import jalview.datamodel.features.FeatureMatcherSet;
import jalview.datamodel.features.FeatureMatcherSetI;
import jalview.gui.Desktop;
-import jalview.io.gff.GffHelperBase;
import jalview.io.gff.GffHelperFactory;
import jalview.io.gff.GffHelperI;
import jalview.schemes.FeatureColour;
*/
public class FeaturesFile extends AlignFile implements FeaturesSourceI
{
+ /*
+ * map-valued attributes are prefixed with this for output to GFF3;
+ * the prefix is removed if found on reading
+ */
+ public static final String MAP_ATTRIBUTE_PREFIX = "jvmap_";
+
private static final String TAB_REGEX = "\\t";
private static final String STARTGROUP = "STARTGROUP";
String phase = sf.getPhase();
out.append(phase == null ? "." : phase);
- // miscellaneous key-values (GFF column 9)
- String attributes = sf.getAttributes();
- if (attributes != null)
+ if (sf.otherDetails != null && !sf.otherDetails.isEmpty())
+ {
+ Map<String, Object> map = sf.otherDetails;
+ formatAttributes(out, map);
+ }
+ }
+
+ /**
+ * A helper method that outputs attributes stored in the map as
+ * semicolon-delimited values e.g.
+ *
+ * <pre>
+ * AC_Male=0;AF_NFE=0.00000e 00;Hom_FIN=0;GQ_MEDIAN=9
+ * </pre>
+ *
+ * A map-valued attribute is formatted as a comma-delimited list within braces,
+ * for example
+ *
+ * <pre>
+ * jvmap_CSQ={ALLELE_NUM=1,UNIPARC=UPI0002841053,Feature=ENST00000585561}
+ * </pre>
+ *
+ * The {@code jvmap_} prefix designates a values map and is removed if the value
+ * is parsed when read in. (The GFF3 specification allows 'semi-structured data'
+ * to be represented provided the attribute name begins with a lower case
+ * letter.)
+ *
+ * @param sb
+ * @param map
+ * @see http://gmod.org/wiki/GFF3#GFF3_Format
+ */
+ void formatAttributes(StringBuilder sb, Map<String, Object> map)
+ {
+ sb.append(TAB);
+ boolean first = true;
+ for (String key : map.keySet())
+ {
+ if (SequenceFeature.STRAND.equals(key)
+ || SequenceFeature.PHASE.equals(key))
+ {
+ /*
+ * values stashed in map but output to their own columns
+ */
+ continue;
+ }
+ {
+ if (!first)
+ {
+ sb.append(";");
+ }
+ }
+ first = false;
+ Object value = map.get(key);
+ if (value instanceof Map<?, ?>)
+ {
+ formatMapAttribute(sb, key, (Map<?, ?>) value);
+ }
+ else
+ {
+ String formatted = StringUtils.urlEncode(value.toString(),
+ GffHelperI.GFF_ENCODABLE);
+ sb.append(key).append("=").append(formatted);
+ }
+ }
+ }
+
+ /**
+ * Formats the map entries as
+ *
+ * <pre>
+ * jvmap_key={key1=value1,key2=value2,...}
+ * </pre>
+ *
+ * and appends this to the string buffer
+ *
+ * @param sb
+ * @param key
+ * @param map
+ */
+ private void formatMapAttribute(StringBuilder sb, String key,
+ Map<?, ?> map)
+ {
+ if (map == null || map.isEmpty())
+ {
+ return;
+ }
+
+ /*
+ * AbstractMap.toString would be a shortcut here, but more reliable
+ * to code the required format in case toString changes in future
+ */
+ sb.append(MAP_ATTRIBUTE_PREFIX).append(key).append("={");
+ boolean first = true;
+ for (Entry<?, ?> entry : map.entrySet())
{
- out.append(TAB).append(attributes);
+ if (!first)
+ {
+ sb.append(",");
+ }
+ first = false;
+ sb.append(entry.getKey().toString()).append("=");
+ String formatted = StringUtils.urlEncode(entry.getValue().toString(),
+ GffHelperI.GFF_ENCODABLE);
+ sb.append(formatted);
}
+ sb.append("}");
}
/**
* format)
*
* @param alignedRegions
- * a list of "Align fromStart toStart fromCount"
+ * a list of "Align fromStart toStart fromCount"
* @param mapIsFromCdna
- * if true, 'from' is dna, else 'from' is protein
+ * if true, 'from' is dna, else 'from' is protein
* @param strand
- * either 1 (forward) or -1 (reverse)
+ * either 1 (forward) or -1 (reverse)
* @return
* @throws IOException
*/
}
/**
- * Process the 'column 9' data of the GFF file. This is less formally defined,
- * and its interpretation will vary depending on the tool that has generated
- * it.
- *
- * @param attributes
- * @param sf
- */
- protected void processGffColumnNine(String attributes, SequenceFeature sf)
- {
- sf.setAttributes(attributes);
-
- /*
- * Parse attributes in column 9 and add them to the sequence feature's
- * 'otherData' table; use Note as a best proxy for description
- */
- char nameValueSeparator = gffVersion == 3 ? '=' : ' ';
- // TODO check we don't break GFF2 values which include commas here
- Map<String, List<String>> nameValues = GffHelperBase
- .parseNameValuePairs(attributes, ";", nameValueSeparator, ",");
- for (Entry<String, List<String>> attr : nameValues.entrySet())
- {
- String values = StringUtils.listToDelimitedString(attr.getValue(),
- "; ");
- sf.setValue(attr.getKey(), values);
- if (NOTE.equals(attr.getKey()))
- {
- sf.setDescription(values);
- }
- }
- }
-
- /**
* After encountering ##fasta in a GFF3 file, process the remainder of the
* file as FAST sequence data. Any placeholder sequences created during
* feature parsing are updated with the actual sequences.
git://source.jalview.org / jalview.git / blobdiff