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("Error parsing feature file: " + ex + "\n" + line);
316 ex.printStackTrace(System.err);
322 * experimental - add any dummy sequences with features to the alignment
323 * - we need them for Ensembl feature extraction - though maybe not otherwise
325 for (SequenceI newseq : newseqs)
327 if (newseq.getFeatures().hasFeatures())
329 align.addSequence(newseq);
336 * Reads input lines from STARTFILTERS to ENDFILTERS and adds a feature type
337 * filter to the map for each line parsed. After exit from this method,
338 * nextLine() should return the line after ENDFILTERS (or we are already at
339 * end of file if ENDFILTERS was missing).
342 * @throws IOException
344 protected void parseFilters(Map<String, FeatureMatcherSetI> filters)
348 while ((line = nextLine()) != null)
350 if (line.toUpperCase(Locale.ROOT).startsWith(ENDFILTERS))
354 String[] tokens = line.split(TAB_REGEX);
355 if (tokens.length != 2)
357 jalview.bin.Console.errPrintln(String.format("Invalid token count %d for %d",
358 tokens.length, line));
362 String featureType = tokens[0];
363 FeatureMatcherSetI fm = FeatureMatcherSet.fromString(tokens[1]);
364 if (fm != null && filters != null)
366 filters.put(featureType, fm);
373 * Try to parse a Jalview format feature specification and add it as a
374 * sequence feature to any matching sequences in the alignment. Returns true
375 * if successful (a feature was added), or false if not.
380 * @param featureColours
382 * @param relaxedIdmatching
383 * @param featureGroup
385 protected boolean parseJalviewFeature(String line, String[] gffColumns,
386 AlignmentI alignment, Map<String, FeatureColourI> featureColours,
387 boolean removeHTML, boolean relaxedIdMatching,
391 * tokens: description seqid seqIndex start end type [score]
393 if (gffColumns.length < 6)
395 jalview.bin.Console.errPrintln("Ignoring feature line '" + line
396 + "' with too few columns (" + gffColumns.length + ")");
399 String desc = gffColumns[0];
400 String seqId = gffColumns[1];
401 SequenceI seq = findSequence(seqId, alignment, null, relaxedIdMatching);
403 if (!ID_NOT_SPECIFIED.equals(seqId))
405 seq = findSequence(seqId, alignment, null, relaxedIdMatching);
411 String seqIndex = gffColumns[2];
414 int idx = Integer.parseInt(seqIndex);
415 seq = alignment.getSequenceAt(idx);
416 } catch (NumberFormatException ex)
418 jalview.bin.Console.errPrintln("Invalid sequence index: " + seqIndex);
424 jalview.bin.Console.outPrintln("Sequence not found: " + line);
428 int startPos = Integer.parseInt(gffColumns[3]);
429 int endPos = Integer.parseInt(gffColumns[4]);
431 String ft = gffColumns[5];
433 if (!featureColours.containsKey(ft))
436 * Perhaps an old style groups file with no colours -
437 * synthesize a colour from the feature type
439 Color colour = ColorUtils.createColourFromName(ft);
440 featureColours.put(ft, new FeatureColour(colour));
442 SequenceFeature sf = null;
443 if (gffColumns.length > 6)
445 float score = Float.NaN;
448 score = Float.valueOf(gffColumns[6]).floatValue();
449 } catch (NumberFormatException ex)
451 sf = new SequenceFeature(ft, desc, startPos, endPos, featureGroup);
453 sf = new SequenceFeature(ft, desc, startPos, endPos, score,
458 sf = new SequenceFeature(ft, desc, startPos, endPos, featureGroup);
461 parseDescriptionHTML(sf, removeHTML);
463 seq.addSequenceFeature(sf);
466 && (seq = alignment.findName(seq, seqId, false)) != null)
468 seq.addSequenceFeature(new SequenceFeature(sf));
474 * clear any temporary handles used to speed up ID matching
476 protected void resetMatcher()
478 lastmatchedAl = null;
