Merge pull request #690 from Mailaender/xml-perf

Fixed XML based formats parsing the whole document during content matching
eclipse-chemclipse · Aug 3, 2021 · 96beeca · 96beeca
2 parents 8d640c8 + 59ae122
commit 96beeca
Show file tree

Hide file tree

Showing 13 changed files with 85 additions and 82 deletions.
diff --git a/...se/chemclipse/msd/converter/supplier/mzdata/converter/ChromatogramMagicNumberMatcher.java b/...se/chemclipse/msd/converter/supplier/mzdata/converter/ChromatogramMagicNumberMatcher.java
@@ -14,18 +14,15 @@
 
 import java.io.File;
 
-import javax.xml.bind.JAXBContext;
-import javax.xml.bind.Unmarshaller;
 import javax.xml.parsers.DocumentBuilder;
 import javax.xml.parsers.DocumentBuilderFactory;
 
 import org.eclipse.chemclipse.converter.core.AbstractMagicNumberMatcher;
 import org.eclipse.chemclipse.converter.core.IMagicNumberMatcher;
-import org.eclipse.chemclipse.msd.converter.supplier.mzdata.internal.io.IFormat;
 import org.eclipse.chemclipse.msd.converter.supplier.mzdata.internal.support.IConstants;
 import org.eclipse.chemclipse.msd.converter.supplier.mzdata.internal.support.SpecificationValidator;
-import org.eclipse.chemclipse.msd.converter.supplier.mzdata.internal.v105.model.MzData;
 import org.w3c.dom.Document;
+import org.w3c.dom.Element;
 import org.w3c.dom.NodeList;
 
 public class ChromatogramMagicNumberMatcher extends AbstractMagicNumberMatcher implements IMagicNumberMatcher {
@@ -45,13 +42,17 @@ public boolean checkFileFormat(File file) {
 			DocumentBuilderFactory documentBuilderFactory = DocumentBuilderFactory.newInstance();
 			DocumentBuilder documentBuilder = documentBuilderFactory.newDocumentBuilder();
 			Document document = documentBuilder.parse(file);
-			NodeList nodeList = document.getElementsByTagName(IConstants.NODE_MZ_DATA);
-			//
-			JAXBContext jaxbContext = JAXBContext.newInstance(IFormat.CONTEXT_PATH_V_105);
-			Unmarshaller unmarshaller = jaxbContext.createUnmarshaller();
-			MzData mzData = (MzData)unmarshaller.unmarshal(nodeList.item(0));
-			if(mzData.getSpectrumList().getCount() > 1) {
-				isValidFormat = true;
+			NodeList root = document.getElementsByTagName(IConstants.NODE_MZ_DATA);
+			if(root.getLength() != 1) {
+				return isValidFormat;
+			}
+			NodeList spectrumList = document.getElementsByTagName(IConstants.NODE_SPECTRUM_LIST);
+			if(spectrumList.getLength() > 0) {
+				Element element = (Element)spectrumList.item(0);
+				int spectrumCount = Integer.parseInt(element.getAttribute("count"));
+				if(spectrumCount > 1) {
+					isValidFormat = true;
+				}
 			}
 		} catch(Exception e) {
 			// Print no exception.

diff --git a/...se/chemclipse/msd/converter/supplier/mzdata/converter/MassSpectrumMagicNumberMatcher.java b/...se/chemclipse/msd/converter/supplier/mzdata/converter/MassSpectrumMagicNumberMatcher.java
@@ -14,18 +14,15 @@
 
 import java.io.File;
 
