="edam_ontology">
@@ -44,17 +44,7 @@
                     <has_size min="1"/>
                 </assert_contents>
             </element>
-            <element name="annotation_data" ftype="tabular">
-                <assert_contents>
-                    <has_size min="1"/>
-                </assert_contents>
-            </element>
-            <element name="gnomon_accepted" ftype="txt">
-                <assert_contents>
-                    <has_size min="1"/>
-                </assert_contents>
-            </element>
-            <element name="complete_cds" ftype="fasta">
+            <element name="complete_genomic_gtf" ftype="gtf">
                 <assert_contents>
                     <has_size min="1"/>
                 </assert_contents>
@@ -64,12 +54,7 @@
                     <has_size min="1"/>
                 </assert_contents>
             </element>
-            <element name="complete_genomic_gtf" ftype="gtf">
-                <assert_contents>
-                    <has_size min="1"/>
-                </assert_contents>
-            </element>
-            <element name="complete_proteins" ftype="fasta">
+            <element name="complete_cds" ftype="fasta">
                 <assert_contents>
                     <has_size min="1"/>
                 </assert_contents>
@@ -79,12 +64,27 @@
                     <has_size min="1"/>
                 </assert_contents>
             </element>
-            <element name="gnomon_contam_report" ftype="fasta">
+            <element name="complete_proteins" ftype="fasta">
+                <assert_contents>
+                    <has_size min="1"/>
+                </assert_contents>
+            </element>
+            <element name="annotation_data" ftype="tabular">
                 <assert_contents>
                     <has_size min="1"/>
                 </assert_contents>
             </element>
-            <element name="gnomon_quality_report" ftype="fasta">
+            <element name="sra_metadata" ftype="tabular">
+                <assert_contents>
+                    <has_size min="1"/>
+                </assert_contents>
+            </element>
+            <element name="gnomon_accept_ftable_annot" ftype="txt">
+                <assert_contents>
+                    <has_size min="0"/>
+                </assert_contents>
+            </element>
+            <element name="gnomon_contam_rpt" ftype="tabular">
                 <assert_contents>
                     <has_size min="1"/>
                 </assert_contents>
@@ -94,27 +94,37 @@
                     <has_size min="1"/>
                 </assert_contents>
             </element>
-            <element name="sra_metadata" ftype="binary">
+            <element name="gnomon_quality_report" ftype="tabular">
                 <assert_contents>
                     <has_size min="1"/>
                 </assert_contents>
             </element>
-            <element name="feature_counts" ftype="xml">
+            <element name="busco_report" ftype="txt">
+                <assert_contents>
+                    <has_size min="1"/>
+                </assert_contents>
+            </element>
+            <element name="busco_report_json" ftype="json">
+                <assert_contents>
+                    <has_size min="1"/>
+                </assert_contents>
+            </element>
+            <element name="stats_feature_counts" ftype="xml">
                 <assert_contents>
                     <is_valid_xml/>
                 </assert_contents>
             </element>
-            <element name="feature_stats" ftype="xml">
+            <element name="stats_feature_stats" ftype="xml">
                 <assert_contents>
                     <is_valid_xml/>
                 </assert_contents>
             </element>
-            <element name="all_unannotated_val" ftype="xml">
+            <element name="validated_all_unannotated" ftype="xml">
                 <assert_contents>
-                    <is_valid_xml/>
+                    <has_size min="0"/>
                 </assert_contents>
             </element>
-            <element name="genome_val" ftype="xml">