483 * Returns a sequence matching the given id, as follows
485 * <li>strict matching is on exact sequence name</li>
486 * <li>relaxed matching allows matching on a token within the sequence name,
488 * <li>first tries to find a match in the alignment sequences</li>
489 * <li>else tries to find a match in the new sequences already generated while
490 * parsing the features file</li>
491 * <li>else creates a new placeholder sequence, adds it to the new sequences
492 * list, and returns it</li>
498 * @param relaxedIdMatching
502 protected SequenceI findSequence(String seqId, AlignmentI align,
503 List<SequenceI> newseqs, boolean relaxedIdMatching)
505 // TODO encapsulate in SequenceIdMatcher, share the matcher
506 // with the GffHelper (removing code duplication)
507 SequenceI match = null;
508 if (relaxedIdMatching)
510 if (lastmatchedAl != align)
512 lastmatchedAl = align;
513 matcher = new SequenceIdMatcher(align.getSequencesArray());
516 matcher.addAll(newseqs);
519 match = matcher.findIdMatch(seqId);
523 match = align.findName(seqId, true);
524 if (match == null && newseqs != null)
526 for (SequenceI m : newseqs)
528 if (seqId.equals(m.getName()))
536 if (match == null && newseqs != null)
538 match = new SequenceDummy(seqId);
539 if (relaxedIdMatching)
541 matcher.addAll(Arrays.asList(new SequenceI[] { match }));
543 // add dummy sequence to the newseqs list
549 public void parseDescriptionHTML(SequenceFeature sf, boolean removeHTML)
551 if (sf.getDescription() == null)
555 ParseHtmlBodyAndLinks parsed = new ParseHtmlBodyAndLinks(
556 sf.getDescription(), removeHTML, newline);
560 sf.setDescription(parsed.getNonHtmlContent());
563 for (String link : parsed.getLinks())
570 * Returns contents of a Jalview format features file, for visible features,
571 * as filtered by type and group. Features with a null group are displayed if
572 * their feature type is visible. Non-positional features may optionally be
573 * included (with no check on type or group).
577 * @param includeNonPositional
578 * if true, include non-positional features (regardless of group or
580 * @param includeComplement
581 * if true, include visible complementary (CDS/protein) positional
582 * features, with locations converted to local sequence coordinates
585 public String printJalviewFormat(SequenceI[] sequences,
586 FeatureRenderer fr, boolean includeNonPositional,
587 boolean includeComplement)
589 Map<String, FeatureColourI> visibleColours = fr
590 .getDisplayedFeatureCols();
591 Map<String, FeatureMatcherSetI> featureFilters = fr.getFeatureFilters();
594 * write out feature colours (if we know them)
596 // TODO: decide if feature links should also be written here ?
597 StringBuilder out = new StringBuilder(256);
598 if (visibleColours != null)
600 for (Entry<String, FeatureColourI> featureColour : visibleColours
603 FeatureColourI colour = featureColour.getValue();
604 out.append(colour.toJalviewFormat(featureColour.getKey()))
609 String[] types = visibleColours == null ? new String[0]
610 : visibleColours.keySet()
611 .toArray(new String[visibleColours.keySet().size()]);
614 * feature filters if any
616 outputFeatureFilters(out, visibleColours, featureFilters);
619 * output features within groups
621 int count = outputFeaturesByGroup(out, fr, types, sequences,
622 includeNonPositional);
624 if (includeComplement)
626 count += outputComplementFeatures(out, fr, sequences);
629 return count > 0 ? out.toString() : "No Features Visible";
633 * Outputs any visible complementary (CDS/peptide) positional features as
634 * Jalview format, within feature group. The coordinates of the linked
635 * features are converted to the corresponding positions of the local
