2 * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3 * Copyright (C) $$Year-Rel$$ The Jalview Authors
5 * This file is part of Jalview.
7 * Jalview is free software: you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation, either version 3
10 * of the License, or (at your option) any later version.
12 * Jalview is distributed in the hope that it will be useful, but
13 * WITHOUT ANY WARRANTY; without even the implied warranty
14 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
15 * PURPOSE. See the GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
19 * The Jalview Authors are detailed in the 'AUTHORS' file.
23 import java.awt.Color;
24 import java.io.IOException;
25 import java.util.ArrayList;
26 import java.util.Arrays;
27 import java.util.Collections;
28 import java.util.HashMap;
29 import java.util.LinkedHashMap;
30 import java.util.List;
31 import java.util.Locale;
33 import java.util.Map.Entry;
34 import java.util.TreeMap;
36 import jalview.analysis.AlignmentUtils;
37 import jalview.analysis.SequenceIdMatcher;
38 import jalview.api.AlignViewportI;
39 import jalview.api.FeatureColourI;
40 import jalview.api.FeatureRenderer;
41 import jalview.api.FeaturesSourceI;
42 import jalview.datamodel.AlignedCodonFrame;
43 import jalview.datamodel.Alignment;
44 import jalview.datamodel.AlignmentI;
45 import jalview.datamodel.MappedFeatures;
46 import jalview.datamodel.SequenceDummy;
47 import jalview.datamodel.SequenceFeature;
48 import jalview.datamodel.SequenceI;
49 import jalview.datamodel.features.FeatureMatcherSet;
50 import jalview.datamodel.features.FeatureMatcherSetI;
51 import jalview.gui.Desktop;
52 import jalview.io.gff.GffHelperFactory;
53 import jalview.io.gff.GffHelperI;
54 import jalview.schemes.FeatureColour;
55 import jalview.util.ColorUtils;
56 import jalview.util.MapList;
57 import jalview.util.ParseHtmlBodyAndLinks;
58 import jalview.util.StringUtils;
61 * Parses and writes features files, which may be in Jalview, GFF2 or GFF3
62 * format. These are tab-delimited formats but with differences in the use of
65 * A Jalview feature file may define feature colours and then declare that the
66 * remainder of the file is in GFF format with the line 'GFF'.
68 * GFF3 files may include alignment mappings for features, which Jalview will
69 * attempt to model, and may include sequence data following a ##FASTA line.
76 public class FeaturesFile extends AlignFile implements FeaturesSourceI
78 private static final String EQUALS = "=";
80 private static final String TAB_REGEX = "\\t";
82 private static final String STARTGROUP = "STARTGROUP";
84 private static final String ENDGROUP = "ENDGROUP";
86 private static final String STARTFILTERS = "STARTFILTERS";
88 private static final String ENDFILTERS = "ENDFILTERS";
90 private static final String ID_NOT_SPECIFIED = "ID_NOT_SPECIFIED";
92 protected static final String GFF_VERSION = "##gff-version";
94 private AlignmentI lastmatchedAl = null;
96 private SequenceIdMatcher matcher = null;
98 protected AlignmentI dataset;
100 protected int gffVersion;
103 * Creates a new FeaturesFile object.
105 public FeaturesFile()
110 * Constructor which does not parse the file immediately
113 * File or String filename
115 * @throws IOException
117 public FeaturesFile(Object file, DataSourceType paste) throws IOException
119 super(false, file, paste);
124 * @throws IOException
126 public FeaturesFile(FileParse source) throws IOException
132 * Constructor that optionally parses the file immediately
134 * @param parseImmediately
137 * @throws IOException
139 public FeaturesFile(boolean parseImmediately, Object file,
140 DataSourceType type) throws IOException
142 super(parseImmediately, file, type);
146 * Parse GFF or sequence features file using case-independent matching,
150 * - alignment/dataset containing sequences that are to be annotated
152 * - hashtable to store feature colour definitions
154 * - process html strings into plain text
155 * @return true if features were added
157 public boolean parse(AlignmentI align,
158 Map<String, FeatureColourI> colours, boolean removeHTML)
160 return parse(align, colours, removeHTML, false);
164 * Extends the default addProperties by also adding peptide-to-cDNA mappings
165 * (if any) derived while parsing a GFF file
168 public void addProperties(AlignmentI al)
170 super.addProperties(al);
171 if (dataset != null && dataset.getCodonFrames() != null)
173 AlignmentI ds = (al.getDataset() == null) ? al : al.getDataset();
174 for (AlignedCodonFrame codons : dataset.getCodonFrames())
