import jalview.analysis.AlignmentUtils;
import jalview.analysis.SequenceIdMatcher;
import jalview.api.AlignViewportI;
+import jalview.api.FeatureColourI;
import jalview.api.FeaturesSourceI;
import jalview.datamodel.AlignedCodonFrame;
import jalview.datamodel.Alignment;
import jalview.io.gff.GffHelperBase;
import jalview.io.gff.GffHelperFactory;
import jalview.io.gff.GffHelperI;
-import jalview.schemes.AnnotationColourGradient;
-import jalview.schemes.GraduatedColor;
-import jalview.schemes.UserColourScheme;
-import jalview.util.Format;
+import jalview.schemes.FeatureColour;
+import jalview.util.ColorUtils;
import jalview.util.MapList;
import jalview.util.ParseHtmlBodyAndLinks;
import jalview.util.StringUtils;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
+import java.util.Collections;
+import java.util.Comparator;
import java.util.HashMap;
-import java.util.Iterator;
+import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
-import java.util.StringTokenizer;
+import java.util.Set;
/**
* Parses and writes features files, which may be in Jalview, GFF2 or GFF3
private static final String NOTE = "Note";
- protected static final String FRAME = "FRAME";
-
protected static final String TAB = "\t";
protected static final String GFF_VERSION = "##gff-version";
+ private static final Comparator<String> SORT_NULL_LAST = new Comparator<String>()
+ {
+ @Override
+ public int compare(String o1, String o2)
+ {
+ if (o1 == null)
+ {
+ return o2 == null ? 0 : 1;
+ }
+ return (o2 == null ? -1 : o1.compareTo(o2));
+ }
+ };
+
private AlignmentI lastmatchedAl = null;
private SequenceIdMatcher matcher = null;
* Constructor which does not parse the file immediately
*
* @param inFile
- * @param type
+ * @param paste
* @throws IOException
*/
- public FeaturesFile(String inFile, String type) throws IOException
+ public FeaturesFile(String inFile, DataSourceType paste)
+ throws IOException
{
- super(false, inFile, type);
+ super(false, inFile, paste);
}
/**
* @param type
* @throws IOException
*/
- public FeaturesFile(boolean parseImmediately, String inFile, String type)
+ public FeaturesFile(boolean parseImmediately, String inFile,
+ DataSourceType type)
throws IOException
{
super(parseImmediately, inFile, type);
* - process html strings into plain text
* @return true if features were added
*/
- public boolean parse(AlignmentI align, Map<String, Object> colours,
- boolean removeHTML)
+ public boolean parse(AlignmentI align,
+ Map<String, FeatureColourI> colours, boolean removeHTML)
{
return parse(align, colours, removeHTML, false);
}
* - when true, ID matches to compound sequence IDs are allowed
* @return true if features were added
*/
- public boolean parse(AlignmentI align, Map<String, Object> colours,
- boolean removeHTML, boolean relaxedIdmatching)
+ public boolean parse(AlignmentI align,
+ Map<String, FeatureColourI> colours, boolean removeHTML,
+ boolean relaxedIdmatching)
{
Map<String, String> gffProps = new HashMap<String, String>();
/*
else if (ft.equalsIgnoreCase("endgroup"))
{
// We should check whether this is the current group,
- // but at present theres no way of showing more than 1 group
+ // but at present there's no way of showing more than 1 group
featureGroup = null;
}
else
{
- parseFeatureColour(line, ft, gffColumns, colours);
+ String colscheme = gffColumns[1];
+ FeatureColourI colour = FeatureColour
+ .parseJalviewFeatureColour(colscheme);
+ if (colour != null)
+ {
+ colours.put(ft, colour);
+ }
}
continue;
}
return false;
}
+ /*
+ * experimental - add any dummy sequences with features to the alignment
+ * - we need them for Ensembl feature extraction - though maybe not otherwise
+ */
+ for (SequenceI newseq : newseqs)
+ {
+ if (newseq.getFeatures().hasFeatures())
+ {
+ align.addSequence(newseq);
+ }
+ }
return true;
}
* @param featureGroup
*/
protected boolean parseJalviewFeature(String line, String[] gffColumns,
- AlignmentI alignment, Map<String, Object> featureColours,
+ AlignmentI alignment, Map<String, FeatureColourI> featureColours,
boolean removeHTML, boolean relaxedIdMatching, String featureGroup)