643 private int outputComplementFeatures(StringBuilder out,
644 FeatureRenderer fr, SequenceI[] sequences)
646 AlignViewportI comp = fr.getViewport().getCodingComplement();
647 FeatureRenderer fr2 = Desktop.getAlignFrameFor(comp)
648 .getFeatureRenderer();
651 * bin features by feature group and sequence
653 Map<String, Map<String, List<SequenceFeature>>> map = new TreeMap<>(
654 String.CASE_INSENSITIVE_ORDER);
657 for (SequenceI seq : sequences)
660 * find complementary features
662 List<SequenceFeature> complementary = findComplementaryFeatures(seq,
664 String seqName = seq.getName();
666 for (SequenceFeature sf : complementary)
668 String group = sf.getFeatureGroup();
669 if (!map.containsKey(group))
671 map.put(group, new LinkedHashMap<>()); // preserves sequence order
673 Map<String, List<SequenceFeature>> groupFeatures = map.get(group);
674 if (!groupFeatures.containsKey(seqName))
676 groupFeatures.put(seqName, new ArrayList<>());
678 List<SequenceFeature> foundFeatures = groupFeatures.get(seqName);
679 foundFeatures.add(sf);
685 * output features by group
687 for (Entry<String, Map<String, List<SequenceFeature>>> groupFeatures : map
691 String group = groupFeatures.getKey();
692 if (!"".equals(group))
694 out.append(STARTGROUP).append(TAB).append(group).append(newline);
696 Map<String, List<SequenceFeature>> seqFeaturesMap = groupFeatures
698 for (Entry<String, List<SequenceFeature>> seqFeatures : seqFeaturesMap
701 String sequenceName = seqFeatures.getKey();
702 for (SequenceFeature sf : seqFeatures.getValue())
704 formatJalviewFeature(out, sequenceName, sf);
707 if (!"".equals(group))
709 out.append(ENDGROUP).append(TAB).append(group).append(newline);
717 * Answers a list of mapped features visible in the (CDS/protein) complement,
718 * with feature positions translated to local sequence coordinates
724 protected List<SequenceFeature> findComplementaryFeatures(SequenceI seq,
728 * avoid duplication of features (e.g. peptide feature
729 * at all 3 mapped codon positions)
731 List<SequenceFeature> found = new ArrayList<>();
732 List<SequenceFeature> complementary = new ArrayList<>();
734 for (int pos = seq.getStart(); pos <= seq.getEnd(); pos++)
736 MappedFeatures mf = fr2.findComplementFeaturesAtResidue(seq, pos);
740 for (SequenceFeature sf : mf.features)
743 * make a virtual feature with local coordinates
745 if (!found.contains(sf))
747 String group = sf.getFeatureGroup();
753 int begin = sf.getBegin();
754 int end = sf.getEnd();
755 int[] range = mf.getMappedPositions(begin, end);
756 SequenceFeature sf2 = new SequenceFeature(sf, range[0],
757 range[1], group, sf.getScore());
758 complementary.add(sf2);
764 return complementary;
768 * Outputs any feature filters defined for visible feature types, sandwiched
769 * by STARTFILTERS and ENDFILTERS lines
773 * @param featureFilters
775 void outputFeatureFilters(StringBuilder out,
776 Map<String, FeatureColourI> visible,
777 Map<String, FeatureMatcherSetI> featureFilters)
779 if (visible == null || featureFilters == null
780 || featureFilters.isEmpty())
785 boolean first = true;
786 for (String featureType : visible.keySet())
788 FeatureMatcherSetI filter = featureFilters.get(featureType);
794 out.append(newline).append(STARTFILTERS).append(newline);
796 out.append(featureType).append(TAB).append(filter.toStableString())
802 out.append(ENDFILTERS).append(newline);
808 * Appends output of visible sequence features within feature groups to the
809 * output buffer. Groups other than the null or empty group are sandwiched by
810 * STARTGROUP and ENDGROUP lines. Answers the number of features written.
