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 jalview.analysis.AlignmentUtils;
24 import jalview.analysis.SequenceIdMatcher;
25 import jalview.api.AlignViewportI;
26 import jalview.api.FeatureColourI;
27 import jalview.api.FeaturesSourceI;
28 import jalview.datamodel.AlignedCodonFrame;
29 import jalview.datamodel.Alignment;
30 import jalview.datamodel.AlignmentI;
31 import jalview.datamodel.SequenceDummy;
32 import jalview.datamodel.SequenceFeature;
33 import jalview.datamodel.SequenceI;
34 import jalview.io.gff.GffHelperBase;
35 import jalview.io.gff.GffHelperFactory;
36 import jalview.io.gff.GffHelperI;
37 import jalview.schemes.FeatureColour;
38 import jalview.util.ColorUtils;
39 import jalview.util.MapList;
40 import jalview.util.ParseHtmlBodyAndLinks;
41 import jalview.util.StringUtils;
43 import java.awt.Color;
44 import java.io.IOException;
45 import java.util.ArrayList;
46 import java.util.Arrays;
47 import java.util.HashMap;
48 import java.util.Iterator;
49 import java.util.List;
51 import java.util.Map.Entry;
54 * Parses and writes features files, which may be in Jalview, GFF2 or GFF3
55 * format. These are tab-delimited formats but with differences in the use of
58 * A Jalview feature file may define feature colours and then declare that the
59 * remainder of the file is in GFF format with the line 'GFF'.
61 * GFF3 files may include alignment mappings for features, which Jalview will
62 * attempt to model, and may include sequence data following a ##FASTA line.
69 public class FeaturesFile extends AlignFile implements FeaturesSourceI
71 private static final String ID_NOT_SPECIFIED = "ID_NOT_SPECIFIED";
73 private static final String NOTE = "Note";
75 protected static final String TAB = "\t";
77 protected static final String GFF_VERSION = "##gff-version";
79 private AlignmentI lastmatchedAl = null;
81 private SequenceIdMatcher matcher = null;
83 protected AlignmentI dataset;
85 protected int gffVersion;
88 * Creates a new FeaturesFile object.
95 * Constructor which does not parse the file immediately
101 public FeaturesFile(String inFile, DataSourceType paste)
104 super(false, inFile, paste);
109 * @throws IOException
111 public FeaturesFile(FileParse source) throws IOException
117 * Constructor that optionally parses the file immediately
119 * @param parseImmediately
122 * @throws IOException
124 public FeaturesFile(boolean parseImmediately, String inFile,
125 DataSourceType type) throws IOException
127 super(parseImmediately, inFile, type);
131 * Parse GFF or sequence features file using case-independent matching,
135 * - alignment/dataset containing sequences that are to be annotated
137 * - hashtable to store feature colour definitions
139 * - process html strings into plain text
140 * @return true if features were added
142 public boolean parse(AlignmentI align,
143 Map<String, FeatureColourI> colours, boolean removeHTML)
145 return parse(align, colours, removeHTML, false);
149 * Extends the default addProperties by also adding peptide-to-cDNA mappings
150 * (if any) derived while parsing a GFF file
153 public void addProperties(AlignmentI al)
155 super.addProperties(al);
156 if (dataset != null && dataset.getCodonFrames() != null)
158 AlignmentI ds = (al.getDataset() == null) ? al : al.getDataset();
159 for (AlignedCodonFrame codons : dataset.getCodonFrames())
161 ds.addCodonFrame(codons);
167 * Parse GFF or Jalview format sequence features file
170 * - alignment/dataset containing sequences that are to be annotated
172 * - hashtable to store feature colour definitions
174 * - process html strings into plain text
175 * @param relaxedIdmatching
176 * - when true, ID matches to compound sequence IDs are allowed
177 * @return true if features were added
179 public boolean parse(AlignmentI align,
180 Map<String, FeatureColourI> colours, boolean removeHTML,
181 boolean relaxedIdmatching)
183 Map<String, String> gffProps = new HashMap<String, String>();
185 * keep track of any sequences we try to create from the data
187 List<SequenceI> newseqs = new ArrayList<SequenceI>();