176 ds.addCodonFrame(codons);
182 * Parse GFF or Jalview format sequence features file
185 * - alignment/dataset containing sequences that are to be annotated
187 * - map to store feature colour definitions
189 * - process html strings into plain text
190 * @param relaxedIdmatching
191 * - when true, ID matches to compound sequence IDs are allowed
192 * @return true if features were added
194 public boolean parse(AlignmentI align,
195 Map<String, FeatureColourI> colours, boolean removeHTML,
196 boolean relaxedIdmatching)
198 return parse(align, colours, null, removeHTML, relaxedIdmatching);
202 * Parse GFF or Jalview format sequence features file
205 * - alignment/dataset containing sequences that are to be annotated
207 * - map to store feature colour definitions
209 * - map to store feature filter definitions
211 * - process html strings into plain text
212 * @param relaxedIdmatching
213 * - when true, ID matches to compound sequence IDs are allowed
214 * @return true if features were added
216 public boolean parse(AlignmentI align,
217 Map<String, FeatureColourI> colours,
218 Map<String, FeatureMatcherSetI> filters, boolean removeHTML,
219 boolean relaxedIdmatching)
221 Map<String, String> gffProps = new HashMap<>();
223 * keep track of any sequences we try to create from the data
225 List<SequenceI> newseqs = new ArrayList<>();
231 String featureGroup = null;
233 while ((line = nextLine()) != null)
235 // skip comments/process pragmas
236 if (line.length() == 0 || line.startsWith("#"))
238 if (line.toLowerCase(Locale.ROOT).startsWith("##"))
240 processGffPragma(line, gffProps, align, newseqs);
245 gffColumns = line.split(TAB_REGEX);
246 if (gffColumns.length == 1)
248 if (line.trim().equalsIgnoreCase("GFF"))
251 * Jalview features file with appended GFF
252 * assume GFF2 (though it may declare ##gff-version 3)
259 if (gffColumns.length > 0 && gffColumns.length < 4)
262 * if 2 or 3 tokens, we anticipate either 'startgroup', 'endgroup' or
263 * a feature type colour specification
265 String ft = gffColumns[0];
266 if (ft.equalsIgnoreCase(STARTFILTERS))
268 parseFilters(filters);
271 if (ft.equalsIgnoreCase(STARTGROUP))
273 featureGroup = gffColumns[1];
275 else if (ft.equalsIgnoreCase(ENDGROUP))
277 // We should check whether this is the current group,
278 // but at present there's no way of showing more than 1 group
283 String colscheme = gffColumns[1];
284 FeatureColourI colour = FeatureColour
285 .parseJalviewFeatureColour(colscheme);
288 colours.put(ft, colour);
295 * if not a comment, GFF pragma, startgroup, endgroup or feature
296 * colour specification, that just leaves a feature details line
297 * in either Jalview or GFF format
301 parseJalviewFeature(line, gffColumns, align, colours, removeHTML,
302 relaxedIdmatching, featureGroup);
306 parseGff(gffColumns, align, relaxedIdmatching, newseqs);
310 } catch (Exception ex)
312 // should report somewhere useful for UI if necessary
313 warningMessage = ((warningMessage == null) ? "" : warningMessage)
314 + "Parsing error at\n" + line;
315 jalview.bin.Console.outPrintln(
316 "Error parsing feature file: " + ex + "\n" + line);
317 ex.printStackTrace(System.err);
323 * experimental - add any dummy sequences with features to the alignment
324 * - we need them for Ensembl feature extraction - though maybe not otherwise
326 for (SequenceI newseq : newseqs)
328 if (newseq.getFeatures().hasFeatures())
330 align.addSequence(newseq);
337 * Reads input lines from STARTFILTERS to ENDFILTERS and adds a feature type
338 * filter to the map for each line parsed. After exit from this method,
339 * nextLine() should return the line after ENDFILTERS (or we are already at
340 * end of file if ENDFILTERS was missing).
343 * @throws IOException
345 protected void parseFilters(Map<String, FeatureMatcherSetI> filters)
349 while ((line = nextLine()) != null)
351 if (line.toUpperCase(Locale.ROOT).startsWith(ENDFILTERS))
355 String[] tokens = line.split(TAB_REGEX);
356 if (tokens.length != 2)
358 jalview.bin.Console.errPrintln(String.format(
359 "Invalid token count %d for %d", tokens.length, line));
363 String featureType = tokens[0];
364 FeatureMatcherSetI fm = FeatureMatcherSet.fromString(tokens[1]);
365 if (fm != null && filters != null)
367 filters.put(featureType, fm);
374 * Try to parse a Jalview format feature specification and add it as a
375 * sequence feature to any matching sequences in the alignment. Returns true
376 * if successful (a feature was added), or false if not.