814 * @param featureTypes
816 * @param includeNonPositional
819 private int outputFeaturesByGroup(StringBuilder out, FeatureRenderer fr,
820 String[] featureTypes, SequenceI[] sequences,
821 boolean includeNonPositional)
823 List<String> featureGroups = fr.getFeatureGroups();
826 * sort groups alphabetically, and ensure that features with a
827 * null or empty group are output after those in named groups
829 List<String> sortedGroups = new ArrayList<>(featureGroups);
830 sortedGroups.remove(null);
831 sortedGroups.remove("");
832 Collections.sort(sortedGroups);
833 sortedGroups.add(null);
834 sortedGroups.add("");
837 List<String> visibleGroups = fr.getDisplayedFeatureGroups();
840 * loop over all groups (may be visible or not);
841 * non-positional features are output even if group is not visible
843 for (String group : sortedGroups)
845 boolean firstInGroup = true;
846 boolean isNullGroup = group == null || "".equals(group);
848 for (int i = 0; i < sequences.length; i++)
850 String sequenceName = sequences[i].getName();
851 List<SequenceFeature> features = new ArrayList<>();
854 * get any non-positional features in this group, if wanted
855 * (for any feature type, whether visible or not)
857 if (includeNonPositional)
859 features.addAll(sequences[i].getFeatures()
860 .getFeaturesForGroup(false, group));
864 * add positional features for visible feature types, but
865 * (for named groups) only if feature group is visible
867 if (featureTypes.length > 0
868 && (isNullGroup || visibleGroups.contains(group)))
870 features.addAll(sequences[i].getFeatures()
871 .getFeaturesForGroup(true, group, featureTypes));
874 for (SequenceFeature sf : features)
876 if (sf.isNonPositional() || fr.isVisible(sf))
884 out.append(STARTGROUP).append(TAB).append(group)
888 firstInGroup = false;
889 formatJalviewFeature(out, sequenceName, sf);
894 if (!isNullGroup && !firstInGroup)
896 out.append(ENDGROUP).append(TAB).append(group).append(newline);
903 * Formats one feature in Jalview format and appends to the string buffer
906 * @param sequenceName
907 * @param sequenceFeature
909 protected void formatJalviewFeature(StringBuilder out,
910 String sequenceName, SequenceFeature sequenceFeature)
912 if (sequenceFeature.description == null
913 || sequenceFeature.description.equals(""))
915 out.append(sequenceFeature.type).append(TAB);
919 if (sequenceFeature.links != null
920 && sequenceFeature.getDescription().indexOf("<html>") == -1)
922 out.append("<html>");
925 out.append(sequenceFeature.description);
926 if (sequenceFeature.links != null)
928 for (int l = 0; l < sequenceFeature.links.size(); l++)
930 String label = sequenceFeature.links.elementAt(l);
931 String href = label.substring(label.indexOf("|") + 1);
932 label = label.substring(0, label.indexOf("|"));
934 if (sequenceFeature.description.indexOf(href) == -1)
936 out.append(" <a href=\"").append(href).append("\">")
937 .append(label).append("</a>");
941 if (sequenceFeature.getDescription().indexOf("</html>") == -1)
943 out.append("</html>");
949 out.append(sequenceName);
950 out.append("\t-1\t");
951 out.append(sequenceFeature.begin);
953 out.append(sequenceFeature.end);
955 out.append(sequenceFeature.type);
956 if (!Float.isNaN(sequenceFeature.score))
959 out.append(sequenceFeature.score);
965 * Parse method that is called when a GFF file is dragged to the desktop
970 AlignViewportI av = getViewport();
973 if (av.getAlignment() != null)
975 dataset = av.getAlignment().getDataset();
979 // working in the applet context ?
980 dataset = av.getAlignment();
985 dataset = new Alignment(new SequenceI[] {});
988 Map<String, FeatureColourI> featureColours = new HashMap<>();
989 boolean parseResult = parse(dataset, featureColours, false, true);
992 // pass error up somehow
996 // update viewport with the dataset data ?