193 String featureGroup = null;
195 while ((line = nextLine()) != null)
197 // skip comments/process pragmas
198 if (line.length() == 0 || line.startsWith("#"))
200 if (line.toLowerCase().startsWith("##"))
202 processGffPragma(line, gffProps, align, newseqs);
207 gffColumns = line.split("\\t"); // tab as regex
208 if (gffColumns.length == 1)
210 if (line.trim().equalsIgnoreCase("GFF"))
213 * Jalview features file with appended GFF
214 * assume GFF2 (though it may declare ##gff-version 3)
221 if (gffColumns.length > 1 && gffColumns.length < 4)
224 * if 2 or 3 tokens, we anticipate either 'startgroup', 'endgroup' or
225 * a feature type colour specification
227 String ft = gffColumns[0];
228 if (ft.equalsIgnoreCase("startgroup"))
230 featureGroup = gffColumns[1];
232 else if (ft.equalsIgnoreCase("endgroup"))
234 // We should check whether this is the current group,
235 // but at present there's no way of showing more than 1 group
240 String colscheme = gffColumns[1];
241 FeatureColourI colour = FeatureColour
242 .parseJalviewFeatureColour(colscheme);
245 colours.put(ft, colour);
252 * if not a comment, GFF pragma, startgroup, endgroup or feature
253 * colour specification, that just leaves a feature details line
254 * in either Jalview or GFF format
258 parseJalviewFeature(line, gffColumns, align, colours, removeHTML,
259 relaxedIdmatching, featureGroup);
263 parseGff(gffColumns, align, relaxedIdmatching, newseqs);
267 } catch (Exception ex)
269 // should report somewhere useful for UI if necessary
270 warningMessage = ((warningMessage == null) ? "" : warningMessage)
271 + "Parsing error at\n" + line;
272 System.out.println("Error parsing feature file: " + ex + "\n" + line);
273 ex.printStackTrace(System.err);
279 * experimental - add any dummy sequences with features to the alignment
280 * - we need them for Ensembl feature extraction - though maybe not otherwise
282 for (SequenceI newseq : newseqs)
284 if (newseq.getSequenceFeatures() != null)
286 align.addSequence(newseq);
293 * Try to parse a Jalview format feature specification and add it as a
294 * sequence feature to any matching sequences in the alignment. Returns true
295 * if successful (a feature was added), or false if not.
300 * @param featureColours
302 * @param relaxedIdmatching
303 * @param featureGroup
305 protected boolean parseJalviewFeature(String line, String[] gffColumns,
306 AlignmentI alignment, Map<String, FeatureColourI> featureColours,
307 boolean removeHTML, boolean relaxedIdMatching,
311 * tokens: description seqid seqIndex start end type [score]
313 if (gffColumns.length < 6)
315 System.err.println("Ignoring feature line '" + line
316 + "' with too few columns (" + gffColumns.length + ")");
319 String desc = gffColumns[0];
320 String seqId = gffColumns[1];
321 SequenceI seq = findSequence(seqId, alignment, null, relaxedIdMatching);
323 if (!ID_NOT_SPECIFIED.equals(seqId))
325 seq = findSequence(seqId, alignment, null, relaxedIdMatching);
331 String seqIndex = gffColumns[2];
334 int idx = Integer.parseInt(seqIndex);
335 seq = alignment.getSequenceAt(idx);
336 } catch (NumberFormatException ex)
338 System.err.println("Invalid sequence index: " + seqIndex);
344 System.out.println("Sequence not found: " + line);
348 int startPos = Integer.parseInt(gffColumns[3]);
349 int endPos = Integer.parseInt(gffColumns[4]);
351 String ft = gffColumns[5];
353 if (!featureColours.containsKey(ft))
356 * Perhaps an old style groups file with no colours -
357 * synthesize a colour from the feature type
359 Color colour = ColorUtils.createColourFromName(ft);
360 featureColours.put(ft, new FeatureColour(colour));
362 SequenceFeature sf = new SequenceFeature(ft, desc, "", startPos, endPos,
364 if (gffColumns.length > 6)
366 float score = Float.NaN;
369 score = new Float(gffColumns[6]).floatValue();
370 // update colourgradient bounds if allowed to
371 } catch (NumberFormatException ex)
378 parseDescriptionHTML(sf, removeHTML);
380 seq.addSequenceFeature(sf);
383 && (seq = alignment.findName(seq, seqId, false)) != null)