-import javax.xml.bind.JAXBContext;
-import javax.xml.bind.Unmarshaller;
 import javax.xml.parsers.DocumentBuilder;
 import javax.xml.parsers.DocumentBuilderFactory;
 
 import org.eclipse.chemclipse.converter.core.AbstractMagicNumberMatcher;
 import org.eclipse.chemclipse.converter.core.IMagicNumberMatcher;
-import org.eclipse.chemclipse.msd.converter.supplier.mzdata.internal.io.IFormat;
 import org.eclipse.chemclipse.msd.converter.supplier.mzdata.internal.support.IConstants;
 import org.eclipse.chemclipse.msd.converter.supplier.mzdata.internal.support.SpecificationValidator;
-import org.eclipse.chemclipse.msd.converter.supplier.mzdata.internal.v105.model.MzData;
 import org.w3c.dom.Document;
+import org.w3c.dom.Element;
 import org.w3c.dom.NodeList;
 
 public class MassSpectrumMagicNumberMatcher extends AbstractMagicNumberMatcher implements IMagicNumberMatcher {
@@ -45,13 +42,17 @@ public boolean checkFileFormat(File file) {
 			DocumentBuilderFactory documentBuilderFactory = DocumentBuilderFactory.newInstance();
 			DocumentBuilder documentBuilder = documentBuilderFactory.newDocumentBuilder();
 			Document document = documentBuilder.parse(file);
-			NodeList nodeList = document.getElementsByTagName(IConstants.NODE_MZ_DATA);
-			//
-			JAXBContext jaxbContext = JAXBContext.newInstance(IFormat.CONTEXT_PATH_V_105);
-			Unmarshaller unmarshaller = jaxbContext.createUnmarshaller();
-			MzData mzData = (MzData)unmarshaller.unmarshal(nodeList.item(0));
-			if(mzData.getSpectrumList().getCount() == 1) {
-				isValidFormat = true;
+			NodeList root = document.getElementsByTagName(IConstants.NODE_MZ_DATA);
+			if(root.getLength() != 1) {
+				return isValidFormat;
+			}
+			NodeList spectrumList = document.getElementsByTagName(IConstants.NODE_SPECTRUM_LIST);
+			if(spectrumList.getLength() > 0) {
+				Element element = (Element)spectrumList.item(0);
+				int spectrumCount = Integer.parseInt(element.getAttribute("count"));
+				if(spectrumCount == 1) {
+					isValidFormat = true;
+				}
 			}
 		} catch(Exception e) {
 			// Print no exception.

diff --git a/...g/eclipse/chemclipse/msd/converter/supplier/mzdata/internal/io/ChromatogramWriter105.java b/...g/eclipse/chemclipse/msd/converter/supplier/mzdata/internal/io/ChromatogramWriter105.java
@@ -95,8 +95,6 @@ public void writeChromatogram(File file, IChromatogramMSD chromatogram, IProgres
 			marshaller.marshal(mzData, file);
 		} catch(JAXBException e) {
 			logger.warn(e);
-		} catch(Exception e) {
-			e.printStackTrace();
 		}
 	}
 

diff --git a/...src/org/eclipse/chemclipse/msd/converter/supplier/mzdata/internal/support/IConstants.java b/...src/org/eclipse/chemclipse/msd/converter/supplier/mzdata/internal/support/IConstants.java
@@ -21,4 +21,5 @@ public interface IConstants {
 	String EXPORT_SCANS = "Export Scans";
 	String SCAN = "Scan";
 	String NODE_MZ_DATA = "mzData";
+	String NODE_SPECTRUM_LIST = "spectrumList";
 }
diff --git a/chemclipse/plugins/org.eclipse.chemclipse.msd.converter.supplier.mzml/META-INF/MANIFEST.MF b/chemclipse/plugins/org.eclipse.chemclipse.msd.converter.supplier.mzml/META-INF/MANIFEST.MF
@@ -19,4 +19,7 @@ Bundle-RequiredExecutionEnvironment: JavaSE-1.8
 Bundle-ActivationPolicy: lazy
 Import-Package: org.apache.commons.lang3.tuple;version="3.1.0",
  org.eclipse.chemclipse.xxd.converter.supplier.io.exception
-Export-Package: org.eclipse.chemclipse.msd.converter.supplier.mzml.preferences
+Export-Package: org.eclipse.chemclipse.msd.converter.supplier.mzml.converter.io,
+ org.eclipse.chemclipse.msd.converter.supplier.mzml.internal.converter,
+ org.eclipse.chemclipse.msd.converter.supplier.mzml.internal.v110.model,
+ org.eclipse.chemclipse.msd.converter.supplier.mzml.preferences
diff --git a/...ipse/chemclipse/msd/converter/supplier/mzml/converter/ChromatogramMagicNumberMatcher.java b/...ipse/chemclipse/msd/converter/supplier/mzml/converter/ChromatogramMagicNumberMatcher.java
@@ -14,18 +14,15 @@
 