381 * @param featureColours
383 * @param relaxedIdmatching
384 * @param featureGroup
386 protected boolean parseJalviewFeature(String line, String[] gffColumns,
387 AlignmentI alignment, Map<String, FeatureColourI> featureColours,
388 boolean removeHTML, boolean relaxedIdMatching,
392 * tokens: description seqid seqIndex start end type [score]
394 if (gffColumns.length < 6)
396 jalview.bin.Console.errPrintln("Ignoring feature line '" + line
397 + "' with too few columns (" + gffColumns.length + ")");
400 String desc = gffColumns[0];
401 String seqId = gffColumns[1];
402 SequenceI seq = findSequence(seqId, alignment, null, relaxedIdMatching);
404 if (!ID_NOT_SPECIFIED.equals(seqId))
406 seq = findSequence(seqId, alignment, null, relaxedIdMatching);
412 String seqIndex = gffColumns[2];
415 int idx = Integer.parseInt(seqIndex);
416 seq = alignment.getSequenceAt(idx);
417 } catch (NumberFormatException ex)
420 .errPrintln("Invalid sequence index: " + seqIndex);
426 jalview.bin.Console.outPrintln("Sequence not found: " + line);
430 int startPos = Integer.parseInt(gffColumns[3]);
431 int endPos = Integer.parseInt(gffColumns[4]);
433 String ft = gffColumns[5];
435 if (!featureColours.containsKey(ft))
438 * Perhaps an old style groups file with no colours -
439 * synthesize a colour from the feature type
441 Color colour = ColorUtils.createColourFromName(ft);
442 featureColours.put(ft, new FeatureColour(colour));
444 SequenceFeature sf = null;
445 if (gffColumns.length > 6)
447 float score = Float.NaN;
450 score = Float.valueOf(gffColumns[6]).floatValue();
451 } catch (NumberFormatException ex)
453 sf = new SequenceFeature(ft, desc, startPos, endPos, featureGroup);
455 sf = new SequenceFeature(ft, desc, startPos, endPos, score,
460 sf = new SequenceFeature(ft, desc, startPos, endPos, featureGroup);
463 parseDescriptionHTML(sf, removeHTML);
465 seq.addSequenceFeature(sf);
468 && (seq = alignment.findName(seq, seqId, false)) != null)
470 seq.addSequenceFeature(new SequenceFeature(sf));
476 * clear any temporary handles used to speed up ID matching
478 protected void resetMatcher()
480 lastmatchedAl = null;
485 * Returns a sequence matching the given id, as follows
487 * <li>strict matching is on exact sequence name</li>
488 * <li>relaxed matching allows matching on a token within the sequence name,
490 * <li>first tries to find a match in the alignment sequences</li>
491 * <li>else tries to find a match in the new sequences already generated while
492 * parsing the features file</li>
493 * <li>else creates a new placeholder sequence, adds it to the new sequences
494 * list, and returns it</li>
500 * @param relaxedIdMatching
504 protected SequenceI findSequence(String seqId, AlignmentI align,
505 List<SequenceI> newseqs, boolean relaxedIdMatching)
507 // TODO encapsulate in SequenceIdMatcher, share the matcher
508 // with the GffHelper (removing code duplication)
509 SequenceI match = null;
510 if (relaxedIdMatching)
512 if (lastmatchedAl != align)
514 lastmatchedAl = align;
515 matcher = new SequenceIdMatcher(align.getSequencesArray());
518 matcher.addAll(newseqs);
521 match = matcher.findIdMatch(seqId);
525 match = align.findName(seqId, true);
526 if (match == null && newseqs != null)
528 for (SequenceI m : newseqs)
530 if (seqId.equals(m.getName()))
538 if (match == null && newseqs != null)
540 match = new SequenceDummy(seqId);
541 if (relaxedIdMatching)
543 matcher.addAll(Arrays.asList(new SequenceI[] { match }));
545 // add dummy sequence to the newseqs list
551 public void parseDescriptionHTML(SequenceFeature sf, boolean removeHTML)
553 if (sf.getDescription() == null)
557 ParseHtmlBodyAndLinks parsed = new ParseHtmlBodyAndLinks(
558 sf.getDescription(), removeHTML, newline);
562 sf.setDescription(parsed.getNonHtmlContent());
565 for (String link : parsed.getLinks())
572 * Returns contents of a Jalview format features file, for visible features,
573 * as filtered by type and group. Features with a null group are displayed if
574 * their feature type is visible. Non-positional features may optionally be
575 * included (with no check on type or group).
