Some debugging and improvements.

Rothamsted · Jul 15, 2017 · 8a880db · 8a880db
1 parent 6322a50
commit 8a880db
Show file tree

Hide file tree

Showing 6 changed files with 49 additions and 50 deletions.
diff --git a/...les/tab-parser-2/src/main/java/net/sourceforge/ondex/plugins/tab_parser_2/TabParser2.java b/...les/tab-parser-2/src/main/java/net/sourceforge/ondex/plugins/tab_parser_2/TabParser2.java
@@ -1,21 +1,19 @@
 package net.sourceforge.ondex.plugins.tab_parser_2;
 
 import java.io.File;
+import java.util.Arrays;
 import java.util.Collections;
-import java.util.stream.Collectors;
-import java.util.stream.Stream;
+import java.util.Optional;
 
 import org.apache.log4j.Logger;
 
-import com.google.common.base.Optional;
+import com.machinezoo.noexception.CheckedExceptionHandler;
 import com.machinezoo.noexception.Exceptions;
 
 import net.sourceforge.ondex.ONDEXPluginArguments;
 import net.sourceforge.ondex.args.ArgumentDefinition;
-import net.sourceforge.ondex.args.BooleanArgumentDefinition;
 import net.sourceforge.ondex.args.FileArgumentDefinition;
-import net.sourceforge.ondex.core.ONDEXConcept;
-import net.sourceforge.ondex.core.ONDEXRelation;
+import net.sourceforge.ondex.args.StringArgumentDefinition;
 import net.sourceforge.ondex.parser.ONDEXParser;
 import net.sourceforge.ondex.plugins.tab_parser_2.config.ConfigParser;
 import net.sourceforge.ondex.tools.subgraph.Subgraph;
@@ -33,7 +31,7 @@
  *
  */
 public class TabParser2 extends ONDEXParser
-{
+{		
   private Logger log = Logger.getLogger ( this.getClass() );
 
 	public String getId ()
@@ -53,7 +51,7 @@ public String getVersion ()
 
 	public ArgumentDefinition<?>[] getArgumentDefinitions ()
 	{
-		return new ArgumentDefinition<?>[] {
+		return new ArgumentDefinition[] {
       new FileArgumentDefinition ( 
       	FileArgumentDefinition.INPUT_FILE, FileArgumentDefinition.INPUT_FILE_DESC, true, true, false, false
       ),
@@ -64,9 +62,16 @@ public ArgumentDefinition<?>[] getArgumentDefinitions ()
       	true, // preExisting 
       	false // isDirectory
       ),
-      new BooleanArgumentDefinition ( PathParser.MERGE_ACC, "Accession-based concept merging in the parser", false, true ),
-      new BooleanArgumentDefinition ( PathParser.MERGE_NAME, "Name-based concept merging in the parser", false, false ),
-      new BooleanArgumentDefinition ( PathParser.MERGE_GDS, "Data source-based concept merging in the parser", false, false )
+      new StringArgumentDefinition ( 
+      	"mergeOptions", 
+      	"Merging options passed to PathParser, valid values are: " 
+      	+ Arrays.toString ( PathParser.validFlags.toArray ( new String[ 0 ] ) )
+      	+ ", default is " + PathParser.MERGE_ACC
+      	+ ", use '-' to specify none", 
+      	false, // required
+      	PathParser.MERGE_ACC, // default 
+      	true  // multiple vals
+      )
 		};
 	}
 
@@ -79,33 +84,21 @@ public void start () throws Exception
     // The mapping parser returns an ONDEX parser straight
     PathParser tabParser = ConfigParser.parseConfigXml ( tabConfigXmlPath, graph, tabInputPath );
 