385 seq.addSequenceFeature(new SequenceFeature(sf));
391 * clear any temporary handles used to speed up ID matching
393 protected void resetMatcher()
395 lastmatchedAl = null;
400 * Returns a sequence matching the given id, as follows
402 * <li>strict matching is on exact sequence name</li>
403 * <li>relaxed matching allows matching on a token within the sequence name,
405 * <li>first tries to find a match in the alignment sequences</li>
406 * <li>else tries to find a match in the new sequences already generated while
407 * parsing the features file</li>
408 * <li>else creates a new placeholder sequence, adds it to the new sequences
409 * list, and returns it</li>
415 * @param relaxedIdMatching
419 protected SequenceI findSequence(String seqId, AlignmentI align,
420 List<SequenceI> newseqs, boolean relaxedIdMatching)
422 // TODO encapsulate in SequenceIdMatcher, share the matcher
423 // with the GffHelper (removing code duplication)
424 SequenceI match = null;
425 if (relaxedIdMatching)
427 if (lastmatchedAl != align)
429 lastmatchedAl = align;
430 matcher = new SequenceIdMatcher(align.getSequencesArray());
433 matcher.addAll(newseqs);
436 match = matcher.findIdMatch(seqId);
440 match = align.findName(seqId, true);
441 if (match == null && newseqs != null)
443 for (SequenceI m : newseqs)
445 if (seqId.equals(m.getName()))
453 if (match == null && newseqs != null)
455 match = new SequenceDummy(seqId);
456 if (relaxedIdMatching)
458 matcher.addAll(Arrays.asList(new SequenceI[] { match }));
460 // add dummy sequence to the newseqs list
466 public void parseDescriptionHTML(SequenceFeature sf, boolean removeHTML)
468 if (sf.getDescription() == null)
472 ParseHtmlBodyAndLinks parsed = new ParseHtmlBodyAndLinks(
473 sf.getDescription(), removeHTML, newline);
475 sf.description = (removeHTML) ? parsed.getNonHtmlContent()
477 for (String link : parsed.getLinks())
485 * generate a features file for seqs includes non-pos features by default.
488 * source of sequence features
490 * hash of feature types and colours
491 * @return features file contents
493 public String printJalviewFormat(SequenceI[] sequences,
494 Map<String, FeatureColourI> visible)
496 return printJalviewFormat(sequences, visible, true, true);
500 * generate a features file for seqs with colours from visible (if any)
505 * hash of Colours for each feature type
507 * when true only feature types in 'visible' will be output
509 * indicates if non-positional features should be output (regardless
511 * @return features file contents
513 public String printJalviewFormat(SequenceI[] sequences,
514 Map<String, FeatureColourI> visible, boolean visOnly,
517 StringBuilder out = new StringBuilder(256);
518 boolean featuresGen = false;
519 if (visOnly && !nonpos && (visible == null || visible.size() < 1))
521 // no point continuing.
522 return "No Features Visible";
525 if (visible != null && visOnly)
527 // write feature colours only if we're given them and we are generating
529 // TODO: decide if feature links should also be written here ?
530 Iterator<String> en = visible.keySet().iterator();
533 String featureType = en.next().toString();
534 FeatureColourI colour = visible.get(featureType);
535 out.append(colour.toJalviewFormat(featureType)).append(newline);
539 // Work out which groups are both present and visible
540 List<String> groups = new ArrayList<String>();
542 boolean isnonpos = false;
544 SequenceFeature[] features;
545 for (int i = 0; i < sequences.length; i++)
547 features = sequences[i].getSequenceFeatures();
548 if (features != null)
550 for (int j = 0; j < features.length; j++)
552 isnonpos = features[j].begin == 0 && features[j].end == 0;
553 if ((!nonpos && isnonpos) || (!isnonpos && visOnly
554 && !visible.containsKey(features[j].type)))
559 if (features[j].featureGroup != null
560 && !groups.contains(features[j].featureGroup))
562 groups.add(features[j].featureGroup);
571 if (groups.size() > 0 && groupIndex < groups.size())
573 group = groups.get(groupIndex);