579 * @param includeNonPositional
580 * if true, include non-positional features (regardless of group or
582 * @param includeComplement
583 * if true, include visible complementary (CDS/protein) positional
584 * features, with locations converted to local sequence coordinates
587 public String printJalviewFormat(SequenceI[] sequences,
588 FeatureRenderer fr, boolean includeNonPositional,
589 boolean includeComplement)
591 Map<String, FeatureColourI> visibleColours = fr
592 .getDisplayedFeatureCols();
593 Map<String, FeatureMatcherSetI> featureFilters = fr.getFeatureFilters();
596 * write out feature colours (if we know them)
598 // TODO: decide if feature links should also be written here ?
599 StringBuilder out = new StringBuilder(256);
600 if (visibleColours != null)
602 for (Entry<String, FeatureColourI> featureColour : visibleColours
605 FeatureColourI colour = featureColour.getValue();
606 out.append(colour.toJalviewFormat(featureColour.getKey()))
611 String[] types = visibleColours == null ? new String[0]
612 : visibleColours.keySet()
613 .toArray(new String[visibleColours.keySet().size()]);
616 * feature filters if any
618 outputFeatureFilters(out, visibleColours, featureFilters);
621 * output features within groups
623 int count = outputFeaturesByGroup(out, fr, types, sequences,
624 includeNonPositional);
626 if (includeComplement)
628 count += outputComplementFeatures(out, fr, sequences);
631 return count > 0 ? out.toString() : "No Features Visible";
635 * Outputs any visible complementary (CDS/peptide) positional features as
636 * Jalview format, within feature group. The coordinates of the linked
637 * features are converted to the corresponding positions of the local
645 private int outputComplementFeatures(StringBuilder out,
646 FeatureRenderer fr, SequenceI[] sequences)
648 AlignViewportI comp = fr.getViewport().getCodingComplement();
649 FeatureRenderer fr2 = Desktop.getAlignFrameFor(comp)
650 .getFeatureRenderer();
653 * bin features by feature group and sequence
655 Map<String, Map<String, List<SequenceFeature>>> map = new TreeMap<>(
656 String.CASE_INSENSITIVE_ORDER);
659 for (SequenceI seq : sequences)
662 * find complementary features
664 List<SequenceFeature> complementary = findComplementaryFeatures(seq,
666 String seqName = seq.getName();
668 for (SequenceFeature sf : complementary)
670 String group = sf.getFeatureGroup();
671 if (!map.containsKey(group))
673 map.put(group, new LinkedHashMap<>()); // preserves sequence order
675 Map<String, List<SequenceFeature>> groupFeatures = map.get(group);
676 if (!groupFeatures.containsKey(seqName))
678 groupFeatures.put(seqName, new ArrayList<>());
680 List<SequenceFeature> foundFeatures = groupFeatures.get(seqName);
681 foundFeatures.add(sf);
687 * output features by group
689 for (Entry<String, Map<String, List<SequenceFeature>>> groupFeatures : map
693 String group = groupFeatures.getKey();
694 if (!"".equals(group))
696 out.append(STARTGROUP).append(TAB).append(group).append(newline);
698 Map<String, List<SequenceFeature>> seqFeaturesMap = groupFeatures
700 for (Entry<String, List<SequenceFeature>> seqFeatures : seqFeaturesMap
703 String sequenceName = seqFeatures.getKey();
704 for (SequenceFeature sf : seqFeatures.getValue())
706 formatJalviewFeature(out, sequenceName, sf);
709 if (!"".equals(group))
711 out.append(ENDGROUP).append(TAB).append(group).append(newline);
719 * Answers a list of mapped features visible in the (CDS/protein) complement,
720 * with feature positions translated to local sequence coordinates
726 protected List<SequenceFeature> findComplementaryFeatures(SequenceI seq,
730 * avoid duplication of features (e.g. peptide feature
731 * at all 3 mapped codon positions)
733 List<SequenceFeature> found = new ArrayList<>();
734 List<SequenceFeature> complementary = new ArrayList<>();
736 for (int pos = seq.getStart(); pos <= seq.getEnd(); pos++)
738 MappedFeatures mf = fr2.findComplementFeaturesAtResidue(seq, pos);
742 for (SequenceFeature sf : mf.features)
745 * make a virtual feature with local coordinates
747 if (!found.contains(sf))
749 String group = sf.getFeatureGroup();
755 int begin = sf.getBegin();
756 int end = sf.getEnd();
757 int[] range = mf.getMappedPositions(begin, end);
758 SequenceFeature sf2 = new SequenceFeature(sf, range[0],
759 range[1], group, sf.getScore());
760 complementary.add(sf2);
766 return complementary;
770 * Outputs any feature filters defined for visible feature types, sandwiched
771 * by STARTFILTERS and ENDFILTERS lines
775 * @param featureFilters
777 void outputFeatureFilters(StringBuilder out,
778 Map<String, FeatureColourI> visible,
779 Map<String, FeatureMatcherSetI> featureFilters)
781 if (visible == null || featureFilters == null
782 || featureFilters.isEmpty())
787 boolean first = true;
788 for (String featureType : visible.keySet())
790 FeatureMatcherSetI filter = featureFilters.get(featureType);
796 out.append(newline).append(STARTFILTERS).append(newline);
798 out.append(featureType).append(TAB).append(filter.toStableString())
804 out.append(ENDFILTERS).append(newline);
810 * Appends output of visible sequence features within feature groups to the
811 * output buffer. Groups other than the null or empty group are sandwiched by
812 * STARTGROUP and ENDGROUP lines. Answers the number of features written.