1000 setSeqs(dataset.getSequencesArray());
1005 * Implementation of unused abstract method
1007 * @return error message
1010 public String print(SequenceI[] sqs, boolean jvsuffix)
1012 jalview.bin.Console.outPrintln("Use printGffFormat() or printJalviewFormat()");
1017 * Returns features output in GFF2 format
1020 * the sequences whose features are to be output
1022 * a map whose keys are the type names of visible features
1023 * @param visibleFeatureGroups
1024 * @param includeNonPositionalFeatures
1025 * @param includeComplement
1028 public String printGffFormat(SequenceI[] sequences, FeatureRenderer fr,
1029 boolean includeNonPositionalFeatures, boolean includeComplement)
1031 FeatureRenderer fr2 = null;
1032 if (includeComplement)
1034 AlignViewportI comp = fr.getViewport().getCodingComplement();
1035 fr2 = Desktop.getAlignFrameFor(comp).getFeatureRenderer();
1038 Map<String, FeatureColourI> visibleColours = fr
1039 .getDisplayedFeatureCols();
1041 StringBuilder out = new StringBuilder(256);
1043 out.append(String.format("%s %d\n", GFF_VERSION,
1044 gffVersion == 0 ? 2 : gffVersion));
1046 String[] types = visibleColours == null ? new String[0]
1047 : visibleColours.keySet()
1048 .toArray(new String[visibleColours.keySet().size()]);
1050 for (SequenceI seq : sequences)
1052 List<SequenceFeature> seqFeatures = new ArrayList<>();
1053 List<SequenceFeature> features = new ArrayList<>();
1054 if (includeNonPositionalFeatures)
1056 features.addAll(seq.getFeatures().getNonPositionalFeatures());
1058 if (visibleColours != null && !visibleColours.isEmpty())
1060 features.addAll(seq.getFeatures().getPositionalFeatures(types));
1062 for (SequenceFeature sf : features)
1064 if (sf.isNonPositional() || fr.isVisible(sf))
1067 * drop features hidden by group visibility, colour threshold,
1068 * or feature filter condition
1070 seqFeatures.add(sf);
1074 if (includeComplement)
1076 seqFeatures.addAll(findComplementaryFeatures(seq, fr2));
1080 * sort features here if wanted
1082 for (SequenceFeature sf : seqFeatures)
1084 formatGffFeature(out, seq, sf);
1085 out.append(newline);
1089 return out.toString();
1093 * Formats one feature as GFF and appends to the string buffer
1095 private void formatGffFeature(StringBuilder out, SequenceI seq,
1098 String source = sf.featureGroup;
1101 source = sf.getDescription();
1104 out.append(seq.getName());
1108 out.append(sf.type);
1110 out.append(sf.begin);
1114 out.append(sf.score);
1117 int strand = sf.getStrand();
1118 out.append(strand == 1 ? "+" : (strand == -1 ? "-" : "."));
1121 String phase = sf.getPhase();
1122 out.append(phase == null ? "." : phase);
1124 if (sf.otherDetails != null && !sf.otherDetails.isEmpty())
1126 Map<String, Object> map = sf.otherDetails;
1127 formatAttributes(out, map);
1132 * A helper method that outputs attributes stored in the map as
1133 * semicolon-delimited values e.g.
1136 * AC_Male=0;AF_NFE=0.00000e 00;Hom_FIN=0;GQ_MEDIAN=9
1139 * A map-valued attribute is formatted as a comma-delimited list within
1140 * braces, for example
1143 * jvmap_CSQ={ALLELE_NUM=1,UNIPARC=UPI0002841053,Feature=ENST00000585561}
1146 * The {@code jvmap_} prefix designates a values map and is removed if the
1147 * value is parsed when read in. (The GFF3 specification allows
1148 * 'semi-structured data' to be represented provided the attribute name begins
1149 * with a lower case letter.)
1153 * @see http://gmod.org/wiki/GFF3#GFF3_Format
1155 void formatAttributes(StringBuilder sb, Map<String, Object> map)
1158 boolean first = true;
1159 for (String key : map.keySet())
1161 if (SequenceFeature.STRAND.equals(key)
1162 || SequenceFeature.PHASE.equals(key))
1165 * values stashed in map but output to their own columns
1176 Object value = map.get(key);
1177 if (value instanceof Map<?, ?>)
1179 formatMapAttribute(sb, key, (Map<?, ?>) value);
1183 String formatted = StringUtils.urlEncode(value.toString(),
1184 GffHelperI.GFF_ENCODABLE);
1185 sb.append(key).append(EQUALS).append(formatted);
1191 * Formats the map entries as
1194 * key=key1=value1,key2=value2,...
