[LMS-2401] parse analysis procedures from GE XML file

SVN: 22256

sanofi/dist/etc/sanofi-dropbox/dropbox-all-in-one-with-library.py

@@ -373,7 +373,8 @@ if incoming.isDirectory():
     analysisFile = util.findFileByExt(incoming, "xml")
     if analysisFile is not None:
         analysisCSVFile = File(analysisFile.getPath() + ".csv")
-        GEExplorerImageAnalysisResultParser(analysisFile.getPath()).writeCSV(analysisCSVFile)
+        geXmlParser = GEExplorerImageAnalysisResultParser(analysisFile.getPath())
+        geXmlParser.writeCSV(analysisCSVFile)
         
         featureProps = Properties()
         featureProps.setProperty("separator", ",")
@@ -381,7 +382,7 @@ if incoming.isDirectory():
         featureProps.setProperty("well-name-col", "Well")
         
         analysisDataSetDetails = factory.createFeatureVectorRegistrationDetails(analysisCSVFile.getPath(), featureProps)
-        analysisProcedureCode = util.extractFileBasename(analysisFile.getName())
+        analysisProcedureCode = geXmlParser.getAnalysisProcedureName()
         analysisDataSetDetails.getDataSetInformation().setAnalysisProcedure(analysisProcedureCode)
         analysisDataSet = transaction.createNewDataSet(analysisDataSetDetails)
         analysisDataSet.setSample(imageDataSet.getSample())

sanofi/sourceTest/java/ch/systemsx/cisd/sanofi/dss/test/SanofiDropboxJythonTest.java

@@ -404,6 +404,8 @@ public class SanofiDropboxJythonTest extends AbstractJythonDataSetHandlerTest
         NewExternalData analysisDataSet = dataSetsRegistered.get(2);
         assertEquals(ANALYSIS_DATA_SET_CODE, analysisDataSet.getCode());
         assertEquals(ANALYSIS_DATA_SET_TYPE, analysisDataSet.getDataSetType());
+        assertEquals("NRF2ProdAnalysis_1.1_MatlabGeneral_1.1",
+                extractAnalysisProcedureCode(analysisDataSet));
 
 
         AssertionUtil
@@ -414,6 +416,14 @@ public class SanofiDropboxJythonTest extends AbstractJythonDataSetHandlerTest
         context.assertIsSatisfied();
     }
 
+    private String extractAnalysisProcedureCode(NewExternalData analysisDataSet)
+    {
+        NewProperty property =
+                EntityHelper.tryFindProperty(analysisDataSet.getDataSetProperties(),
+                        ScreeningConstants.ANALYSIS_PROCEDURE);
+        return (property != null) ? property.getValue() : null;
+    }
+
     @Test
     public void testFatalErrorSentToAdmin() throws IOException
     {

screening/source/java/ch/systemsx/cisd/openbis/dss/etl/custom/geexplorer/GEExplorerImageAnalysisResultParser.java

@@ -31,6 +31,7 @@ import javax.xml.parsers.ParserConfigurationException;
 import javax.xml.parsers.SAXParser;
 import javax.xml.parsers.SAXParserFactory;
 
+import org.apache.commons.lang.StringUtils;
 import org.xml.sax.Attributes;
 import org.xml.sax.SAXException;
 import org.xml.sax.helpers.DefaultHandler;
@@ -49,17 +50,23 @@ public class GEExplorerImageAnalysisResultParser extends DefaultHandler
 
     private static final String WELL_TAG = "Well";
 
-    private static final String TABLE_TAG = "Table";
-
     private static final String DATA_TAG = "Data";
 
+    private static final String ANALYSIS_TAG = "AutoLeadAnalysisProtocol";
+
+    private static final String ASSAY_TAG = "AutoLeadAssay";
+
     private enum State
     {
-        INIT, DATA, TABLE, WELL, FINISHED
+        INIT, ANALYSIS, DATA, TABLE, WELL, FINISHED
     }
 
     private State state = State.INIT;
 
+    private String analysisProtocolWithVersion = StringUtils.EMPTY;
+
+    private String assayWithVersion = StringUtils.EMPTY;
+
     private String currentWellId;
 
     private HashSet<String> wells = new HashSet<String>();
@@ -99,8 +106,16 @@ public class GEExplorerImageAnalysisResultParser extends DefaultHandler
                 }
                 break;
             case DATA:
-                if (TABLE_TAG.equals(qName) && "Wells Summary".equals(attributes.getValue("title")))
+                if (ANALYSIS_TAG.equals(qName))
+                {
+                    analysisProtocolWithVersion = parseNameAndVersion(attributes);
+                    state = State.ANALYSIS;
+                }
+                break;
+            case ANALYSIS:
+                if (ASSAY_TAG.equals(qName))
                 {
+                    assayWithVersion = parseNameAndVersion(attributes);
                     state = State.TABLE;
                 }
                 break;
@@ -224,4 +239,35 @@ public class GEExplorerImageAnalysisResultParser extends DefaultHandler
         out.close();
     }
 
+    public String getAnalysisProcedureName()
+    {
+        if (StringUtils.isEmpty(assayWithVersion))
+        {
+            return analysisProtocolWithVersion;
+        } else
+        {
+            return analysisProtocolWithVersion + "_" + assayWithVersion;
+        }
+    }
+
+    private String parseNameAndVersion(Attributes attributes)
+    {
+        String name = parseAttribute("name", attributes);
+        String version = parseAttribute("version", attributes);
+        if (StringUtils.isEmpty(version))
+        {
+            return name;
+        } else
+        {
+            return name + "_" + version;
+        }
+    }
+
+    private String parseAttribute(String attrName, Attributes attributes)
+    {
+        String attrValue = attributes.getValue(attrName);
+        return StringUtils.trimToEmpty(attrValue);
+    }
+
+
 }