575 out.append("STARTGROUP").append(TAB);
584 for (int i = 0; i < sequences.length; i++)
586 features = sequences[i].getSequenceFeatures();
587 if (features != null)
589 for (SequenceFeature sequenceFeature : features)
591 isnonpos = sequenceFeature.begin == 0
592 && sequenceFeature.end == 0;
593 if ((!nonpos && isnonpos) || (!isnonpos && visOnly
594 && !visible.containsKey(sequenceFeature.type)))
596 // skip if feature is nonpos and we ignore them or if we only
597 // output visible and it isn't non-pos and it's not visible
601 if (group != null && (sequenceFeature.featureGroup == null
602 || !sequenceFeature.featureGroup.equals(group)))
607 if (group == null && sequenceFeature.featureGroup != null)
611 // we have features to output
613 if (sequenceFeature.description == null
614 || sequenceFeature.description.equals(""))
616 out.append(sequenceFeature.type).append(TAB);
620 if (sequenceFeature.links != null && sequenceFeature
621 .getDescription().indexOf("<html>") == -1)
623 out.append("<html>");
626 out.append(sequenceFeature.description);
627 if (sequenceFeature.links != null)
629 for (int l = 0; l < sequenceFeature.links.size(); l++)
631 String label = sequenceFeature.links.elementAt(l);
632 String href = label.substring(label.indexOf("|") + 1);
633 label = label.substring(0, label.indexOf("|"));
635 if (sequenceFeature.description.indexOf(href) == -1)
638 " <a href=\"" + href + "\">" + label + "</a>");
642 if (sequenceFeature.getDescription()
643 .indexOf("</html>") == -1)
645 out.append("</html>");
651 out.append(sequences[i].getName());
652 out.append("\t-1\t");
653 out.append(sequenceFeature.begin);
655 out.append(sequenceFeature.end);
657 out.append(sequenceFeature.type);
658 if (!Float.isNaN(sequenceFeature.score))
661 out.append(sequenceFeature.score);
670 out.append("ENDGROUP").append(TAB);
680 } while (groupIndex < groups.size() + 1);
684 return "No Features Visible";
687 return out.toString();
691 * Parse method that is called when a GFF file is dragged to the desktop
696 AlignViewportI av = getViewport();
699 if (av.getAlignment() != null)
701 dataset = av.getAlignment().getDataset();
705 // working in the applet context ?
706 dataset = av.getAlignment();
711 dataset = new Alignment(new SequenceI[] {});
714 Map<String, FeatureColourI> featureColours = new HashMap<String, FeatureColourI>();
715 boolean parseResult = parse(dataset, featureColours, false, true);
718 // pass error up somehow
722 // update viewport with the dataset data ?
726 setSeqs(dataset.getSequencesArray());
731 * Implementation of unused abstract method
733 * @return error message
736 public String print(SequenceI[] sqs, boolean jvsuffix)
738 System.out.println("Use printGffFormat() or printJalviewFormat()");
743 * Returns features output in GFF2 format, including hidden and non-positional
747 * the sequences whose features are to be output
749 * a map whose keys are the type names of visible features
752 public String printGffFormat(SequenceI[] sequences,
753 Map<String, FeatureColourI> visible)
755 return printGffFormat(sequences, visible, true, true);
759 * Returns features output in GFF2 format
762 * the sequences whose features are to be output
764 * a map whose keys are the type names of visible features
765 * @param outputVisibleOnly
766 * @param includeNonPositionalFeatures
769 public String printGffFormat(SequenceI[] sequences,
770 Map<String, FeatureColourI> visible, boolean outputVisibleOnly,
771 boolean includeNonPositionalFeatures)
773 StringBuilder out = new StringBuilder(256);
774 int version = gffVersion == 0 ? 2 : gffVersion;
775 out.append(String.format("%s %d\n", GFF_VERSION, version));
778 for (SequenceI seq : sequences)
780 SequenceFeature[] features = seq.getSequenceFeatures();
781 if (features != null)
783 for (SequenceFeature sf : features)
785 isnonpos = sf.begin == 0 && sf.end == 0;
786 if (!includeNonPositionalFeatures && isnonpos)
789 * ignore non-positional features if not wanted
793 // TODO why the test !isnonpos here?
794 // what about not visible non-positional features?