1197 * and appends this to the string buffer
1203 private void formatMapAttribute(StringBuilder sb, String key,
1206 if (map == null || map.isEmpty())
1212 * AbstractMap.toString would be a shortcut here, but more reliable
1213 * to code the required format in case toString changes in future
1215 sb.append(key).append(EQUALS);
1216 boolean first = true;
1217 for (Entry<?, ?> entry : map.entrySet())
1224 sb.append(entry.getKey().toString()).append(EQUALS);
1225 String formatted = StringUtils.urlEncode(entry.getValue().toString(),
1226 GffHelperI.GFF_ENCODABLE);
1227 sb.append(formatted);
1232 * Returns a mapping given list of one or more Align descriptors (exonerate
1235 * @param alignedRegions
1236 * a list of "Align fromStart toStart fromCount"
1237 * @param mapIsFromCdna
1238 * if true, 'from' is dna, else 'from' is protein
1240 * either 1 (forward) or -1 (reverse)
1242 * @throws IOException
1244 protected MapList constructCodonMappingFromAlign(
1245 List<String> alignedRegions, boolean mapIsFromCdna, int strand)
1250 throw new IOException(
1251 "Invalid strand for a codon mapping (cannot be 0)");
1253 int regions = alignedRegions.size();
1254 // arrays to hold [start, end] for each aligned region
1255 int[] fromRanges = new int[regions * 2]; // from dna
1256 int[] toRanges = new int[regions * 2]; // to protein
1257 int fromRangesIndex = 0;
1258 int toRangesIndex = 0;
1260 for (String range : alignedRegions)
1263 * Align mapFromStart mapToStart mapFromCount
1264 * e.g. if mapIsFromCdna
1265 * Align 11270 143 120
1267 * 120 bases from pos 11270 align to pos 143 in peptide
1268 * if !mapIsFromCdna this would instead be
1269 * Align 143 11270 40
1271 String[] tokens = range.split(" ");
1272 if (tokens.length != 3)
1274 throw new IOException("Wrong number of fields for Align");
1281 fromStart = Integer.parseInt(tokens[0]);
1282 toStart = Integer.parseInt(tokens[1]);
1283 fromCount = Integer.parseInt(tokens[2]);
1284 } catch (NumberFormatException nfe)
1286 throw new IOException(
1287 "Invalid number in Align field: " + nfe.getMessage());
1291 * Jalview always models from dna to protein, so adjust values if the
1292 * GFF mapping is from protein to dna
1297 int temp = fromStart;
1298 fromStart = toStart;
1301 fromRanges[fromRangesIndex++] = fromStart;
1302 fromRanges[fromRangesIndex++] = fromStart + strand * (fromCount - 1);
1305 * If a codon has an intron gap, there will be contiguous 'toRanges';
1306 * this is handled for us by the MapList constructor.
1307 * (It is not clear that exonerate ever generates this case)
1309 toRanges[toRangesIndex++] = toStart;
1310 toRanges[toRangesIndex++] = toStart + (fromCount - 1) / 3;
1313 return new MapList(fromRanges, toRanges, 3, 1);
1317 * Parse a GFF format feature. This may include creating a 'dummy' sequence to
1318 * hold the feature, or for its mapped sequence, or both, to be resolved
1319 * either later in the GFF file (##FASTA section), or when the user loads
1320 * additional sequences.
1324 * @param relaxedIdMatching
1328 protected SequenceI parseGff(String[] gffColumns, AlignmentI alignment,
1329 boolean relaxedIdMatching, List<SequenceI> newseqs)
1332 * GFF: seqid source type start end score strand phase [attributes]
1334 if (gffColumns.length < 5)
1336 jalview.bin.Console.errPrintln("Ignoring GFF feature line with too few columns ("
1337 + gffColumns.length + ")");
1342 * locate referenced sequence in alignment _or_
1343 * as a forward or external reference (SequenceDummy)
1345 String seqId = gffColumns[0];
1346 SequenceI seq = findSequence(seqId, alignment, newseqs,
1349 SequenceFeature sf = null;
1350 GffHelperI helper = GffHelperFactory.getHelper(gffColumns);
1355 sf = helper.processGff(seq, gffColumns, alignment, newseqs,
1359 seq.addSequenceFeature(sf);
1360 while ((seq = alignment.findName(seq, seqId, true)) != null)
1362 seq.addSequenceFeature(new SequenceFeature(sf));
1365 } catch (IOException e)
1367 jalview.bin.Console.errPrintln("GFF parsing failed with: " + e.getMessage());
1376 * After encountering ##fasta in a GFF3 file, process the remainder of the
1377 * file as FAST sequence data. Any placeholder sequences created during
1378 * feature parsing are updated with the actual sequences.
1382 * @throws IOException
1384 protected void processAsFasta(AlignmentI align, List<SequenceI> newseqs)
1390 } catch (IOException q)
1393 // Opening a FastaFile object with the remainder of this object's dataIn.