+            <element name="validated_genome" ftype="xml">
                 <assert_contents>
                     <is_valid_xml/>
                 </assert_contents>
@@ -126,11 +136,6 @@
                     <has_size min="1"/>
                 </assert_contents>
             </element>
-            <element name="run_params" ftype="yaml">
-                <assert_contents>
-                    <has_size min="1"/>
-                </assert_contents>
-            </element>
             <element name="run_report" ftype="html">
                 <assert_contents>
                     <has_size min="1"/>
@@ -146,6 +151,11 @@
                     <has_size min="1"/>
                 </assert_contents>
             </element>
+            <element name="run_params" ftype="yaml">
+                <assert_contents>
+                    <has_size min="1"/>
+                </assert_contents>
+            </element>
         </output_collection>
     </xml>
 </macros>
--- a/ncbi_egapx.xml	Mon Jun 30 23:21:01 2025 +0000
+++ b/ncbi_egapx.xml	Fri Jul 18 00:58:39 2025 +0000
@@ -268,28 +268,30 @@
         <data name="complete_genomic_gff" format="gff" label="Final annotation for ${on_string}" from_work_dir="egapx_out/complete.genomic.gff"/>
         <collection name="output_files" type="list" label="EGAPx output for ${on_string}">
             <data name="annotated_genome" format="asn1" label="Final annotation" from_work_dir="egapx_out/annotated_genome.asn"/>
-            <data name="annotation_data" format="tabular" label="Annotation structured comment" from_work_dir="egapx_out/annotation_data.cmt"/>
-            <data name="gnomon_accepted" format="txt" label="Gnomon accepted annotation models" from_work_dir="egapx_out/annot_builder_output/accept.ftable_annot"/>
+            <data name="complete_genomic_gtf" format="gtf" label="Final annotation" from_work_dir="egapx_out/complete.genomic.gtf"/>
+            <data name="complete_genomic_fna" format="fasta" label="Full genome sequences" from_work_dir="egapx_out/complete.genomic.fna"/>
             <data name="complete_cds" format="fasta" label="Annotated CDS" from_work_dir="egapx_out/complete.cds.fna"/>
-            <data name="complete_genomic_fna" format="fasta" label="Full genome sequences" from_work_dir="egapx_out/complete.genomic.fna"/>
-            <data name="complete_genomic_gtf" format="gtf" label="Final annotation" from_work_dir="egapx_out/complete.genomic.gtf"/>
-            <data name="complete_proteins" format="fasta" label="Annotated protein products" from_work_dir="egapx_out/complete.proteins.faa"/>
             <data name="complete_transcripts" format="fasta" label="Annotated transcripts" from_work_dir="egapx_out/complete.transcripts.fna"/>
-            <data name="gnomon_contam_report" format="tabular" label="Gnomon contamination report" from_work_dir="egapx_out/GNOMON/contam_rpt.tsv"/>
-            <data name="gnomon_quality_report" format="tabular" label="Gnomon quality report" from_work_dir="egapx_out/GNOMON/new.gnomon_quality_report.txt"/>
-            <data name="gnomon_report" format="txt" label="Gnomon report" from_work_dir="egapx_out/GNOMON/new.gnomon_report.txt"/>
+            <data name="complete_proteins" format="fasta" label="Annotated protein products" from_work_dir="egapx_out/complete.proteins.faa"/>
+            <data name="annotation_data" format="tabular" label="Annotation structured comment" from_work_dir="egapx_out/annotation_data.cmt"/>
             <data name="sra_metadata" format="tabular" label="SRA run metadata" from_work_dir="egapx_out/sra_metadata.dat"/>
-            <data name="feature_counts" format="xml" label="Feature counts" from_work_dir="egapx_out/stats/feature_counts.xml"/>
-            <data name="feature_stats" format="xml" label="Feature stats" from_work_dir="egapx_out/stats/feature_stats.xml"/>
-            <data name="all_unannotated_val" format="xml" label="all_unannotated" from_work_dir="egapx_out/validated/all_unannotated.val"/>
-            <data name="genome_val" format="xml" label="all_unannotated" from_work_dir="egapx_out/validated/genome.val"/>