795 if (!isnonpos && outputVisibleOnly
796 && !visible.containsKey(sf.type))
799 * ignore not visible features if not wanted
804 source = sf.featureGroup;
807 source = sf.getDescription();
810 out.append(seq.getName());
816 out.append(sf.begin);
820 out.append(sf.score);
823 int strand = sf.getStrand();
824 out.append(strand == 1 ? "+" : (strand == -1 ? "-" : "."));
827 String phase = sf.getPhase();
828 out.append(phase == null ? "." : phase);
830 // miscellaneous key-values (GFF column 9)
831 String attributes = sf.getAttributes();
832 if (attributes != null)
834 out.append(TAB).append(attributes);
842 return out.toString();
846 * Returns a mapping given list of one or more Align descriptors (exonerate
849 * @param alignedRegions
850 * a list of "Align fromStart toStart fromCount"
851 * @param mapIsFromCdna
852 * if true, 'from' is dna, else 'from' is protein
854 * either 1 (forward) or -1 (reverse)
856 * @throws IOException
858 protected MapList constructCodonMappingFromAlign(
859 List<String> alignedRegions, boolean mapIsFromCdna, int strand)
864 throw new IOException(
865 "Invalid strand for a codon mapping (cannot be 0)");
867 int regions = alignedRegions.size();
868 // arrays to hold [start, end] for each aligned region
869 int[] fromRanges = new int[regions * 2]; // from dna
870 int[] toRanges = new int[regions * 2]; // to protein
871 int fromRangesIndex = 0;
872 int toRangesIndex = 0;
874 for (String range : alignedRegions)
877 * Align mapFromStart mapToStart mapFromCount
878 * e.g. if mapIsFromCdna
879 * Align 11270 143 120
881 * 120 bases from pos 11270 align to pos 143 in peptide
882 * if !mapIsFromCdna this would instead be
885 String[] tokens = range.split(" ");
886 if (tokens.length != 3)
888 throw new IOException("Wrong number of fields for Align");
895 fromStart = Integer.parseInt(tokens[0]);
896 toStart = Integer.parseInt(tokens[1]);
897 fromCount = Integer.parseInt(tokens[2]);
898 } catch (NumberFormatException nfe)
900 throw new IOException(
901 "Invalid number in Align field: " + nfe.getMessage());
905 * Jalview always models from dna to protein, so adjust values if the
906 * GFF mapping is from protein to dna
911 int temp = fromStart;
915 fromRanges[fromRangesIndex++] = fromStart;
916 fromRanges[fromRangesIndex++] = fromStart + strand * (fromCount - 1);
919 * If a codon has an intron gap, there will be contiguous 'toRanges';
920 * this is handled for us by the MapList constructor.
921 * (It is not clear that exonerate ever generates this case)
923 toRanges[toRangesIndex++] = toStart;
924 toRanges[toRangesIndex++] = toStart + (fromCount - 1) / 3;
927 return new MapList(fromRanges, toRanges, 3, 1);
931 * Parse a GFF format feature. This may include creating a 'dummy' sequence to
932 * hold the feature, or for its mapped sequence, or both, to be resolved
933 * either later in the GFF file (##FASTA section), or when the user loads
934 * additional sequences.
938 * @param relaxedIdMatching
942 protected SequenceI parseGff(String[] gffColumns, AlignmentI alignment,
943 boolean relaxedIdMatching, List<SequenceI> newseqs)
946 * GFF: seqid source type start end score strand phase [attributes]
948 if (gffColumns.length < 5)
950 System.err.println("Ignoring GFF feature line with too few columns ("
951 + gffColumns.length + ")");
956 * locate referenced sequence in alignment _or_
957 * as a forward or external reference (SequenceDummy)
959 String seqId = gffColumns[0];
960 SequenceI seq = findSequence(seqId, alignment, newseqs,
963 SequenceFeature sf = null;
964 GffHelperI helper = GffHelperFactory.getHelper(gffColumns);
969 sf = helper.processGff(seq, gffColumns, alignment, newseqs,
973 seq.addSequenceFeature(sf);
974 while ((seq = alignment.findName(seq, seqId, true)) != null)
976 seq.addSequenceFeature(new SequenceFeature(sf));
979 } catch (IOException e)
981 System.err.println("GFF parsing failed with: " + e.getMessage());
990 * Process the 'column 9' data of the GFF file. This is less formally defined,
991 * and its interpretation will vary depending on the tool that has generated
997 protected void processGffColumnNine(String attributes, SequenceFeature sf)
999 sf.setAttributes(attributes);
1002 * Parse attributes in column 9 and add them to the sequence feature's
1003 * 'otherData' table; use Note as a best proxy for description
1005 char nameValueSeparator = gffVersion == 3 ? '=' : ' ';
1006 // TODO check we don't break GFF2 values which include commas here
1007 Map<String, List<String>> nameValues = GffHelperBase
1008 .parseNameValuePairs(attributes, ";", nameValueSeparator, ",");
1009 for (Entry<String, List<String>> attr : nameValues.entrySet())
1011 String values = StringUtils.listToDelimitedString(attr.getValue(),
1013 sf.setValue(attr.getKey(), values);
1014 if (NOTE.equals(attr.getKey()))
1016 sf.setDescription(values);
1022 * After encountering ##fasta in a GFF3 file, process the remainder of the
1023 * file as FAST sequence data. Any placeholder sequences created during
1024 * feature parsing are updated with the actual sequences.