816 * @param featureTypes
818 * @param includeNonPositional
821 private int outputFeaturesByGroup(StringBuilder out, FeatureRenderer fr,
822 String[] featureTypes, SequenceI[] sequences,
823 boolean includeNonPositional)
825 List<String> featureGroups = fr.getFeatureGroups();
828 * sort groups alphabetically, and ensure that features with a
829 * null or empty group are output after those in named groups
831 List<String> sortedGroups = new ArrayList<>(featureGroups);
832 sortedGroups.remove(null);
833 sortedGroups.remove("");
834 Collections.sort(sortedGroups);
835 sortedGroups.add(null);
836 sortedGroups.add("");
839 List<String> visibleGroups = fr.getDisplayedFeatureGroups();
842 * loop over all groups (may be visible or not);
843 * non-positional features are output even if group is not visible
845 for (String group : sortedGroups)
847 boolean firstInGroup = true;
848 boolean isNullGroup = group == null || "".equals(group);
850 for (int i = 0; i < sequences.length; i++)
852 String sequenceName = sequences[i].getName();
853 List<SequenceFeature> features = new ArrayList<>();
856 * get any non-positional features in this group, if wanted
857 * (for any feature type, whether visible or not)
859 if (includeNonPositional)
861 features.addAll(sequences[i].getFeatures()
862 .getFeaturesForGroup(false, group));
866 * add positional features for visible feature types, but
867 * (for named groups) only if feature group is visible
869 if (featureTypes.length > 0
870 && (isNullGroup || visibleGroups.contains(group)))
872 features.addAll(sequences[i].getFeatures()
873 .getFeaturesForGroup(true, group, featureTypes));
876 for (SequenceFeature sf : features)
878 if (sf.isNonPositional() || fr.isVisible(sf))
886 out.append(STARTGROUP).append(TAB).append(group)
890 firstInGroup = false;
891 formatJalviewFeature(out, sequenceName, sf);
896 if (!isNullGroup && !firstInGroup)
898 out.append(ENDGROUP).append(TAB).append(group).append(newline);
905 * Formats one feature in Jalview format and appends to the string buffer
908 * @param sequenceName
909 * @param sequenceFeature
911 protected void formatJalviewFeature(StringBuilder out,
912 String sequenceName, SequenceFeature sequenceFeature)
914 if (sequenceFeature.description == null
915 || sequenceFeature.description.equals(""))
917 out.append(sequenceFeature.type).append(TAB);
921 if (sequenceFeature.links != null
922 && sequenceFeature.getDescription().indexOf("<html>") == -1)
924 out.append("<html>");
927 out.append(sequenceFeature.description);
928 if (sequenceFeature.links != null)
930 for (int l = 0; l < sequenceFeature.links.size(); l++)
932 String label = sequenceFeature.links.elementAt(l);
933 String href = label.substring(label.indexOf("|") + 1);
934 label = label.substring(0, label.indexOf("|"));
936 if (sequenceFeature.description.indexOf(href) == -1)
938 out.append(" <a href=\"").append(href).append("\">")
939 .append(label).append("</a>");
943 if (sequenceFeature.getDescription().indexOf("</html>") == -1)
945 out.append("</html>");
951 out.append(sequenceName);
952 out.append("\t-1\t");
953 out.append(sequenceFeature.begin);
955 out.append(sequenceFeature.end);
957 out.append(sequenceFeature.type);
958 if (!Float.isNaN(sequenceFeature.score))
961 out.append(sequenceFeature.score);
967 * Parse method that is called when a GFF file is dragged to the desktop
972 AlignViewportI av = getViewport();
975 if (av.getAlignment() != null)
977 dataset = av.getAlignment().getDataset();
981 // working in the applet context ?
