X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fio%2FFeaturesFile.java;h=73783e5b408a0dac903dabe9d10d480626651a16;hb=37de9310bec3501cbc6381e0c3dcb282fcaad812;hp=804cefc682381bc37a41d3cf15849d3317dbe8b2;hpb=25aaaa87042b3f507ad4348120df7dd073182759;p=jalview.git diff --git a/src/jalview/io/FeaturesFile.java b/src/jalview/io/FeaturesFile.java index 804cefc..73783e5 100755 --- a/src/jalview/io/FeaturesFile.java +++ b/src/jalview/io/FeaturesFile.java @@ -1,6 +1,6 @@ /* - * Jalview - A Sequence Alignment Editor and Viewer (Version 2.8.2b1) - * Copyright (C) 2014 The Jalview Authors + * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$) + * Copyright (C) $$Year-Rel$$ The Jalview Authors * * This file is part of Jalview. * @@ -20,30 +20,68 @@ */ package jalview.io; -import java.io.*; -import java.util.*; - +import jalview.analysis.AlignmentUtils; import jalview.analysis.SequenceIdMatcher; -import jalview.datamodel.*; -import jalview.schemes.*; -import jalview.util.Format; +import jalview.api.AlignViewportI; +import jalview.api.FeatureColourI; +import jalview.api.FeaturesSourceI; +import jalview.datamodel.AlignedCodonFrame; +import jalview.datamodel.Alignment; +import jalview.datamodel.AlignmentI; +import jalview.datamodel.SequenceDummy; +import jalview.datamodel.SequenceFeature; +import jalview.datamodel.SequenceI; +import jalview.io.gff.GffHelperBase; +import jalview.io.gff.GffHelperFactory; +import jalview.io.gff.GffHelperI; +import jalview.schemes.FeatureColour; +import jalview.schemes.UserColourScheme; +import jalview.util.MapList; +import jalview.util.ParseHtmlBodyAndLinks; +import jalview.util.StringUtils; + +import java.io.IOException; +import java.util.ArrayList; +import java.util.Arrays; +import java.util.HashMap; +import java.util.Iterator; +import java.util.List; +import java.util.Map; +import java.util.Map.Entry; /** - * Parse and create Jalview Features files Detects GFF format features files and - * parses. Does not implement standard print() - call specific printFeatures or - * printGFF. Uses AlignmentI.findSequence(String id) to find the sequence object - * for the features annotation - this normally works on an exact match. + * Parses and writes features files, which may be in Jalview, GFF2 or GFF3 + * format. These are tab-delimited formats but with differences in the use of + * columns. + * + * A Jalview feature file may define feature colours and then declare that the + * remainder of the file is in GFF format with the line 'GFF'. + * + * GFF3 files may include alignment mappings for features, which Jalview will + * attempt to model, and may include sequence data following a ##FASTA line. + * * * @author AMW - * @version $Revision$ + * @author jbprocter + * @author gmcarstairs */ -public class FeaturesFile extends AlignFile +public class FeaturesFile extends AlignFile implements FeaturesSourceI { - /** - * work around for GFF interpretation bug where source string becomes - * description rather than a group - */ - private boolean doGffSource = true; + private static final String ID_NOT_SPECIFIED = "ID_NOT_SPECIFIED"; + + private static final String NOTE = "Note"; + + protected static final String TAB = "\t"; + + protected static final String GFF_VERSION = "##gff-version"; + + private AlignmentI lastmatchedAl = null; + + private SequenceIdMatcher matcher = null; + + protected AlignmentI dataset; + + protected int gffVersion; /** * Creates a new FeaturesFile object. @@ -53,47 +91,43 @@ public class FeaturesFile extends AlignFile } /** - * Creates a new FeaturesFile object. + * Constructor which does not parse the file immediately * * @param inFile - * DOCUMENT ME! * @param type - * DOCUMENT ME! - * * @throws IOException - * DOCUMENT ME! */ public FeaturesFile(String inFile, String type) throws IOException { - super(inFile, type); + super(false, inFile, type); } + /** + * @param source + * @throws IOException + */ public FeaturesFile(FileParse source) throws IOException { super(source); } /** - * Parse GFF or sequence features file using case-independent matching, - * discarding URLs + * Constructor that optionally parses the file immediately * - * @param align - * - alignment/dataset containing sequences that are to be annotated - * @param colours - * - hashtable to store feature colour definitions - * @param removeHTML - * - process html strings into plain text - * @return true if features were added + * @param parseImmediately + * @param inFile + * @param type + * @throws IOException */ - public boolean parse(AlignmentI align, Hashtable colours, - boolean removeHTML) + public FeaturesFile(boolean parseImmediately, String inFile, String type) + throws IOException { - return parse(align, colours, null, removeHTML, false); + super(parseImmediately, inFile, type); } /** - * Parse GFF or sequence features file optionally using case-independent - * matching, discarding URLs + * Parse GFF or sequence features file using case-independent matching, + * discarding URLs * * @param align * - alignment/dataset containing sequences that are to be annotated @@ -101,552 +135,327 @@ public class FeaturesFile extends AlignFile * - hashtable to store feature colour definitions * @param removeHTML * - process html strings into plain text - * @param relaxedIdmatching - * - when true, ID matches to compound sequence IDs are allowed * @return true if features were added */ - public boolean parse(AlignmentI align, Map colours, boolean removeHTML, - boolean relaxedIdMatching) + public boolean parse(AlignmentI align, + Map colours, boolean removeHTML) { - return parse(align, colours, null, removeHTML, relaxedIdMatching); + return parse(align, colours, removeHTML, false); } /** - * Parse GFF or sequence features file optionally using case-independent - * matching - * - * @param align - * - alignment/dataset containing sequences that are to be annotated - * @param colours - * - hashtable to store feature colour definitions - * @param featureLink - * - hashtable to store associated URLs - * @param removeHTML - * - process html strings into plain text - * @return true if features were added + * Extends the default addProperties by also adding peptide-to-cDNA mappings + * (if any) derived while parsing a GFF file */ - public boolean parse(AlignmentI align, Map colours, Map featureLink, - boolean removeHTML) + @Override + public void addProperties(AlignmentI al) { - return parse(align, colours, featureLink, removeHTML, false); + super.addProperties(al); + if (dataset != null && dataset.getCodonFrames() != null) + { + AlignmentI ds = (al.getDataset() == null) ? al : al.getDataset(); + for (AlignedCodonFrame codons : dataset.getCodonFrames()) + { + ds.addCodonFrame(codons); + } + } } /** - * Parse GFF or sequence features file + * Parse GFF or Jalview format sequence features file * * @param align * - alignment/dataset containing sequences that are to be annotated * @param colours * - hashtable to store feature colour definitions - * @param featureLink - * - hashtable to store associated URLs * @param removeHTML * - process html strings into plain text * @param relaxedIdmatching * - when true, ID matches to compound sequence IDs are allowed * @return true if features were added */ - public boolean parse(AlignmentI align, Map colours, Map featureLink, - boolean removeHTML, boolean relaxedIdmatching) + public boolean parse(AlignmentI align, + Map colours, boolean removeHTML, + boolean relaxedIdmatching) { + Map gffProps = new HashMap(); + /* + * keep track of any sequences we try to create from the data + */ + List newseqs = new ArrayList(); String line = null; try { - SequenceI seq = null; - String