{
/*
* Perhaps an old style groups file with no colours -
* synthesize a colour from the feature type
*/
- UserColourScheme ucs = new UserColourScheme(ft);
- featureColours.put(ft, ucs.findColour('A'));
+ Color colour = ColorUtils.createColourFromName(ft);
+ featureColours.put(ft, new FeatureColour(colour));
}
SequenceFeature sf = new SequenceFeature(ft, desc, "", startPos,
endPos, featureGroup);
}
/**
- * Process a feature type colour specification
- *
- * @param line
- * the current input line (for error messages only)
- * @param featureType
- * the first token on the line
- * @param gffColumns
- * holds tokens on the line
- * @param colours
- * map to which to add derived colour specification
- */
- protected void parseFeatureColour(String line, String featureType,
- String[] gffColumns, Map<String, Object> colours)
- {
- Object colour = null;
- String colscheme = gffColumns[1];
- if (colscheme.indexOf("|") > -1
- || colscheme.trim().equalsIgnoreCase("label"))
- {
- colour = parseGraduatedColourScheme(line, colscheme);
- }
- else
- {
- UserColourScheme ucs = new UserColourScheme(colscheme);
- colour = ucs.findColour('A');
- }
- if (colour != null)
- {
- colours.put(featureType, colour);
- }
- }
-
- /**
- * Parse a Jalview graduated colour descriptor
- *
- * @param line
- * @param colourDescriptor
- * @return
- */
- protected GraduatedColor parseGraduatedColourScheme(String line,
- String colourDescriptor)
- {
- // Parse '|' separated graduated colourscheme fields:
- // [label|][mincolour|maxcolour|[absolute|]minvalue|maxvalue|thresholdtype|thresholdvalue]
- // can either provide 'label' only, first is optional, next two
- // colors are required (but may be
- // left blank), next is optional, nxt two min/max are required.
- // first is either 'label'
- // first/second and third are both hexadecimal or word equivalent
- // colour.
- // next two are values parsed as floats.
- // fifth is either 'above','below', or 'none'.
- // sixth is a float value and only required when fifth is either
- // 'above' or 'below'.
- StringTokenizer gcol = new StringTokenizer(colourDescriptor, "|", true);
- // set defaults
- float min = Float.MIN_VALUE, max = Float.MAX_VALUE;
- boolean labelCol = false;
- // Parse spec line
- String mincol = gcol.nextToken();
- if (mincol == "|")
- {
- System.err
- .println("Expected either 'label' or a colour specification in the line: "
- + line);
- return null;
- }
- String maxcol = null;
- if (mincol.toLowerCase().indexOf("label") == 0)
- {
- labelCol = true;
- mincol = (gcol.hasMoreTokens() ? gcol.nextToken() : null); // skip '|'
- mincol = (gcol.hasMoreTokens() ? gcol.nextToken() : null);
- }
- String abso = null, minval, maxval;
- if (mincol != null)
- {
- // at least four more tokens
- if (mincol.equals("|"))
- {
- mincol = "";
- }
- else
- {
- gcol.nextToken(); // skip next '|'
- }
- // continue parsing rest of line
- maxcol = gcol.nextToken();
- if (maxcol.equals("|"))
- {
- maxcol = "";
- }
- else
- {
- gcol.nextToken(); // skip next '|'
- }
- abso = gcol.nextToken();
- gcol.nextToken(); // skip next '|'
- if (abso.toLowerCase().indexOf("abso") != 0)
- {
- minval = abso;
- abso = null;
- }
- else
- {
- minval = gcol.nextToken();
- gcol.nextToken(); // skip next '|'
- }
- maxval = gcol.nextToken();
- if (gcol.hasMoreTokens())
- {
- gcol.nextToken(); // skip next '|'
- }
- try
- {
- if (minval.length() > 0)
- {
- min = Float.valueOf(minval);
- }
- } catch (Exception e)
- {
- System.err
- .println("Couldn't parse the minimum value for graduated colour for type ("
- + colourDescriptor
- + ") - did you misspell 'auto' for the optional automatic colour switch ?");
- e.printStackTrace();
- }
- try
- {
- if (maxval.length() > 0)
- {
- max = Float.valueOf(maxval);
- }
- } catch (Exception e)
- {
- System.err
- .println("Couldn't parse the maximum value for graduated colour for type ("
- + colourDescriptor + ")");
- e.printStackTrace();
- }
- }
- else
- {
- // add in some dummy min/max colours for the label-only
- // colourscheme.