-    // Merging flags
-    //
-    String[] mergeFlags = (String[]) Stream.of ( PathParser.MERGE_ACC, PathParser.MERGE_NAME, PathParser.MERGE_GDS )
-    .filter ( flag ->
-    	// Keeps it if it's true, else this kind of merging won't happen, because it won't be in the final 
-    	// processing options.
-    	Boolean.TRUE.equals ( 
-    		Exceptions
-        // Wraps InvalidPluginArgumentException for getUniqueValue() into a RuntimeException   			
-    		.wrap ( ex -> new IllegalArgumentException ( "Plug-in argument error: " + ex.getMessage (), ex ) )
-    		.get ( () -> (Boolean) args.getUniqueValue ( flag ) )
-    	)
-    )
-    .collect ( Collectors.toList () )    
-    .toArray ( new String [ 0 ] );
-
-    tabParser.setProcessingOptions ( mergeFlags );
+    String[] mergeFlags = (String[]) args.getObjectValueArray ( "mergeOptions" );
+    if ( ! ( mergeFlags == null || mergeFlags.length == 1 && "-".equals ( mergeFlags [ 0 ] ) ) )
+    	// Consider them if non null and non default
+    	tabParser.setProcessingOptions ( mergeFlags );
 
     Subgraph newGraph = tabParser.parse ();
-    int nconcepts = Optional.fromNullable ( newGraph.getConcepts () ).or ( Collections.<ONDEXConcept>emptySet () ).size ();
-    int nrelations = Optional.fromNullable ( newGraph.getRelations () ).or ( Collections.<ONDEXRelation>emptySet () ).size ();
+    int nconcepts = Optional.ofNullable ( newGraph.getConcepts () ).orElse ( Collections.emptySet () ).size ();
+    int nrelations = Optional.ofNullable ( newGraph.getRelations () ).orElse ( Collections.emptySet () ).size ();
 
     log.info ( String.format ( 
     	"Got %d concepts and %d relations from '%s'", 
     	nconcepts, nrelations, new File ( tabInputPath ).getAbsolutePath () 
     ));
 	}
+
 
 	public String[] requiresValidators ()
 	{

diff --git a/modules/tab-parser-2/src/test/resources/duped_relations/duped_rels.tsv b/modules/tab-parser-2/src/test/resources/duped_relations/duped_rels.tsv
@@ -2,4 +2,4 @@ Ensembl Gene ID	UniProt/SwissProt Accession	Chromosome Name	Gene Start (bp)	Gene
 ENSG00000137845	O14672	15	58588807	58749978	Note 1
 ENSG00000137845	O14672	15	58588807	58749978	Note 2
 ENSG00000159082	O43426	21	32628759	32728048	Note 3
-ENSG00000115317	O43464	2	74529377	74533348	
+ENSG00000115317	O43464	2	74529377	74533348	
diff --git a/modules/tab-parser-2/src/test/resources/duped_relations/duped_rels_cfg.xml b/modules/tab-parser-2/src/test/resources/duped_relations/duped_rels_cfg.xml
@@ -12,6 +12,7 @@
 			<column index = '0' />
 		</accession>
 		<name><column index = '0' /></name>
+		<parser-id><column index = '0' /></parser-id>		
 		<data-source>ENSEMBL</data-source>
 		<attribute name="TAXID" type = "TEXT" indexed = 'false'>9606</attribute>
 		<attribute name="Chromosome" type = "INTEGER">
@@ -31,7 +32,9 @@
 		<accession data-source="UNIPROTKB"><!-- ambiguous is false by default -->
 			<column index = '1'/>
 		</accession>
-		<data-source>ENSEMBL</data-source>
+		<name><column index = '1' /></name>
+		<parser-id><column index = '1' /></parser-id>				
+		<data-source>UNIPROTKB</data-source>
 	</concept>
 
 

