X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fws%2Fdbsources%2FEmblXmlSource.java;h=b3b99ca1ed5b156e26e609b06b40bcd6fa0a1a64;hb=a1984b1c8c273ed33c7ce9283039f4027dcae2de;hp=6b6f2ecea0d69f65fcc3a623bc5d93c412cad02a;hpb=fe3cd724aecdeb06a130a502ce3a967ad643f458;p=jalview.git

diff --git a/src/jalview/ws/dbsources/EmblXmlSource.java b/src/jalview/ws/dbsources/EmblXmlSource.java
index 6b6f2ec..b3b99ca 100644
--- a/src/jalview/ws/dbsources/EmblXmlSource.java
+++ b/src/jalview/ws/dbsources/EmblXmlSource.java
@@ -28,6 +28,7 @@ import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.Hashtable;
 import java.util.List;
+import java.util.Locale;
 import java.util.Map;
 import java.util.Map.Entry;
 
@@ -63,12 +64,6 @@ import jalview.xml.binding.embl.EntryType.Feature.Qualifier;
 import jalview.xml.binding.embl.ROOT;
 import jalview.xml.binding.embl.XrefType;
 
-/**
- * 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- * @deprecated endpoint withdrawn August 2020 (JAL-3692), use EmblFlatfileSource
- */
 public abstract class EmblXmlSource extends EbiFileRetrievedProxy
 {
   private static final Regex ACCESSION_REGEX = new Regex("^[A-Z]+[0-9]+");
@@ -102,14 +97,14 @@ public abstract class EmblXmlSource extends EbiFileRetrievedProxy
     try
     {
       reply = dbFetch.fetchDataAsFile(
-              emprefx.toLowerCase() + ":" + query.trim(), "display=xml",
-              "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);
@@ -453,9 +448,8 @@ public abstract class EmblXmlSource extends EbiFileRetrievedProxy
         else
         {
           // final product length truncation check
-          int[] cdsRanges = adjustForProteinLength(translationLength,
-                  exons);
-          dnaToProteinMapping = new Mapping(product, cdsRanges,
+          int[] exons2 = adjustForProteinLength(translationLength, exons);
+          dnaToProteinMapping = new Mapping(product, exons2,
                   new int[]
                   { 1, translationLength }, 3, 1);
           if (product != null)
@@ -574,6 +568,7 @@ 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 +658,7 @@ public abstract class EmblXmlSource extends EbiFileRetrievedProxy
       return listToArray(ranges);
     } catch (ParseException e)
     {
-      Cache.log.warn(
+      Cache.warn(
               String.format("Not parsing inexact CDS location %s in ENA %s",
                       location, accession));
       return new int[] {};
@@ -756,8 +751,7 @@ 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 +768,9 @@ 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;
     }