From: Jim Procter <j.procter@dundee.ac.uk>
Date: Thu, 23 Sep 2021 16:24:16 +0000 (+0100)
Subject: JAL-3821 ported to refactored FlatFile parser
X-Git-Tag: Release_2_11_2_0~32^2^2~1
X-Git-Url: http://source.jalview.org/gitweb/?a=commitdiff_plain;h=9ae61a06b581ffba65614a2f08acc36e7f7685e7;p=jalview.git

JAL-3821 ported to refactored FlatFile parser
---

diff --git a/src/jalview/io/FlatFile.java b/src/jalview/io/FlatFile.java
index 55fdd37..9e5c652 100644
--- a/src/jalview/io/FlatFile.java
+++ b/src/jalview/io/FlatFile.java
@@ -7,6 +7,7 @@ import java.util.Arrays;
 import java.util.HashMap;
 import java.util.Hashtable;
 import java.util.List;
+import java.util.Locale;
 import java.util.Map;
 import java.util.Map.Entry;
 import java.util.TreeMap;
@@ -159,6 +160,13 @@ public abstract class FlatFile extends AlignFile
   }
 
   /*
+   * when true, interpret the mol_type 'source' feature attribute
+   * and generate an RNA sequence from the DNA record
+   */
+  protected boolean produceRna=true;
+    
+
+  /*
    * values parsed from the data file
    */
   protected String sourceDb;
@@ -173,6 +181,8 @@ public abstract class FlatFile extends AlignFile
 
   protected List<DBRefEntry> dbrefs;
 
+  protected boolean sequenceStringIsRNA=false;
+
   protected String sequenceString;
 
   protected Map<String, CdsData> cds;
@@ -199,6 +209,52 @@ public abstract class FlatFile extends AlignFile
   }
 
   /**
+   * process attributes for 'source' until the next FT feature entry
+   * only interested in 'mol_type'
+   * @param tokens
+   * @return
+   * @throws IOException
+   */
+  private String parseSourceQualifiers(String[] tokens) throws IOException
+  {
+    if (!"source".equals(tokens[0]))
+    {
+      throw (new RuntimeException("Not given a 'source' qualifier line"));
+    }
+    // search for mol_type attribute
+
+    StringBuilder sb = new StringBuilder().append(tokens[1]); // extent of
+                                                              // sequence
+
+    String line = parseFeatureQualifier(sb, false);
+    while (line != null)
+    {
+      if (!line.startsWith("FT    ")) // four spaces, end of this feature table
+                                      // entry
+      {
+        return line;
+      }
+
+      // case sensitive ?
+      int p = line.indexOf("\\mol_type");
+      int qs = line.indexOf("\"", p);
+      int qe = line.indexOf("\"", qs + 1);
+      String qualifier=line.substring(qs,qe).toLowerCase(Locale.ROOT);
+      if (qualifier.indexOf("rna") > -1)
+      {
+        sequenceStringIsRNA = true;
+      }
+      if (qualifier.indexOf("dna") > -1)
+      {
+        sequenceStringIsRNA = false;
+      }
+      line=parseFeatureQualifier(sb, false);
+    }
+    return line;
+  }
+
+   
+  /**
    * Parses one (GenBank or EMBL format) CDS feature, saves the parsed data, and
    * returns the next line
    * 
@@ -335,6 +391,12 @@ public abstract class FlatFile extends AlignFile
     {
       name = this.sourceDb + "|" + name;
     }
+
+    if (produceRna && sequenceStringIsRNA)
+    {
+      sequenceString = sequenceString.replace('T', 'U').replace('t', 'u');
+    }
+
     SequenceI seq = new Sequence(name, this.sequenceString);
     seq.setDescription(this.description);
 
@@ -738,11 +800,14 @@ public abstract class FlatFile extends AlignFile
   protected String parseFeature(String line) throws IOException
   {
     String[] tokens = line.trim().split(WHITESPACE);
-    if (tokens.length < 2 || !"CDS".equals(tokens[0]))
+    if (tokens.length < 2 || (!"CDS".equals(tokens[0]) && (!"source".equals(tokens[0]))))
     {
       return nextLine();
     }
-
+    if (tokens[0].equals("source"))
+    {
+      return parseSourceQualifiers(tokens);
+    }
     return parseCDSFeature(tokens[1]);
   }
 }