diff --git a/modules/tab-parser-2/src/test/resources/duped_relations/duped_rels_wf.xml b/modules/tab-parser-2/src/test/resources/duped_relations/duped_rels_wf.xml
@@ -8,26 +8,22 @@
     <Parser name="tabParser2">
       <Arg name="InputFile">/Users/brandizi/Documents/Work/RRes/ondex_git/ondex-full/ondex-knet-builder/modules/tab-parser-2/src/test/resources/duped_relations/duped_rels.tsv</Arg>
       <Arg name="configFile">/Users/brandizi/Documents/Work/RRes/ondex_git/ondex-full/ondex-knet-builder/modules/tab-parser-2/src/test/resources/duped_relations/duped_rels_cfg.xml</Arg>
-      <Arg name="MERGE_ON_ACCESSIONS">false</Arg>
-      <Arg name="MERGE_ON_NAMES">false</Arg>
-      <Arg name="MERGE_ON_GDS">false</Arg>
+      <Arg name="mergeOptions">-</Arg>
       <Arg name="graphId">default</Arg>
     </Parser>
-
-		<!-- Mapping name="lowmemoryaccessionbased">
-		  <Arg name="IgnoreAmbiguity">false</Arg>
-		  <!- - Arg name="RelationType">collapse_me</Arg - ->
-		  <Arg name="WithinDataSourceMapping">false</Arg>
-			<Arg name="graphId">default</Arg>
-	  </Mapping -->
-
-		<!-- Transformer name="relationcollapser">
+    <Mapping name="lowmemoryaccessionbased">
+      <Arg name="IgnoreAmbiguity">false</Arg>
+      <Arg name="RelationType">same-as</Arg>
+      <Arg name="WithinDataSourceMapping">true</Arg>
+      <Arg name="graphId">default</Arg>
+      <Arg name="EquivalentConceptClass">Gene,Gene;Protein,Protein</Arg>
+    </Mapping>
+    <Transformer name="relationcollapser">
       <Arg name="CloneAttributes">true</Arg>
       <Arg name="CopyTagReferences">true</Arg>
       <Arg name="graphId">default</Arg>
-      <Arg name="RelationType">enc</Arg>
-    </Transformer -->
-
+      <Arg name="RelationType">same-as</Arg>
+    </Transformer>
   </Workflow>
 </Ondex>
 
diff --git a/modules/tab-parser-2/src/test/resources/ondex_tutorial_2016/gene_example_import_wf.xml b/modules/tab-parser-2/src/test/resources/ondex_tutorial_2016/gene_example_import_wf.xml
@@ -6,8 +6,11 @@
       <Arg name="graphId">default</Arg>
     </Graph>
     <Parser name="tabParser2">
-      <Arg name="InputFile">/Users/brandizi/Documents/Work/RRes/ondex-code/modules/tab-parser-2/src/test/resources/ondex_tutorial_2016/gene_example.tsv</Arg>
-      <Arg name="configFile">/Users/brandizi/Documents/Work/RRes/ondex-code/modules/tab-parser-2/src/test/resources/ondex_tutorial_2016/gene_example_parser_cfg.xml</Arg>
+      <Arg name="InputFile">/Users/brandizi/Documents/Work/RRes/ondex_git/ondex-full/ondex-knet-builder/modules/tab-parser-2/src/test/resources/ondex_tutorial_2016/gene_example.tsv</Arg>
+      <Arg name="configFile">/Users/brandizi/Documents/Work/RRes/ondex_git/ondex-full/ondex-knet-builder/modules/tab-parser-2/src/test/resources/ondex_tutorial_2016/gene_example_parser_cfg.xml</Arg>
+      <Arg name="MERGE_ON_ACCESSIONS">true</Arg>
+      <Arg name="MERGE_ON_NAMES">false</Arg>
+      <Arg name="MERGE_ON_GDS">false</Arg>
       <Arg name="graphId">default</Arg>
     </Parser>
   </Workflow>

diff --git a/modules/tab-parser-2/src/test/resources/ondex_tutorial_2016/gene_example_parser_cfg.xml b/modules/tab-parser-2/src/test/resources/ondex_tutorial_2016/gene_example_parser_cfg.xml
@@ -12,6 +12,8 @@
 			<column index = '0' />
 		</accession>
 		<name><column index = '0' /></name>
+		<parser-id><column index = '0' /></parser-id>
+
 		<data-source>ENSEMBL</data-source>
 		<attribute name="TAXID" type = "TEXT" indexed = 'false'>9606</attribute>
 		<attribute name="Chromosome" type = "INTEGER">
@@ -31,7 +33,9 @@
 		<accession data-source="UNIPROTKB"><!-- ambiguous is false by default -->
 			<column index = '1'/>
 		</accession>
-		<data-source>ENSEMBL</data-source>
+		<data-source>UNIPROTKB</data-source>
+		<name><column index = '1' /></name>
+		<parser-id><column index = '1' /></parser-id>		
 	</concept>