type, desc, token = null; - - int index, start, end; - float score; - StringTokenizer st; - SequenceFeature sf; - String featureGroup = null, groupLink = null; - Map typeLink = new Hashtable(); - /** - * when true, assume GFF style features rather than Jalview style. - */ - boolean GFFFile = true; + String[] gffColumns; + String featureGroup = null; + while ((line = nextLine()) != null) { - if (line.startsWith("#")) + // skip comments/process pragmas + if (line.length() == 0 || line.startsWith("#")) { + if (line.toLowerCase().startsWith("##")) + { + processGffPragma(line, gffProps, align, newseqs); + } continue; } - st = new StringTokenizer(line, "\t"); - if (st.countTokens() == 1) + gffColumns = line.split("\\t"); // tab as regex + if (gffColumns.length == 1) { if (line.trim().equalsIgnoreCase("GFF")) { - // Start parsing file as if it might be GFF again. - GFFFile = true; + /* + * Jalview features file with appended GFF + * assume GFF2 (though it may declare ##gff-version 3) + */ + gffVersion = 2; continue; } } - if (st.countTokens() > 1 && st.countTokens() < 4) + + if (gffColumns.length > 1 && gffColumns.length < 4) { - GFFFile = false; - type = st.nextToken(); - if (type.equalsIgnoreCase("startgroup")) + /* + * if 2 or 3 tokens, we anticipate either 'startgroup', 'endgroup' or + * a feature type colour specification + */ + String ft = gffColumns[0]; + if (ft.equalsIgnoreCase("startgroup")) { - featureGroup = st.nextToken(); - if (st.hasMoreElements()) - { - groupLink = st.nextToken(); - featureLink.put(featureGroup, groupLink); - } + featureGroup = gffColumns[1]; } - else if (type.equalsIgnoreCase("endgroup")) + else if (ft.equalsIgnoreCase("endgroup")) { // We should check whether this is the current group, - // but at present theres no way of showing more than 1 group - st.nextToken(); + // but at present there's no way of showing more than 1 group featureGroup = null; - groupLink = null; } else { - Object colour = null; - String colscheme = st.nextToken(); - if (colscheme.indexOf("|") > -1 - || colscheme.trim().equalsIgnoreCase("label")) - { - // Parse '|' separated graduated colourscheme fields: - // [label|][mincolour|maxcolour|[absolute|]minvalue|maxvalue|thresholdtype|thresholdvalue] - // can either provide 'label' only, first is optional, next two - // colors are required (but may be - // left blank), next is optional, nxt two min/max are required. - // first is either 'label' - // first/second and third are both hexadecimal or word equivalent - // colour. - // next two are values parsed as floats. - // fifth is either 'above','below', or 'none'. - // sixth is a float value and only required when fifth is either - // 'above' or 'below'. - StringTokenizer gcol = new StringTokenizer(colscheme, "|", - true); - // set defaults - int threshtype = AnnotationColourGradient.NO_THRESHOLD; - float min = Float.MIN_VALUE, max = Float.MAX_VALUE, threshval = Float.NaN; - boolean labelCol = false; - // Parse spec line - String mincol = gcol.nextToken(); - if (mincol == "|") - { - System.err - .println("Expected either 'label' or a colour specification in the line: " - + line); - continue; - } - String maxcol = null; - if (mincol.toLowerCase().indexOf("label") == 0) - { - labelCol = true; - mincol = (gcol.hasMoreTokens() ? gcol.nextToken() : null); // skip - // '|' - mincol = (gcol.hasMoreTokens() ? gcol.nextToken() : null); - } - String abso = null, minval, maxval; - if (mincol != null) - { - // at least four more tokens - if (mincol.equals("|")) - { - mincol = ""; - } - else - { - gcol.nextToken(); // skip next '|' - } - // continue parsing rest of line - maxcol = gcol.nextToken(); - if (maxcol.equals("|")) - { - maxcol = ""; - } - else - { - gcol.nextToken(); // skip next '|' - } - abso = gcol.nextToken(); - gcol.nextToken(); // skip next '|' - if (abso.toLowerCase().indexOf("abso") != 0) - { - minval = abso; - abso = null; - } - else - { - minval = gcol.nextToken(); - gcol.nextToken(); // skip next '|' - } - maxval = gcol.nextToken(); - if (gcol.hasMoreTokens()) - { - gcol.nextToken(); // skip next '|' - } - try - { - if (minval.length() > 0) - { - min = new Float(minval).floatValue(); - } - } catch (Exception e) - { - System.err - .println("Couldn't parse the minimum value for graduated colour for type (" - + colscheme - + ") - did you misspell 'auto' for the optional automatic colour switch ?"); - e.printStackTrace(); - } - try - { - if (maxval.length() > 0) - { - max = new Float(maxval).floatValue(); - } - } catch (Exception e) - { - System.err - .println("Couldn't parse the maximum value for graduated colour for type (" - + colscheme + ")"); - e.printStackTrace(); - } - } - else - { - // add in some dummy min/max colours for the label-only - // colourscheme. - mincol = "FFFFFF"; - maxcol = "000000"; - } - try - { - colour = new jalview.schemes.GraduatedColor( - new UserColourScheme(mincol).findColour('A'), - new UserColourScheme(maxcol).findColour('A'), min, - max); - } catch (Exception e) - { - System.err - .println("Couldn't parse the graduated colour scheme (" - + colscheme + ")"); - e.printStackTrace(); - } - if (colour != null) - { - ((jalview.schemes.GraduatedColor) colour) - .setColourByLabel(labelCol); - ((jalview.schemes.GraduatedColor) colour) - .setAutoScaled(abso == null); - // add in any additional parameters - String ttype = null, tval = null; - if (gcol.hasMoreTokens()) - { - // threshold type and possibly a threshold value - ttype = gcol.nextToken(); - if (ttype.toLowerCase().startsWith("below")) - { - ((jalview.schemes.GraduatedColor) colour) - .setThreshType(AnnotationColourGradient.BELOW_THRESHOLD); - } - else if (ttype.toLowerCase().startsWith("above")) - { - ((jalview.schemes.GraduatedColor) colour) - .setThreshType(AnnotationColourGradient.ABOVE_THRESHOLD); - } - else - { - ((jalview.schemes.GraduatedColor) colour) - .setThreshType(AnnotationColourGradient.NO_THRESHOLD); - if (!ttype.toLowerCase().startsWith("no")) - { - System.err - .println("Ignoring unrecognised threshold type : " - + ttype); - } - } - } - if (((GraduatedColor) colour).getThreshType() != AnnotationColourGradient.NO_THRESHOLD) - { - try - { - gcol.nextToken(); - tval = gcol.nextToken(); - ((jalview.schemes.GraduatedColor) colour) - .setThresh(new Float(tval).floatValue()); - } catch (Exception e) - { - System.err - .println("Couldn't parse threshold value as a float: (" - + tval + ")"); - e.printStackTrace(); - } - } - // parse the thresh-is-min token ? - if (gcol.hasMoreTokens()) - { - System.err - .println("Ignoring additional tokens in parameters in graduated colour specification\n"); - while (gcol.hasMoreTokens()) - { - System.err.println("|" + gcol.nextToken()); - } - System.err.println("\n"); - } - } - } - else - { - UserColourScheme ucs = new UserColourScheme(colscheme); - colour = ucs.findColour('A'); - } + String colscheme = gffColumns[1]; + FeatureColourI colour = FeatureColour + .parseJalviewFeatureColour(colscheme); if (colour != null) { - colours.put(type, colour); - } - if (st.hasMoreElements()) - { - String link = st.nextToken(); - typeLink.put(type, link); - if (featureLink == null) - { - featureLink = new Hashtable(); - } - featureLink.put(type, link); + colours.put(ft, colour); } } continue; } - String seqId = ""; - while (st.hasMoreElements()) - { - - if (GFFFile) - { - // Still possible this is an old Jalview file, - // which does not have type colours at the beginning - seqId = token = st.nextToken(); - seq = findName(align, seqId, relaxedIdmatching); - if (seq != null) - { - desc = st.nextToken(); - String group = null; - if (doGffSource && desc.indexOf(' ') == -1) - { - // could