1028 * @throws IOException
1030 protected void processAsFasta(AlignmentI align, List<SequenceI> newseqs)
1036 } catch (IOException q)
1039 FastaFile parser = new FastaFile(this);
1040 List<SequenceI> includedseqs = parser.getSeqs();
1042 SequenceIdMatcher smatcher = new SequenceIdMatcher(newseqs);
1045 * iterate over includedseqs, and replacing matching ones with newseqs
1046 * sequences. Generic iterator not used here because we modify
1047 * includedseqs as we go
1049 for (int p = 0, pSize = includedseqs.size(); p < pSize; p++)
1051 // search for any dummy seqs that this sequence can be used to update
1052 SequenceI includedSeq = includedseqs.get(p);
1053 SequenceI dummyseq = smatcher.findIdMatch(includedSeq);
1054 if (dummyseq != null && dummyseq instanceof SequenceDummy)
1056 // probably have the pattern wrong
1057 // idea is that a flyweight proxy for a sequence ID can be created for
1058 // 1. stable reference creation
1059 // 2. addition of annotation
1060 // 3. future replacement by a real sequence
1061 // current pattern is to create SequenceDummy objects - a convenience
1062 // constructor for a Sequence.
1063 // problem is that when promoted to a real sequence, all references
1064 // need to be updated somehow. We avoid that by keeping the same object.
1065 ((SequenceDummy) dummyseq).become(includedSeq);
1066 dummyseq.createDatasetSequence();
1069 * Update mappings so they are now to the dataset sequence
1071 for (AlignedCodonFrame mapping : align.getCodonFrames())
1073 mapping.updateToDataset(dummyseq);
1077 * replace parsed sequence with the realised forward reference
1079 includedseqs.set(p, dummyseq);
1082 * and remove from the newseqs list
1084 newseqs.remove(dummyseq);
1089 * finally add sequences to the dataset
1091 for (SequenceI seq : includedseqs)
1093 // experimental: mapping-based 'alignment' to query sequence
1094 AlignmentUtils.alignSequenceAs(seq, align,
1095 String.valueOf(align.getGapCharacter()), false, true);
1097 // rename sequences if GFF handler requested this
1098 // TODO a more elegant way e.g. gffHelper.postProcess(newseqs) ?
1099 SequenceFeature[] sfs = seq.getSequenceFeatures();
1102 String newName = (String) sfs[0].getValue(GffHelperI.RENAME_TOKEN);
1103 if (newName != null)
1105 seq.setName(newName);
1108 align.addSequence(seq);
1113 * Process a ## directive
1119 * @throws IOException
1121 protected void processGffPragma(String line, Map<String, String> gffProps,
1122 AlignmentI align, List<SequenceI> newseqs) throws IOException
1125 if ("###".equals(line))
1127 // close off any open 'forward references'
1131 String[] tokens = line.substring(2).split(" ");
1132 String pragma = tokens[0];
1133 String value = tokens.length == 1 ? null : tokens[1];
1135 if ("gff-version".equalsIgnoreCase(pragma))
1141 // value may be e.g. "3.1.2"
1142 gffVersion = Integer.parseInt(value.split("\\.")[0]);
1143 } catch (NumberFormatException e)
1149 else if ("sequence-region".equalsIgnoreCase(pragma))
1151 // could capture <seqid start end> if wanted here
1153 else if ("feature-ontology".equalsIgnoreCase(pragma))
1155 // should resolve against the specified feature ontology URI
1157 else if ("attribute-ontology".equalsIgnoreCase(pragma))
1159 // URI of attribute ontology - not currently used in GFF3
1161 else if ("source-ontology".equalsIgnoreCase(pragma))
1163 // URI of source ontology - not currently used in GFF3
1165 else if ("species-build".equalsIgnoreCase(pragma))
1167 // save URI of specific NCBI taxon version of annotations
1168 gffProps.put("species-build", value);
1170 else if ("fasta".equalsIgnoreCase(pragma))
1172 // process the rest of the file as a fasta file and replace any dummy
1174 processAsFasta(align, newseqs);
1178 System.err.println("Ignoring unknown pragma: " + line);