Merge branch 'develop' into update_212_Dec_merge_with_21125_chamges

[jalview.git] / src / jalview / ws / dbsources / EmblXmlSource.java

diff --git a/src/jalview/ws/dbsources/EmblXmlSource.java b/src/jalview/ws/dbsources/EmblXmlSource.java
index 6b6f2ec..ba46a6c 100644 (file)
--- a/src/jalview/ws/dbsources/EmblXmlSource.java
+++ b/src/jalview/ws/dbsources/EmblXmlSource.java
@@ -20,6 +20,8 @@
  */
 package jalview.ws.dbsources;
 
+import java.util.Locale;
+
 import java.io.File;
 import java.io.FileInputStream;
 import java.io.InputStream;
@@ -39,10 +41,9 @@ import javax.xml.stream.XMLInputFactory;
 import javax.xml.stream.XMLStreamException;
 import javax.xml.stream.XMLStreamReader;
 
-import com.stevesoft.pat.Regex;
 
 import jalview.analysis.SequenceIdMatcher;
-import jalview.bin.Cache;
+import jalview.bin.Console;
 import jalview.datamodel.Alignment;
 import jalview.datamodel.AlignmentI;
 import jalview.datamodel.DBRefEntry;
@@ -56,6 +57,8 @@ import jalview.util.DBRefUtils;
 import jalview.util.DnaUtils;
 import jalview.util.MapList;
 import jalview.util.MappingUtils;
+import jalview.util.MessageManager;
+import jalview.util.Platform;
 import jalview.ws.ebi.EBIFetchClient;
 import jalview.xml.binding.embl.EntryType;
 import jalview.xml.binding.embl.EntryType.Feature;
@@ -63,15 +66,19 @@ import jalview.xml.binding.embl.EntryType.Feature.Qualifier;
 import jalview.xml.binding.embl.ROOT;
 import jalview.xml.binding.embl.XrefType;
 
+import com.stevesoft.pat.Regex;
+
 /**
  * Provides XML binding and parsing of EMBL or EMBLCDS records retrieved from
  * (e.g.) {@code https://www.ebi.ac.uk/ena/data/view/x53828&display=xml}.
  * 
  * @deprecated endpoint withdrawn August 2020 (JAL-3692), use EmblFlatfileSource
  */
+@Deprecated
 public abstract class EmblXmlSource extends EbiFileRetrievedProxy
 {
-  private static final Regex ACCESSION_REGEX = new Regex("^[A-Z]+[0-9]+");
+  // TODO: delete class or update tyhis validator for 2.12 style Platform.regex
+    private static final Regex ACCESSION_REGEX = Platform.newRegex("^[A-Z]+[0-9]+");
 
   /*
    * JAL-1856 Embl returns this text for query not found
@@ -102,14 +109,14 @@ public abstract class EmblXmlSource extends EbiFileRetrievedProxy
     try
     {
       reply = dbFetch.fetchDataAsFile(
-              emprefx.toLowerCase() + ":" + query.trim(), "display=xml",
+              emprefx.toLowerCase(Locale.ROOT) + ":" + query.trim(), "display=xml",
               "xml");
     } catch (Exception e)
     {
       stopQuery();
       throw new Exception(
               String.format("EBI EMBL XML retrieval failed for %s:%s",
-                      emprefx.toLowerCase(), query.trim()),
+                      emprefx.toLowerCase(Locale.ROOT), query.trim()),
               e);
     }
     return getEmblSequenceRecords(emprefx, query, reply);
@@ -574,6 +581,7 @@ public abstract class EmblXmlSource extends EbiFileRetrievedProxy
               proteinSeq = new Sequence(proteinSeqName,
                       product.getSequenceAsString());
               matcher.add(proteinSeq);
+              proteinSeq.setDescription(product.getDescription());
               peptides.add(proteinSeq);
             }
             dnaToProteinMapping.setTo(proteinSeq);
@@ -663,7 +671,7 @@ public abstract class EmblXmlSource extends EbiFileRetrievedProxy
       return listToArray(ranges);
     } catch (ParseException e)
     {
-      Cache.log.warn(
+      Console.warn(
               String.format("Not parsing inexact CDS location %s in ENA %s",
                       location, accession));
       return new int[] {};
@@ -756,8 +764,7 @@ public abstract class EmblXmlSource extends EbiFileRetrievedProxy
 
   /**
    * Truncates (if necessary) the exon intervals to match 3 times the length of
-   * the protein; also accepts 3 bases longer (for stop codon not included in
-   * protein)
+   * the protein (including truncation for stop codon included in exon)
    * 
    * @param proteinLength
    * @param exon
@@ -774,11 +781,9 @@ public abstract class EmblXmlSource extends EbiFileRetrievedProxy
     int exonLength = MappingUtils.getLength(Arrays.asList(exon));
 
     /*
-     * if exon length matches protein, or is shorter, or longer by the 
-     * length of a stop codon (3 bases), then leave it unchanged
+     * if exon length matches protein, or is shorter, then leave it unchanged
      */
-    if (expectedCdsLength >= exonLength
-            || expectedCdsLength == exonLength - 3)
+    if (expectedCdsLength >= exonLength)
     {
       return exon;
     }