- mincol = "FFFFFF";
- maxcol = "000000";
- }
-
- GraduatedColor colour = null;
- try
- {
- colour = new GraduatedColor(
- new UserColourScheme(mincol).findColour('A'),
- new UserColourScheme(maxcol).findColour('A'), min, max);
- } catch (Exception e)
- {
- System.err.println("Couldn't parse the graduated colour scheme ("
- + colourDescriptor + ")");
- e.printStackTrace();
- }
- if (colour != null)
- {
- colour.setColourByLabel(labelCol);
- colour.setAutoScaled(abso == null);
- // add in any additional parameters
- String ttype = null, tval = null;
- if (gcol.hasMoreTokens())
- {
- // threshold type and possibly a threshold value
- ttype = gcol.nextToken();
- if (ttype.toLowerCase().startsWith("below"))
- {
- colour.setThreshType(AnnotationColourGradient.BELOW_THRESHOLD);
- }
- else if (ttype.toLowerCase().startsWith("above"))
- {
- colour.setThreshType(AnnotationColourGradient.ABOVE_THRESHOLD);
- }
- else
- {
- colour.setThreshType(AnnotationColourGradient.NO_THRESHOLD);
- if (!ttype.toLowerCase().startsWith("no"))
- {
- System.err.println("Ignoring unrecognised threshold type : "
- + ttype);
- }
- }
- }
- if (colour.getThreshType() != AnnotationColourGradient.NO_THRESHOLD)
- {
- try
- {
- gcol.nextToken();
- tval = gcol.nextToken();
- colour.setThresh(new Float(tval).floatValue());
- } catch (Exception e)
- {
- System.err.println("Couldn't parse threshold value as a float: ("
- + tval + ")");
- e.printStackTrace();
- }
- }
- // parse the thresh-is-min token ?
- if (gcol.hasMoreTokens())
- {
- System.err
- .println("Ignoring additional tokens in parameters in graduated colour specification\n");
- while (gcol.hasMoreTokens())
- {
- System.err.println("|" + gcol.nextToken());
- }
- System.err.println("\n");
- }
- }
- return colour;
- }
-
- /**
* clear any temporary handles used to speed up ID matching
*/
protected void resetMatcher()
ParseHtmlBodyAndLinks parsed = new ParseHtmlBodyAndLinks(
sf.getDescription(), removeHTML, newline);
- sf.description = (removeHTML) ? parsed.getNonHtmlContent()
- : sf.description;
+ if (removeHTML)
+ {
+ sf.setDescription(parsed.getNonHtmlContent());
+ }
+
for (String link : parsed.getLinks())
{
sf.addLink(link);
* @return features file contents
*/
public String printJalviewFormat(SequenceI[] sequences,
- Map<String, Object> visible)
+ Map<String, FeatureColourI> visible)
{
return printJalviewFormat(sequences, visible, true, true);
}
* @return features file contents
*/
public String printJalviewFormat(SequenceI[] sequences,
- Map<String, Object> visible, boolean visOnly, boolean nonpos)
+ Map<String, FeatureColourI> visible, boolean visOnly,
+ boolean nonpos)
{
- StringBuilder out = new StringBuilder(256);
- boolean featuresGen = false;
- if (visOnly && !nonpos && (visible == null || visible.size() < 1))
+ if (visOnly && !nonpos && (visible == null || visible.isEmpty()))
{
// no point continuing.
return "No Features Visible";
}
- if (visible != null && visOnly)
+ /*
+ * write out feature colours (if we know them)
+ */
+ // TODO: decide if feature links should also be written here ?
+ StringBuilder out = new StringBuilder(256);
+ if (visible != null && visOnly) // todo why visOnly test?
{
- // write feature colours only if we're given them and we are generating
- // viewed features
- // TODO: decide if feature links should also be written here ?