 import java.io.File;
 
-import javax.xml.bind.JAXBContext;
-import javax.xml.bind.Unmarshaller;
 import javax.xml.parsers.DocumentBuilder;
 import javax.xml.parsers.DocumentBuilderFactory;
 
 import org.eclipse.chemclipse.converter.core.AbstractMagicNumberMatcher;
 import org.eclipse.chemclipse.converter.core.IMagicNumberMatcher;
-import org.eclipse.chemclipse.msd.converter.supplier.mzml.converter.io.IFormat;
 import org.eclipse.chemclipse.msd.converter.supplier.mzml.internal.converter.IConstants;
 import org.eclipse.chemclipse.msd.converter.supplier.mzml.internal.converter.SpecificationValidator;
-import org.eclipse.chemclipse.msd.converter.supplier.mzml.internal.v110.model.MzML;
 import org.w3c.dom.Document;
+import org.w3c.dom.Element;
 import org.w3c.dom.NodeList;
 
 public class ChromatogramMagicNumberMatcher extends AbstractMagicNumberMatcher implements IMagicNumberMatcher {
@@ -45,13 +42,17 @@ public boolean checkFileFormat(File file) {
 			DocumentBuilderFactory documentBuilderFactory = DocumentBuilderFactory.newInstance();
 			DocumentBuilder documentBuilder = documentBuilderFactory.newDocumentBuilder();
 			Document document = documentBuilder.parse(file);
-			NodeList nodeList = document.getElementsByTagName(IConstants.NODE_MZML);
-			//
-			JAXBContext jaxbContext = JAXBContext.newInstance(IFormat.CONTEXT_PATH_V_110);
-			Unmarshaller unmarshaller = jaxbContext.createUnmarshaller();
-			MzML mzML = (MzML)unmarshaller.unmarshal(nodeList.item(0));
-			if(mzML.getRun().getChromatogramList() != null) {
-				isValidFormat = true;
+			NodeList root = document.getElementsByTagName(IConstants.NODE_MZML);
+			if(root.getLength() != 1) {
+				return isValidFormat;
+			}
+			NodeList chromatogramList = document.getElementsByTagName(IConstants.NODE_CHROMATOGRAM_LIST);
+			if(chromatogramList.getLength() > 0) {
+				Element element = (Element)chromatogramList.item(0);
+				int chromatogramCount = Integer.parseInt(element.getAttribute("count"));
+				if(chromatogramCount > 0) {
+					isValidFormat = true;
+				}
 			}
 		} catch(Exception e) {
 			// fail silently

diff --git a/...ipse/chemclipse/msd/converter/supplier/mzml/converter/MassSpectrumMagicNumberMatcher.java b/...ipse/chemclipse/msd/converter/supplier/mzml/converter/MassSpectrumMagicNumberMatcher.java
@@ -14,19 +14,15 @@
 
 import java.io.File;
 