982 dataset = av.getAlignment();
987 dataset = new Alignment(new SequenceI[] {});
990 Map<String, FeatureColourI> featureColours = new HashMap<>();
991 boolean parseResult = parse(dataset, featureColours, false, true);
994 // pass error up somehow
998 // update viewport with the dataset data ?
1002 setSeqs(dataset.getSequencesArray());
1007 * Implementation of unused abstract method
1009 * @return error message
1012 public String print(SequenceI[] sqs, boolean jvsuffix)
1015 .outPrintln("Use printGffFormat() or printJalviewFormat()");
1020 * Returns features output in GFF2 format
1023 * the sequences whose features are to be output
1025 * a map whose keys are the type names of visible features
1026 * @param visibleFeatureGroups
1027 * @param includeNonPositionalFeatures
1028 * @param includeComplement
1031 public String printGffFormat(SequenceI[] sequences, FeatureRenderer fr,
1032 boolean includeNonPositionalFeatures, boolean includeComplement)
1034 FeatureRenderer fr2 = null;
1035 if (includeComplement)
1037 AlignViewportI comp = fr.getViewport().getCodingComplement();
1038 fr2 = Desktop.getAlignFrameFor(comp).getFeatureRenderer();
1041 Map<String, FeatureColourI> visibleColours = fr
1042 .getDisplayedFeatureCols();
1044 StringBuilder out = new StringBuilder(256);
1046 out.append(String.format("%s %d\n", GFF_VERSION,
1047 gffVersion == 0 ? 2 : gffVersion));
1049 String[] types = visibleColours == null ? new String[0]
1050 : visibleColours.keySet()
1051 .toArray(new String[visibleColours.keySet().size()]);
1053 for (SequenceI seq : sequences)
1055 List<SequenceFeature> seqFeatures = new ArrayList<>();
1056 List<SequenceFeature> features = new ArrayList<>();
1057 if (includeNonPositionalFeatures)
1059 features.addAll(seq.getFeatures().getNonPositionalFeatures());
1061 if (visibleColours != null && !visibleColours.isEmpty())
1063 features.addAll(seq.getFeatures().getPositionalFeatures(types));
1065 for (SequenceFeature sf : features)
1067 if (sf.isNonPositional() || fr.isVisible(sf))
1070 * drop features hidden by group visibility, colour threshold,
1071 * or feature filter condition
1073 seqFeatures.add(sf);
1077 if (includeComplement)
1079 seqFeatures.addAll(findComplementaryFeatures(seq, fr2));
1083 * sort features here if wanted
1085 for (SequenceFeature sf : seqFeatures)
1087 formatGffFeature(out, seq, sf);
1088 out.append(newline);
1092 return out.toString();
1096 * Formats one feature as GFF and appends to the string buffer
1098 private void formatGffFeature(StringBuilder out, SequenceI seq,
1101 String source = sf.featureGroup;
1104 source = sf.getDescription();
1107 out.append(seq.getName());
1111 out.append(sf.type);
1113 out.append(sf.begin);
1117 out.append(sf.score);
1120 int strand = sf.getStrand();
1121 out.append(strand == 1 ? "+" : (strand == -1 ? "-" : "."));
1124 String phase = sf.getPhase();
1125 out.append(phase == null ? "." : phase);
1127 if (sf.otherDetails != null && !sf.otherDetails.isEmpty())
1129 Map<String, Object> map = sf.otherDetails;
1130 formatAttributes(out, map);
1135 * A helper method that outputs attributes stored in the map as
1136 * semicolon-delimited values e.g.
1139 * AC_Male=0;AF_NFE=0.00000e 00;Hom_FIN=0;GQ_MEDIAN=9
1142 * A map-valued attribute is formatted as a comma-delimited list within
1143 * braces, for example
1146 * jvmap_CSQ={ALLELE_NUM=1,UNIPARC=UPI0002841053,Feature=ENST00000585561}
1149 * The {@code jvmap_} prefix designates a values map and is removed if the
1150 * value is parsed when read in. (The GFF3 specification allows
1151 * 'semi-structured data' to be represented provided the attribute name begins
1152 * with a lower case letter.)
1156 * @see http://gmod.org/wiki/GFF3#GFF3_Format
1158 void formatAttributes(StringBuilder sb, Map<String, Object> map)
1161 boolean first = true;
1162 for (String key : map.keySet())
1164 if (SequenceFeature.STRAND.equals(key)
1165 || SequenceFeature.PHASE.equals(key))
1168 * values stashed in map but output to their own columns
1179 Object value = map.get(key);
1180 if (value instanceof Map<?, ?>)
1182 formatMapAttribute(sb, key, (Map<?, ?>) value);
1186 String formatted = StringUtils.urlEncode(value.toString(),
1187 GffHelperI.GFF_ENCODABLE);
1188 sb.append(key).append(EQUALS).append(formatted);
1194 * Formats the map entries as
1197 * key=key1=value1,key2=value2,...
