Merge pull request #3 from sureshhewabi/master

save output as a file & multiple peak files check

pom.xml

@@ -12,12 +12,19 @@
         <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
         <archive.repo.version>1.0.2</archive.repo.version>
         <px.submission.core.version>2.0.28</px.submission.core.version>
+        <ms.data.core.api.version>2.0.31-SNAPSHOT</ms.data.core.api.version>
         <javax.activation.version>1.2.0</javax.activation.version>
         <jaxb.api.version>2.3.1</jaxb.api.version>
         <pia.version>1.3.15</pia.version>
+        <cpdetector>1.0.7</cpdetector>
     </properties>
 
     <dependencies>
+        <dependency>
+            <groupId>cpdetector</groupId>
+            <artifactId>cpdetector</artifactId>
+            <version>${cpdetector}</version>
+        </dependency>
         <dependency>
             <groupId>de.mpc.pia</groupId>
             <artifactId>pia</artifactId>
@@ -55,7 +62,11 @@
                     <groupId>com.sun.xml.bind</groupId>
                     <artifactId>jaxb-xjc</artifactId>
                 </exclusion>
-
+                <!--           exclude this as this fails to download from maven a lot. Instead include it separately-->
+                <exclusion>
+                    <groupId>net.sourceforge.cpdetector</groupId>
+                    <artifactId>cpdetector</artifactId>
+                </exclusion>
             </exclusions>
         </dependency>
 
@@ -270,6 +281,17 @@
             <id>nexus-ebi-snapshot-repo</id>
             <url>http://www.ebi.ac.uk/Tools/maven/repos/content/groups/ebi-snapshots/</url>
         </repository>
+        <repository>
+            <id>nexus-ebi-intact-repo</id>
+            <url>https://www.ebi.ac.uk/intact/maven/nexus/content/repositories/ebi-repo/</url>
+        </repository>
+        <!-- UGent Genesis repository -->
+        <repository>
+            <id>genesis-maven2-repository</id>
+            <name>Genesis maven2 repository</name>
+            <url>http://genesis.ugent.be/maven2</url>
+            <layout>default</layout>
+        </repository>
         <repository>
             <id>pst-release</id>
             <name>EBI Nexus Repository</name>

src/main/java/uk/ac/ebi/pride/toolsuite/px_validator/SubmissionToolValidator.java

@@ -4,6 +4,7 @@
 import org.apache.log4j.Logger;
 import uk.ac.ebi.pride.toolsuite.px_validator.utils.IReport;
 
+import java.io.File;
 import java.util.*;
 
 import static uk.ac.ebi.pride.toolsuite.px_validator.utils.Utility.*;
@@ -30,8 +31,14 @@ public static void main(String[] args) {
       if (args.length > 0) {
         if (cmd.hasOption(ARG_VALIDATION)) {
           IReport report = Validator.startValidation(cmd);
-          if (report != null)
-            log.info(report.toString());
+          if (report != null) {
+            if(cmd.hasOption(ARG_REPORTFILE)){
+              File outputFile  = cmd.hasOption(ARG_REPORTFILE) ? new File(cmd.getOptionValue(ARG_REPORTFILE)) : null;
+              outputReport(report, outputFile);
+            }else {
+              log.info(report.toString());
+            }
+          }
         }  else {
           log.error("Did not find validation command from arguments ");
           Arrays.stream(args).forEach(log::error);

src/main/java/uk/ac/ebi/pride/toolsuite/px_validator/Validator.java

@@ -84,7 +84,7 @@ private static List<File> getFilesToValidate(File file) {
    * @param cmd the command line arguments.
    * @return List of peak files.
    */
-  private static List<File> getPeakFiles(CommandLine cmd) {
+  public static List<File> getPeakFiles(CommandLine cmd) {
     List<File> peakFiles = new ArrayList<>();
     if (cmd.hasOption(ARG_PEAK) || cmd.hasOption(ARG_PEAKS)) {
       String[] peakFilesString = cmd.hasOption(ARG_PEAK) ? cmd.getOptionValues(ARG_PEAK)

src/main/java/uk/ac/ebi/pride/toolsuite/px_validator/utils/Report.java

@@ -34,7 +34,7 @@ public String toString() {
         for(Map.Entry er: errors)
             error.append(er.getKey().toString()).append(" == Level Error: ").append(er.getValue()).append(" ==\n");
         if(errors.isEmpty()){
-            error.append("No errors found -- OK").append("\n");
+            error.append("Errors found: None").append("\n");
         }
         return error.toString();
     }

src/main/java/uk/ac/ebi/pride/toolsuite/px_validator/utils/ResultReport.java

@@ -8,6 +8,7 @@ public class ResultReport extends Report {
     private int numberOfProteins;
     private int numberOfPeptides;
     private int numberOfPSMs;
+    private int numberOfPeakFiles;
 
     public ResultReport() {
 
@@ -25,10 +26,15 @@ public void setNumberOfPSMs(int numberOfPSMs) {
         this.numberOfPSMs = numberOfPSMs;
     }
 