-import javax.xml.bind.JAXBContext;
-import javax.xml.bind.Unmarshaller;
 import javax.xml.parsers.DocumentBuilder;
 import javax.xml.parsers.DocumentBuilderFactory;
 
 import org.eclipse.chemclipse.converter.core.AbstractMagicNumberMatcher;
 import org.eclipse.chemclipse.converter.core.IMagicNumberMatcher;
-import org.eclipse.chemclipse.msd.converter.supplier.mzml.converter.io.IFormat;
 import org.eclipse.chemclipse.msd.converter.supplier.mzml.internal.converter.IConstants;
 import org.eclipse.chemclipse.msd.converter.supplier.mzml.internal.converter.SpecificationValidator;
-import org.eclipse.chemclipse.msd.converter.supplier.mzml.internal.v110.model.MzML;
-import org.eclipse.chemclipse.msd.converter.supplier.mzml.internal.v110.model.SpectrumListType;
 import org.w3c.dom.Document;
+import org.w3c.dom.Element;
 import org.w3c.dom.NodeList;
 
 public class MassSpectrumMagicNumberMatcher extends AbstractMagicNumberMatcher implements IMagicNumberMatcher {
@@ -46,14 +42,17 @@ public boolean checkFileFormat(File file) {
 			DocumentBuilderFactory documentBuilderFactory = DocumentBuilderFactory.newInstance();
 			DocumentBuilder documentBuilder = documentBuilderFactory.newDocumentBuilder();
 			Document document = documentBuilder.parse(file);
-			NodeList nodeList = document.getElementsByTagName(IConstants.NODE_MZML);
-			//
-			JAXBContext jaxbContext = JAXBContext.newInstance(IFormat.CONTEXT_PATH_V_110);
-			Unmarshaller unmarshaller = jaxbContext.createUnmarshaller();
-			MzML mzML = (MzML)unmarshaller.unmarshal(nodeList.item(0));
-			SpectrumListType spectrumList = mzML.getRun().getSpectrumList();
-			if(spectrumList != null && spectrumList.getCount().intValue() == 1) {
-				isValidFormat = true;
+			NodeList root = document.getElementsByTagName(IConstants.NODE_MZML);
+			if(root.getLength() != 1) {
+				return isValidFormat;
+			}
+			NodeList spectrumList = document.getElementsByTagName(IConstants.NODE_SPECTRUM_LIST);
+			if(spectrumList.getLength() > 0) {
+				Element element = (Element)spectrumList.item(0);
+				int spectrumCount = Integer.parseInt(element.getAttribute("count"));
+				if(spectrumCount == 1) {
+					isValidFormat = true;
+				}
 			}
 		} catch(Exception e) {
 			// fail silently

diff --git a/...src/org/eclipse/chemclipse/msd/converter/supplier/mzml/internal/converter/IConstants.java b/...src/org/eclipse/chemclipse/msd/converter/supplier/mzml/internal/converter/IConstants.java
@@ -22,4 +22,6 @@ public interface IConstants {
 	//
 	String NODE_RUN = "run";
 	String NODE_MZML = "mzML";
+	String NODE_CHROMATOGRAM_LIST = "chromatogramList";
+	String NODE_SPECTRUM_LIST = "spectrumList";
 }
diff --git a/...ipse/chemclipse/msd/converter/supplier/mzml/internal/io/ChromatogramReaderVersion110.java b/...ipse/chemclipse/msd/converter/supplier/mzml/internal/io/ChromatogramReaderVersion110.java
@@ -77,12 +77,16 @@ public IChromatogramOverview readOverview(File file, IProgressMonitor monitor) t
 			//
 			RunType run = XmlReader.getMzML(file, contextPath).getRun();
 			for(ChromatogramType chromatogramType : run.getChromatogramList().getChromatogram()) {
-				for(BinaryDataArrayType binaryDataArrayType : chromatogramType.getBinaryDataArrayList().getBinaryDataArray()) {
-					Pair<String, double[]> binaryData = BinaryReader.parseBinaryData(binaryDataArrayType);
-					if(binaryData.getKey().equals("time")) {
-						retentionTimes = binaryData.getValue();
-					} else if(binaryData.getKey().equals("intensity")) {
-						intensities = binaryData.getValue();
+				if(chromatogramType.getId().equals("TIC")) {
+					if(chromatogramType.getCvParam().stream().anyMatch(n -> n.getAccession().equals("MS:1000235") && n.getName().equals("total ion current chromatogram"))) {
+						for(BinaryDataArrayType binaryDataArrayType : chromatogramType.getBinaryDataArrayList().getBinaryDataArray()) {
+							Pair<String, double[]> binaryData = BinaryReader.parseBinaryData(binaryDataArrayType);
+							if(binaryData.getKey().equals("time")) {
+								retentionTimes = binaryData.getValue();
+							} else if(binaryData.getKey().equals("intensity")) {
+								intensities = binaryData.getValue();
+							}
+						}
 					}
 				}
 			}