1200 * and appends this to the string buffer
1206 private void formatMapAttribute(StringBuilder sb, String key,
1209 if (map == null || map.isEmpty())
1215 * AbstractMap.toString would be a shortcut here, but more reliable
1216 * to code the required format in case toString changes in future
1218 sb.append(key).append(EQUALS);
1219 boolean first = true;
1220 for (Entry<?, ?> entry : map.entrySet())
1227 sb.append(entry.getKey().toString()).append(EQUALS);
1228 String formatted = StringUtils.urlEncode(entry.getValue().toString(),
1229 GffHelperI.GFF_ENCODABLE);
1230 sb.append(formatted);
1235 * Returns a mapping given list of one or more Align descriptors (exonerate
1238 * @param alignedRegions
1239 * a list of "Align fromStart toStart fromCount"
1240 * @param mapIsFromCdna
1241 * if true, 'from' is dna, else 'from' is protein
1243 * either 1 (forward) or -1 (reverse)
1245 * @throws IOException
1247 protected MapList constructCodonMappingFromAlign(
1248 List<String> alignedRegions, boolean mapIsFromCdna, int strand)
1253 throw new IOException(
1254 "Invalid strand for a codon mapping (cannot be 0)");
1256 int regions = alignedRegions.size();
1257 // arrays to hold [start, end] for each aligned region
1258 int[] fromRanges = new int[regions * 2]; // from dna
1259 int[] toRanges = new int[regions * 2]; // to protein
1260 int fromRangesIndex = 0;
1261 int toRangesIndex = 0;
1263 for (String range : alignedRegions)
1266 * Align mapFromStart mapToStart mapFromCount
1267 * e.g. if mapIsFromCdna
1268 * Align 11270 143 120
1270 * 120 bases from pos 11270 align to pos 143 in peptide
1271 * if !mapIsFromCdna this would instead be
1272 * Align 143 11270 40
1274 String[] tokens = range.split(" ");
1275 if (tokens.length != 3)
1277 throw new IOException("Wrong number of fields for Align");
1284 fromStart = Integer.parseInt(tokens[0]);
1285 toStart = Integer.parseInt(tokens[1]);
1286 fromCount = Integer.parseInt(tokens[2]);
1287 } catch (NumberFormatException nfe)
1289 throw new IOException(
1290 "Invalid number in Align field: " + nfe.getMessage());
1294 * Jalview always models from dna to protein, so adjust values if the
1295 * GFF mapping is from protein to dna
1300 int temp = fromStart;
1301 fromStart = toStart;
1304 fromRanges[fromRangesIndex++] = fromStart;
1305 fromRanges[fromRangesIndex++] = fromStart + strand * (fromCount - 1);
1308 * If a codon has an intron gap, there will be contiguous 'toRanges';
1309 * this is handled for us by the MapList constructor.
1310 * (It is not clear that exonerate ever generates this case)
1312 toRanges[toRangesIndex++] = toStart;
1313 toRanges[toRangesIndex++] = toStart + (fromCount - 1) / 3;
1316 return new MapList(fromRanges, toRanges, 3, 1);
1320 * Parse a GFF format feature. This may include creating a 'dummy' sequence to
1321 * hold the feature, or for its mapped sequence, or both, to be resolved
1322 * either later in the GFF file (##FASTA section), or when the user loads
1323 * additional sequences.
1327 * @param relaxedIdMatching
1331 protected SequenceI parseGff(String[] gffColumns, AlignmentI alignment,
1332 boolean relaxedIdMatching, List<SequenceI> newseqs)
1335 * GFF: seqid source type start end score strand phase [attributes]
1337 if (gffColumns.length < 5)
1340 .errPrintln("Ignoring GFF feature line with too few columns ("
1341 + gffColumns.length + ")");
1346 * locate referenced sequence in alignment _or_
1347 * as a forward or external reference (SequenceDummy)
1349 String seqId = gffColumns[0];
1350 SequenceI seq = findSequence(seqId, alignment, newseqs,
1353 SequenceFeature sf = null;
1354 GffHelperI helper = GffHelperFactory.getHelper(gffColumns);
1359 sf = helper.processGff(seq, gffColumns, alignment, newseqs,
1363 seq.addSequenceFeature(sf);
1364 while ((seq = alignment.findName(seq, seqId, true)) != null)
1366 seq.addSequenceFeature(new SequenceFeature(sf));
1369 } catch (IOException e)
1372 .errPrintln("GFF parsing failed with: " + e.getMessage());
1381 * After encountering ##fasta in a GFF3 file, process the remainder of the
1382 * file as FAST sequence data. Any placeholder sequences created during
1383 * feature parsing are updated with the actual sequences.