+    public void setNumberOfPeakFiles(int numberOfPeakFiles) {
+        this.numberOfPeakFiles = numberOfPeakFiles;
+    }
+
     @Override
     public String toString() {
-        return super.toString() + "\nIdentification results Report: \nNumber of reported proteins  -- " + numberOfProteins + "\n" +
-                "Number of reported peptides  -- " + numberOfPeptides + "\n" +
-                "Number of reported PSMs  -- " + numberOfPSMs + "\n";
+        return super.toString() + "Number of reported proteins : " + numberOfProteins + "\n" +
+                "Number of reported peptides : " + numberOfPeptides + "\n" +
+                "Number of reported PSMs : " + numberOfPSMs + "\n" +
+                "Number of peak files : " + numberOfPeakFiles + "\n";
     }
 }

src/main/java/uk/ac/ebi/pride/toolsuite/px_validator/utils/Utility.java

@@ -104,4 +104,29 @@ public static File createNewTempFile(File file) {
     }
     return tempFile;
   }
+
+  public static void outputReport(IReport report, File reportFile){
+    log.info(report.toString());
+    if (reportFile!=null) {
+      try {
+        log.info("Writing report to: " + reportFile.getAbsolutePath());
+        java.nio.file.Files.write(reportFile.toPath(), report.toString().getBytes());
+      } catch (IOException ioe) {
+        log.error("Problem when writing report file: ", ioe);
+      }
+    }
+  }
+
+  public static String getRealFileName(String fileName) {
+    String name = fileName;
+
+    if (name.contains("/") || name.contains("\\")) {
+      String[] parts = name.split("/");
+      name = parts[parts.length - 1];
+      parts = name.split("\\\\");
+      name = parts[parts.length - 1];
+    }
+
+    return name;
+  }
 }

src/main/java/uk/ac/ebi/pride/toolsuite/px_validator/validators/MzIdValidator.java

@@ -2,8 +2,10 @@
 
 import de.mpc.pia.intermediate.compiler.PIASimpleCompiler;
 import de.mpc.pia.intermediate.compiler.parser.InputFileParserFactory;
+import de.mpc.pia.modeller.PIAModeller;
 import org.apache.commons.cli.CommandLine;
 import org.xml.sax.SAXException;
+import uk.ac.ebi.jmzidml.model.mzidml.SpectraData;
 import uk.ac.ebi.pride.data.validation.ValidationMessage;
 import uk.ac.ebi.pride.tools.ErrorHandlerIface;
 import uk.ac.ebi.pride.tools.GenericSchemaValidator;
@@ -13,11 +15,15 @@
 import java.io.*;
 import java.net.URI;
 import java.net.URISyntaxException;
+import java.util.ArrayList;
 import java.util.List;
+import java.util.Map;
+import java.util.stream.Collectors;
 
 public class MzIdValidator implements Validator{
 
     private File file;
+    private List<File> peakFiles;
     private static final String MZID_SCHEMA = "https://storage.googleapis.com/google-code-archive-downloads/v2/code.google.com/psi-pi/mzIdentML1.1.0.xsd";
 
     public static Validator getInstance(CommandLine cmd) throws Exception {
@@ -35,6 +41,7 @@ private MzIdValidator(CommandLine cmd) throws Exception{
         }else{
             throw new IOException("In order to validate a mzid file the argument -mzid should be provided");
         }
+        peakFiles = uk.ac.ebi.pride.toolsuite.px_validator.Validator.getPeakFiles(cmd);
     }
 
     @Override
@@ -50,13 +57,23 @@ public IReport validate() {
         piaCompiler.buildClusterList();
         piaCompiler.buildIntermediateStructure();
 
+        List<SpectraData> spectrumFiles = getPeakFiles(piaCompiler);
+        List<String> peakFilesError = validatePeakFiles(spectrumFiles);
+        if(peakFilesError.size() > 0){
+            for(String error: peakFilesError){
+                report.addException(new IOException(error), ValidationMessage.Type.ERROR);
+            }
+        }
+
         int numProteins = piaCompiler.getNrAccessions();
         int numPeptides = piaCompiler.getNrPeptides();
         int numPSMs = piaCompiler.getNrPeptideSpectrumMatches();
+        int numPeakFiles = spectrumFiles.size();
 
         ((ResultReport) report).setNumberOfPeptides(numPeptides);
         ((ResultReport) report).setNumberOfProteins(numProteins);
         ((ResultReport) report).setNumberOfPSMs(numPSMs);
+        ((ResultReport) report).setNumberOfPeakFiles(numPeakFiles);
         return report;
     }
 
@@ -77,4 +94,47 @@ private static IReport validateMzidSchema(File mzIdentML) {
         }
         return report;
   }
+
+  private List<SpectraData> getPeakFiles(PIASimpleCompiler piaCompiler){
+      PIAModeller piaModeller = null;
+      try {
+          if (piaCompiler.getAllPeptideSpectrumMatcheIDs() != null
+                  && !piaCompiler.getAllPeptideSpectrumMatcheIDs().isEmpty()) {
+              File inferenceTempFile = File.createTempFile("assay", ".tmp");
+              piaCompiler.writeOutXML(inferenceTempFile);
+              piaCompiler.finish();
+              piaModeller = new PIAModeller(inferenceTempFile.getAbsolutePath());
+
+              if (inferenceTempFile.exists()) {
+                  inferenceTempFile.deleteOnExit();
+              }
+          }
+      } catch (IOException e) {
+          e.printStackTrace();
+      }
+      piaModeller.getSpectraData();
+
+      return piaModeller.getSpectraData()
+              .entrySet().stream().map(Map.Entry::getValue)
+              .collect(Collectors.toList());
+  }
+
+  private List<String> validatePeakFiles(List<SpectraData> spectraDataFiles){
+      List<String> peakFileErrors = new ArrayList<>();
+
+      for(File peakFile: this.peakFiles){
+          boolean isFound = false;
+          for (SpectraData spectraData : spectraDataFiles) {
+              String filename = Utility.getRealFileName(spectraData.getLocation());
+              if(filename.equals(peakFile.getName())){
+                  isFound = true;
+                  break;
+              }
+          }
+          if(!isFound){
+              peakFileErrors.add(peakFile.getName() + " does not found as a reference in MzIdentML");
+          }
+      }
+      return peakFileErrors;
+  }
 }