diff --git a/...ipse/chemclipse/msd/converter/supplier/mzml/internal/io/ChromatogramWriterVersion110.java b/...ipse/chemclipse/msd/converter/supplier/mzml/internal/io/ChromatogramWriterVersion110.java
@@ -21,6 +21,7 @@
 import javax.xml.bind.JAXBContext;
 import javax.xml.bind.JAXBException;
 import javax.xml.bind.Marshaller;
+import javax.xml.datatype.DatatypeConfigurationException;
 import javax.xml.datatype.DatatypeFactory;
 
 import org.eclipse.chemclipse.converter.exceptions.FileIsNotWriteableException;
@@ -169,8 +170,8 @@ public void writeChromatogram(File file, IChromatogramMSD chromatogram, IProgres
 			marshaller.marshal(mzML, file);
 		} catch(JAXBException e) {
 			logger.warn(e);
-		} catch(Exception e) {
-			e.printStackTrace();
+		} catch(DatatypeConfigurationException e) {
+			logger.warn(e);
 		}
 	}
 }
diff --git a/...pse/chemclipse/msd/converter/supplier/mzxml/converter/ChromatogramMagicNumberMatcher.java b/...pse/chemclipse/msd/converter/supplier/mzxml/converter/ChromatogramMagicNumberMatcher.java
@@ -14,17 +14,13 @@
 
 import java.io.File;
 
-import javax.xml.bind.JAXBContext;
-import javax.xml.bind.Unmarshaller;
 import javax.xml.parsers.DocumentBuilder;
 import javax.xml.parsers.DocumentBuilderFactory;
 
 import org.eclipse.chemclipse.converter.core.AbstractMagicNumberMatcher;
 import org.eclipse.chemclipse.converter.core.IMagicNumberMatcher;
 import org.eclipse.chemclipse.msd.converter.supplier.mzxml.internal.io.IConstants;
-import org.eclipse.chemclipse.msd.converter.supplier.mzxml.internal.io.IFormat;
 import org.eclipse.chemclipse.msd.converter.supplier.mzxml.internal.io.SpecificationValidator;
-import org.eclipse.chemclipse.msd.converter.supplier.mzxml.internal.v32.model.MsRun;
 import org.w3c.dom.Document;
 import org.w3c.dom.NodeList;
 
@@ -45,16 +41,15 @@ public boolean checkFileFormat(File file) {
 			DocumentBuilderFactory documentBuilderFactory = DocumentBuilderFactory.newInstance();
 			DocumentBuilder documentBuilder = documentBuilderFactory.newDocumentBuilder();
 			Document document = documentBuilder.parse(file);
-			NodeList nodeList = document.getElementsByTagName(IConstants.NODE_MS_RUN);
-			//
-			JAXBContext jaxbContext = JAXBContext.newInstance(IFormat.CONTEXT_PATH_V_320);
-			Unmarshaller unmarshaller = jaxbContext.createUnmarshaller();
-			MsRun msRun = (MsRun)unmarshaller.unmarshal(nodeList.item(0));
-			if(msRun.getScan().size() > 1) {
-				isValidFormat = true;
+			NodeList root = document.getElementsByTagName(IConstants.NODE_MZXML);
+			if(root.getLength() != 1) {
+				return isValidFormat;
 			}
+			NodeList scanList = document.getElementsByTagName(IConstants.NODE_SCAN);
+			if(scanList.getLength() > 1)
+				isValidFormat = true;
 		} catch(Exception e) {
-			e.printStackTrace();
+			// fail silently
 		}
 		return isValidFormat;
 	}