1387 * @throws IOException
1389 protected void processAsFasta(AlignmentI align, List<SequenceI> newseqs)
1395 } catch (IOException q)
1398 // Opening a FastaFile object with the remainder of this object's dataIn.
1399 // Tell the constructor to NOT close the dataIn when finished.
1400 FastaFile parser = new FastaFile(this, false);
1401 List<SequenceI> includedseqs = parser.getSeqs();
1403 SequenceIdMatcher smatcher = new SequenceIdMatcher(newseqs);
1406 * iterate over includedseqs, and replacing matching ones with newseqs
1407 * sequences. Generic iterator not used here because we modify
1408 * includedseqs as we go
1410 for (int p = 0, pSize = includedseqs.size(); p < pSize; p++)
1412 // search for any dummy seqs that this sequence can be used to update
1413 SequenceI includedSeq = includedseqs.get(p);
1414 SequenceI dummyseq = smatcher.findIdMatch(includedSeq);
1415 if (dummyseq != null && dummyseq instanceof SequenceDummy)
1417 // probably have the pattern wrong
1418 // idea is that a flyweight proxy for a sequence ID can be created for
1419 // 1. stable reference creation
1420 // 2. addition of annotation
1421 // 3. future replacement by a real sequence
1422 // current pattern is to create SequenceDummy objects - a convenience
1423 // constructor for a Sequence.
1424 // problem is that when promoted to a real sequence, all references
1425 // need to be updated somehow. We avoid that by keeping the same object.
1426 ((SequenceDummy) dummyseq).become(includedSeq);
1427 dummyseq.createDatasetSequence();
1430 * Update mappings so they are now to the dataset sequence
1432 for (AlignedCodonFrame mapping : align.getCodonFrames())
1434 mapping.updateToDataset(dummyseq);
1438 * replace parsed sequence with the realised forward reference
1440 includedseqs.set(p, dummyseq);
1443 * and remove from the newseqs list
1445 newseqs.remove(dummyseq);
1450 * finally add sequences to the dataset
1452 for (SequenceI seq : includedseqs)
1454 // experimental: mapping-based 'alignment' to query sequence
1455 AlignmentUtils.alignSequenceAs(seq, align,
1456 String.valueOf(align.getGapCharacter()), false, true);
1458 // rename sequences if GFF handler requested this
1459 // TODO a more elegant way e.g. gffHelper.postProcess(newseqs) ?
1460 List<SequenceFeature> sfs = seq.getFeatures().getPositionalFeatures();
1463 String newName = (String) sfs.get(0)
1464 .getValue(GffHelperI.RENAME_TOKEN);
1465 if (newName != null)
1467 seq.setName(newName);
1470 align.addSequence(seq);
1475 * Process a ## directive
1481 * @throws IOException
1483 protected void processGffPragma(String line, Map<String, String> gffProps,
1484 AlignmentI align, List<SequenceI> newseqs) throws IOException
1487 if ("###".equals(line))
1489 // close off any open 'forward references'
1493 String[] tokens = line.substring(2).split(" ");
1494 String pragma = tokens[0];
1495 String value = tokens.length == 1 ? null : tokens[1];
1497 if ("gff-version".equalsIgnoreCase(pragma))
1503 // value may be e.g. "3.1.2"
1504 gffVersion = Integer.parseInt(value.split("\\.")[0]);
1505 } catch (NumberFormatException e)
1511 else if ("sequence-region".equalsIgnoreCase(pragma))
1513 // could capture <seqid start end> if wanted here
1515 else if ("feature-ontology".equalsIgnoreCase(pragma))
1517 // should resolve against the specified feature ontology URI
1519 else if ("attribute-ontology".equalsIgnoreCase(pragma))
1521 // URI of attribute ontology - not currently used in GFF3
1523 else if ("source-ontology".equalsIgnoreCase(pragma))
1525 // URI of source ontology - not currently used in GFF3
1527 else if ("species-build".equalsIgnoreCase(pragma))
1529 // save URI of specific NCBI taxon version of annotations
1530 gffProps.put("species-build", value);
1532 else if ("fasta".equalsIgnoreCase(pragma))
1534 // process the rest of the file as a fasta file and replace any dummy
1536 processAsFasta(align, newseqs);
1540 jalview.bin.Console.errPrintln("Ignoring unknown pragma: " + line);