- Iterator<String> en = visible.keySet().iterator();
- String featureType, color;
- while (en.hasNext())
+ for (Entry<String, FeatureColourI> featureColour : visible.entrySet())
{
- featureType = en.next().toString();
-
- if (visible.get(featureType) instanceof GraduatedColor)
- {
- GraduatedColor gc = (GraduatedColor) visible.get(featureType);
- color = (gc.isColourByLabel() ? "label|" : "")
- + Format.getHexString(gc.getMinColor()) + "|"
- + Format.getHexString(gc.getMaxColor())
- + (gc.isAutoScale() ? "|" : "|abso|") + gc.getMin() + "|"
- + gc.getMax() + "|";
- if (gc.getThreshType() != AnnotationColourGradient.NO_THRESHOLD)
- {
- if (gc.getThreshType() == AnnotationColourGradient.BELOW_THRESHOLD)
- {
- color += "below";
- }
- else
- {
- if (gc.getThreshType() != AnnotationColourGradient.ABOVE_THRESHOLD)
- {
- System.err.println("WARNING: Unsupported threshold type ("
- + gc.getThreshType() + ") : Assuming 'above'");
- }
- color += "above";
- }
- // add the value
- color += "|" + gc.getThresh();
- }
- else
- {
- color += "none";
- }
- }
- else if (visible.get(featureType) instanceof Color)
- {
- color = Format.getHexString((Color) visible.get(featureType));
- }
- else
- {
- // legacy support for integer objects containing colour triplet values
- color = Format.getHexString(new Color(Integer.parseInt(visible
- .get(featureType).toString())));
- }
- out.append(featureType);
- out.append(TAB);
- out.append(color);
- out.append(newline);
+ FeatureColourI colour = featureColour.getValue();
+ out.append(colour.toJalviewFormat(featureColour.getKey())).append(
+ newline);
}
}
+
// Work out which groups are both present and visible
- List<String> groups = new ArrayList<String>();
- int groupIndex = 0;
- boolean isnonpos = false;
+ Set<String> groups = new HashSet<String>();
+ String[] types = visible == null ? null : visible.keySet().toArray(
+ new String[visible.keySet().size()]);
- SequenceFeature[] features;
for (int i = 0; i < sequences.length; i++)
{
- features = sequences[i].getSequenceFeatures();
- if (features != null)
+ groups.addAll(sequences[i].getFeatures()
+ .getFeatureGroups(true, types));
+ if (nonpos)
{
- for (int j = 0; j < features.length; j++)
- {
- isnonpos = features[j].begin == 0 && features[j].end == 0;
- if ((!nonpos && isnonpos)
- || (!isnonpos && visOnly && !visible
- .containsKey(features[j].type)))
- {
- continue;
- }
-
- if (features[j].featureGroup != null
- && !groups.contains(features[j].featureGroup))
- {
- groups.add(features[j].featureGroup);
- }
- }
+ groups.addAll(sequences[i].getFeatures().getFeatureGroups(false,
+ types));
}
}
- String group = null;
- do
+ /*
+ * sort distinct groups so null group is output last
+ */
+ List<String> sortedGroups = new ArrayList<String>(groups);
+ Collections.sort(sortedGroups, SORT_NULL_LAST);
+
+ // TODO check where null group should be output
+ boolean foundSome = false;
+ for (String group : sortedGroups)
{
- if (groups.size() > 0 && groupIndex < groups.size())
+ if (group != null)
{
- group = groups.get(groupIndex);
out.append(newline);
out.append("STARTGROUP").append(TAB);
out.append(group);
out.append(newline);
}
- else
- {
- group = null;
- }
+ /*
+ * output features within groups (non-positional first if wanted)
+ */
for (int i = 0; i < sequences.length; i++)
{
- features = sequences[i].getSequenceFeatures();
- if (features != null)
+ List<SequenceFeature> features = new ArrayList<SequenceFeature>();
+ if (nonpos)
{
- for (int j = 0; j < features.length; j++)
- {
- isnonpos = features[j].begin == 0 && features[j].end == 0;
- if ((!nonpos && isnonpos)
- || (!isnonpos && visOnly && !visible
- .containsKey(features[j].type)))
- {
- // skip if feature is nonpos and we ignore them or if we only
- // output visible and it isn't non-pos and it's not visible
- continue;
- }
+ features.addAll(sequences[i].getFeatures().getFeaturesForGroup(
+ false, group, types));
+ }
+ features.addAll(sequences[i].getFeatures().getFeaturesForGroup(
+ true, group, types));
- if (group != null
- && (features[j].featureGroup == null || !features[j].featureGroup
- .equals(group)))
+ for (SequenceFeature sequenceFeature : features)
+ {
+ // we have features to output
+ foundSome = true;
+ if (sequenceFeature.description == null
+ || sequenceFeature.description.equals(""))
+ {
+ out.append(sequenceFeature.type).append(TAB);