diff --git a/...pse/chemclipse/msd/converter/supplier/mzxml/converter/MassSpectrumMagicNumberMatcher.java b/...pse/chemclipse/msd/converter/supplier/mzxml/converter/MassSpectrumMagicNumberMatcher.java
@@ -14,17 +14,13 @@
 
 import java.io.File;
 
-import javax.xml.bind.JAXBContext;
-import javax.xml.bind.Unmarshaller;
 import javax.xml.parsers.DocumentBuilder;
 import javax.xml.parsers.DocumentBuilderFactory;
 
 import org.eclipse.chemclipse.converter.core.AbstractMagicNumberMatcher;
 import org.eclipse.chemclipse.converter.core.IMagicNumberMatcher;
 import org.eclipse.chemclipse.msd.converter.supplier.mzxml.internal.io.IConstants;
-import org.eclipse.chemclipse.msd.converter.supplier.mzxml.internal.io.IFormat;
 import org.eclipse.chemclipse.msd.converter.supplier.mzxml.internal.io.SpecificationValidator;
-import org.eclipse.chemclipse.msd.converter.supplier.mzxml.internal.v32.model.MsRun;
 import org.w3c.dom.Document;
 import org.w3c.dom.NodeList;
 
@@ -45,16 +41,15 @@ public boolean checkFileFormat(File file) {
 			DocumentBuilderFactory documentBuilderFactory = DocumentBuilderFactory.newInstance();
 			DocumentBuilder documentBuilder = documentBuilderFactory.newDocumentBuilder();
 			Document document = documentBuilder.parse(file);
-			NodeList nodeList = document.getElementsByTagName(IConstants.NODE_MS_RUN);
-			//
-			JAXBContext jaxbContext = JAXBContext.newInstance(IFormat.CONTEXT_PATH_V_320);
-			Unmarshaller unmarshaller = jaxbContext.createUnmarshaller();
-			MsRun msRun = (MsRun)unmarshaller.unmarshal(nodeList.item(0));
-			if(msRun.getScan().size() == 1) {
-				isValidFormat = true;
+			NodeList root = document.getElementsByTagName(IConstants.NODE_MZXML);
+			if(root.getLength() != 1) {
+				return isValidFormat;
 			}
+			NodeList scanList = document.getElementsByTagName(IConstants.NODE_SCAN);
+			if(scanList.getLength() == 1)
+				isValidFormat = true;
 		} catch(Exception e) {
-			e.printStackTrace();
+			// fail silently
 		}
 		return isValidFormat;
 	}

diff --git a/...mzxml/src/org/eclipse/chemclipse/msd/converter/supplier/mzxml/internal/io/IConstants.java b/...mzxml/src/org/eclipse/chemclipse/msd/converter/supplier/mzxml/internal/io/IConstants.java
@@ -1,5 +1,5 @@
 /*******************************************************************************
- * Copyright (c) 2008, 2018 Lablicate GmbH.
+ * Copyright (c) 2008, 2021 Lablicate GmbH.
  * 
  * All rights reserved.
  * This program and the accompanying materials are made available under the
@@ -20,5 +20,7 @@ public interface IConstants {
 	String EXPORT_SCANS = "Export Scans";
 	String SCAN = "Scan";
 	//
+	String NODE_MZXML = "mzXML";
 	String NODE_MS_RUN = "msRun";
+	String NODE_SCAN = "scan";
 }