1394 // Tell the constructor to NOT close the dataIn when finished.
1395 FastaFile parser = new FastaFile(this, false);
1396 List<SequenceI> includedseqs = parser.getSeqs();
1398 SequenceIdMatcher smatcher = new SequenceIdMatcher(newseqs);
1401 * iterate over includedseqs, and replacing matching ones with newseqs
1402 * sequences. Generic iterator not used here because we modify
1403 * includedseqs as we go
1405 for (int p = 0, pSize = includedseqs.size(); p < pSize; p++)
1407 // search for any dummy seqs that this sequence can be used to update
1408 SequenceI includedSeq = includedseqs.get(p);
1409 SequenceI dummyseq = smatcher.findIdMatch(includedSeq);
1410 if (dummyseq != null && dummyseq instanceof SequenceDummy)
1412 // probably have the pattern wrong
1413 // idea is that a flyweight proxy for a sequence ID can be created for
1414 // 1. stable reference creation
1415 // 2. addition of annotation
1416 // 3. future replacement by a real sequence
1417 // current pattern is to create SequenceDummy objects - a convenience
1418 // constructor for a Sequence.
1419 // problem is that when promoted to a real sequence, all references
1420 // need to be updated somehow. We avoid that by keeping the same object.
1421 ((SequenceDummy) dummyseq).become(includedSeq);
1422 dummyseq.createDatasetSequence();
1425 * Update mappings so they are now to the dataset sequence
1427 for (AlignedCodonFrame mapping : align.getCodonFrames())
1429 mapping.updateToDataset(dummyseq);
1433 * replace parsed sequence with the realised forward reference
1435 includedseqs.set(p, dummyseq);
1438 * and remove from the newseqs list
1440 newseqs.remove(dummyseq);
1445 * finally add sequences to the dataset
1447 for (SequenceI seq : includedseqs)
1449 // experimental: mapping-based 'alignment' to query sequence
1450 AlignmentUtils.alignSequenceAs(seq, align,
1451 String.valueOf(align.getGapCharacter()), false, true);
1453 // rename sequences if GFF handler requested this
1454 // TODO a more elegant way e.g. gffHelper.postProcess(newseqs) ?
1455 List<SequenceFeature> sfs = seq.getFeatures().getPositionalFeatures();
1458 String newName = (String) sfs.get(0)
1459 .getValue(GffHelperI.RENAME_TOKEN);
1460 if (newName != null)
1462 seq.setName(newName);
1465 align.addSequence(seq);
1470 * Process a ## directive
1476 * @throws IOException
1478 protected void processGffPragma(String line, Map<String, String> gffProps,
1479 AlignmentI align, List<SequenceI> newseqs) throws IOException
1482 if ("###".equals(line))
1484 // close off any open 'forward references'
1488 String[] tokens = line.substring(2).split(" ");
1489 String pragma = tokens[0];
1490 String value = tokens.length == 1 ? null : tokens[1];
1492 if ("gff-version".equalsIgnoreCase(pragma))
1498 // value may be e.g. "3.1.2"
1499 gffVersion = Integer.parseInt(value.split("\\.")[0]);
1500 } catch (NumberFormatException e)
1506 else if ("sequence-region".equalsIgnoreCase(pragma))
1508 // could capture <seqid start end> if wanted here
1510 else if ("feature-ontology".equalsIgnoreCase(pragma))
1512 // should resolve against the specified feature ontology URI
1514 else if ("attribute-ontology".equalsIgnoreCase(pragma))
1516 // URI of attribute ontology - not currently used in GFF3
1518 else if ("source-ontology".equalsIgnoreCase(pragma))
1520 // URI of source ontology - not currently used in GFF3
1522 else if ("species-build".equalsIgnoreCase(pragma))
1524 // save URI of specific NCBI taxon version of annotations
1525 gffProps.put("species-build", value);
1527 else if ("fasta".equalsIgnoreCase(pragma))
1529 // process the rest of the file as a fasta file and replace any dummy
1531 processAsFasta(align, newseqs);
1535 jalview.bin.Console.errPrintln("Ignoring unknown pragma: " + line);