+ }
+ else
+ {
+ if (sequenceFeature.links != null
+ && sequenceFeature.getDescription().indexOf("<html>") == -1)
{
- continue;
+ out.append("<html>");
}
- if (group == null && features[j].featureGroup != null)
+ out.append(sequenceFeature.description);
+ if (sequenceFeature.links != null)
{
- continue;
- }
- // we have features to output
- featuresGen = true;
- if (features[j].description == null
- || features[j].description.equals(""))
- {
- out.append(features[j].type).append(TAB);
- }
- else
- {
- if (features[j].links != null
- && features[j].getDescription().indexOf("<html>") == -1)
+ for (int l = 0; l < sequenceFeature.links.size(); l++)
{
- out.append("<html>");
- }
+ String label = sequenceFeature.links.elementAt(l);
+ String href = label.substring(label.indexOf("|") + 1);
+ label = label.substring(0, label.indexOf("|"));
- out.append(features[j].description + " ");
- if (features[j].links != null)
- {
- for (int l = 0; l < features[j].links.size(); l++)
+ if (sequenceFeature.description.indexOf(href) == -1)
{
- String label = features[j].links.elementAt(l).toString();
- String href = label.substring(label.indexOf("|") + 1);
- label = label.substring(0, label.indexOf("|"));
-
- if (features[j].description.indexOf(href) == -1)
- {
- out.append("<a href=\"" + href + "\">" + label + "</a>");
- }
- }
-
- if (features[j].getDescription().indexOf("</html>") == -1)
- {
- out.append("</html>");
+ out.append(" <a href=\"" + href + "\">" + label + "</a>");
}
}
- out.append(TAB);
+ if (sequenceFeature.getDescription().indexOf("</html>") == -1)
+ {
+ out.append("</html>");
+ }
}
- out.append(sequences[i].getName());
- out.append("\t-1\t");
- out.append(features[j].begin);
+
out.append(TAB);
- out.append(features[j].end);
+ }
+ out.append(sequences[i].getName());
+ out.append("\t-1\t");
+ out.append(sequenceFeature.begin);
+ out.append(TAB);
+ out.append(sequenceFeature.end);
+ out.append(TAB);
+ out.append(sequenceFeature.type);
+ if (!Float.isNaN(sequenceFeature.score))
+ {
out.append(TAB);
- out.append(features[j].type);
- if (!Float.isNaN(features[j].score))
- {
- out.append(TAB);
- out.append(features[j].score);
- }
- out.append(newline);
+ out.append(sequenceFeature.score);
}
+ out.append(newline);
}
}
out.append("ENDGROUP").append(TAB);
out.append(group);
out.append(newline);
- groupIndex++;
- }
- else
- {
- break;
}
-
- } while (groupIndex < groups.size() + 1);
-
- if (!featuresGen)
- {
- return "No Features Visible";
}
- return out.toString();
+ return foundSome ? out.toString() : "No Features Visible";
}
/**
dataset = new Alignment(new SequenceI[] {});
}
- boolean parseResult = parse(dataset, null, false, true);
+ Map<String, FeatureColourI> featureColours = new HashMap<String, FeatureColourI>();
+ boolean parseResult = parse(dataset, featureColours, false, true);
if (!parseResult)
{
// pass error up somehow
* @return error message
*/
@Override
- public String print()
+ public String print(SequenceI[] sqs, boolean jvsuffix)
{
- return "Use printGffFormat() or printJalviewFormat()";
+ System.out.println("Use printGffFormat() or printJalviewFormat()");
+ return null;
}
/**
* @return
*/
public String printGffFormat(SequenceI[] sequences,
- Map<String, Object> visible)
+ Map<String, FeatureColourI> visible)
{
return printGffFormat(sequences, visible, true, true);
}
* @return
*/
public String printGffFormat(SequenceI[] sequences,
- Map<String, Object> visible, boolean outputVisibleOnly,
+ Map<String, FeatureColourI> visible, boolean outputVisibleOnly,
boolean includeNonPositionalFeatures)
{
StringBuilder out = new StringBuilder(256);
- out.append(String.format("%s %d\n", GFF_VERSION, gffVersion));
+ int version = gffVersion == 0 ? 2 : gffVersion;
+ out.append(String.format("%s %d\n", GFF_VERSION, version));
String source;
boolean isnonpos;
for (SequenceI seq : sequences)
out.append(strand == 1 ? "+" : (strand == -1 ? "-" : "."));
out.append(TAB);
- out.append(sf.getValue(FRAME, "."));
+ String phase = sf.getPhase();
+ out.append(phase == null ? "." : phase);
// miscellaneous key-values (GFF column 9)
String attributes = sf.getAttributes();
* replace parsed sequence with the realised forward reference
*/
includedseqs.set(p, dummyseq);
+
+ /*
+ * and remove from the newseqs list
+ */
+ newseqs.remove(dummyseq);
}
}
git://source.jalview.org / jalview.git / blobdiff