+            <data name="gnomon_accept_ftable_annot" format="txt" label="Gnomon accepted annotation models" from_work_dir="egapx_out/annot_builder_output/accept.ftable_annot"/>
+            <data name="gnomon_contam_rpt" format="tabular" label="Gnomon contamination report" from_work_dir="egapx_out/GNOMON/contam_rpt.tsv"/>
+            <data name="gnomon_report" format="txt" label="Gnomon report" from_work_dir="egapx_out/GNOMON/new.gnomon_report.txt"/>
+            <data name="gnomon_quality_report" format="tabular" label="Gnomon quality report" from_work_dir="egapx_out/GNOMON/new.gnomon_quality_report.txt"/>
+            <data name="busco_report" format="txt" label="BUSCO report" from_work_dir="egapx_out/busco/short_summary*.txt"/>
+            <data name="busco_report_json" format="json" label="BUSCO report (json)" from_work_dir="egapx_out/busco/short_summary*.json"/>
+            <data name="stats_feature_counts" format="xml" label="Feature counts" from_work_dir="egapx_out/stats/feature_counts.xml"/>
+            <data name="stats_feature_stats" format="xml" label="Feature stats" from_work_dir="egapx_out/stats/feature_stats.xml"/>
+            <data name="validated_all_unannotated" format="xml" label="validation all_unannotated" from_work_dir="egapx_out/validated/all_unannotated.val"/>
+            <data name="validated_genome" format="xml" label="valudation genome" from_work_dir="egapx_out/validated/genome.val"/>
         </collection>
         <collection name="nextflow_stats" type="list" label="EGAPx nextflow stats for ${on_string}">
             <data name="nextflow_log" format="txt" label="Nextflow execution log" from_work_dir="egapx_out/nextflow/nextflow.log"/>
-            <data name="run_params" format="yaml" label="Nextflow run parameters" from_work_dir="egapx_out/nextflow/run_params.yaml"/>
             <data name="run_report" format="html" label="Nextflow execution report" from_work_dir="egapx_out/nextflow/run.report.html"/>
             <data name="run_timeline" format="html" label="Nextflow execution timeline" from_work_dir="egapx_out/nextflow/run.timeline.html"/>
             <data name="run_trace" format="tabular" label="Nextflow trace file" from_work_dir="egapx_out/nextflow/run.trace.txt"/>
+            <data name="run_params" format="yaml" label="Nextflow run parameters" from_work_dir="egapx_out/nextflow/run_params.yaml"/>
         </collection>
         <collection name="star_alignments" type="list" label="EGAPx STAR alignments for ${on_string}">
             <discover_datasets pattern="(?:.+/)?bam\.[0-9A-Za-z]{10}/(?P&lt;designation&gt;.+)\.bam" format="bam" directory="egapx_out" recurse="true" match_relative_path="true"/>
@@ -297,7 +299,7 @@
         </collection>
     </outputs>
     <tests>
-        <test expect_num_outputs="20" expect_test_failure="true">
+        <test expect_num_outputs="27" expect_test_failure="true">
             <conditional name="cond_input_style">
                 <param name="input_style" value="fillform"/>
                 <conditional name="cond_genome_style">
@@ -315,7 +317,7 @@
             </section>
             <expand macro="test_outputs"/>
         </test>
-        <test expect_num_outputs="21" expect_test_failure="true">
+        <test expect_num_outputs="27" expect_test_failure="true">
             <conditional name="cond_input_style">
                 <param name="input_style" value="history"/>
                 <param name="yamlin" value="input.yaml"/>
--- a/test-data/input.yaml	Mon Jun 30 23:21:01 2025 +0000
+++ b/test-data/input.yaml	Fri Jul 18 00:58:39 2025 +0000
@@ -1,6 +1,6 @@
 genome: https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/020/809/275/GCF_020809275.1_ASM2080927v1/GCF_020809275.1_ASM2080927v1_genomic.fna.gz
 taxid: 6954
-reads:
+short_reads:
   - https://ftp.ncbi.nlm.nih.gov/genomes/TOOLS/EGAP/sample_data/Dermatophagoides_farinae_small/SRR8506572_1.gz
   - https://ftp.ncbi.nlm.nih.gov/genomes/TOOLS/EGAP/sample_data/Dermatophagoides_farinae_small/SRR8506572_2.gz
   - https://ftp.ncbi.nlm.nih.gov/genomes/TOOLS/EGAP/sample_data/Dermatophagoides_farinae_small/SRR9005248_1.gz