also be a source term rather than description line - group = new String(desc); - } - type = st.nextToken(); - try - { - String stt = st.nextToken(); - if (stt.length() == 0 || stt.equals("-")) - { - start = 0; - } - else - { - start = Integer.parseInt(stt); - } - } catch (NumberFormatException ex) - { - start = 0; - } - try - { - String stt = st.nextToken(); - if (stt.length() == 0 || stt.equals("-")) - { - end = 0; - } - else - { - end = Integer.parseInt(stt); - } - } catch (NumberFormatException ex) - { - end = 0; - } - // TODO: decide if non positional feature assertion for input data - // where end==0 is generally valid - if (end == 0) - { - // treat as non-positional feature, regardless. - start = 0; - } - try - { - score = new Float(st.nextToken()).floatValue(); - } catch (NumberFormatException ex) - { - score = 0; - } - sf = new SequenceFeature(type, desc, start, end, score, group); - - try - { - sf.setValue("STRAND", st.nextToken()); - sf.setValue("FRAME", st.nextToken()); - } catch (Exception ex) - { - } - - if (st.hasMoreTokens()) - { - StringBuffer attributes = new StringBuffer(); - while (st.hasMoreTokens()) - { - attributes.append("\t" + st.nextElement()); - } - // TODO validate and split GFF2 attributes field ? parse out - // ([A-Za-z][A-Za-z0-9_]*) ; and add as - // sf.setValue(attrib, val); - sf.setValue("ATTRIBUTES", attributes.toString()); - } - - seq.addSequenceFeature(sf); - while ((seq = align.findName(seq, seqId, true)) != null) - { - seq.addSequenceFeature(new SequenceFeature(sf)); - } - break; - } - } - - if (GFFFile && seq == null) - { - desc = token; - } - else - { - desc = st.nextToken(); - } - if (!st.hasMoreTokens()) - { - System.err - .println("DEBUG: Run out of tokens when trying to identify the destination for the feature.. giving up."); - // in all probability, this isn't a file we understand, so bail - // quietly. - return false; - } - - token = st.nextToken(); - - if (!token.equals("ID_NOT_SPECIFIED")) - { - seq = findName(align, seqId = token, relaxedIdmatching); - st.nextToken(); - } - else - { - seqId = null; - try - { - index = Integer.parseInt(st.nextToken()); - seq = align.getSequenceAt(index); - } catch (NumberFormatException ex) - { - seq = null; - } - } - - if (seq == null) - { - System.out.println("Sequence not found: " + line); - break; - } - - start = Integer.parseInt(st.nextToken()); - end = Integer.parseInt(st.nextToken()); - - type = st.nextToken(); - - if (!colours.containsKey(type)) - { - // Probably the old style groups file - UserColourScheme ucs = new UserColourScheme(type); - colours.put(type, ucs.findColour('A')); - } - sf = new SequenceFeature(type, desc, "", start, end, featureGroup); - if (st.hasMoreTokens()) - { - try - { - score = new Float(st.nextToken()).floatValue(); - // update colourgradient bounds if allowed to - } catch (NumberFormatException ex) - { - score = 0; - } - sf.setScore(score); - } - if (groupLink != null && removeHTML) - { - sf.addLink(groupLink); - sf.description += "%LINK%"; - } - if (typeLink.containsKey(type) && removeHTML) - { - sf.addLink(typeLink.get(type).toString()); - sf.description += "%LINK%"; - } - - parseDescriptionHTML(sf, removeHTML); - - seq.addSequenceFeature(sf); - - while (seqId != null - && (seq = align.findName(seq, seqId, false)) != null) - { - seq.addSequenceFeature(new SequenceFeature(sf)); - } - // If we got here, its not a GFFFile - GFFFile = false; + /* + * if not a comment, GFF pragma, startgroup, endgroup or feature + * colour specification, that just leaves a feature details line + * in either Jalview or GFF format + */ + if (gffVersion == 0) + { + parseJalviewFeature(line, gffColumns, align, colours, removeHTML, + relaxedIdmatching, featureGroup); + } + else + { + parseGff(gffColumns, align, relaxedIdmatching, newseqs); } } resetMatcher(); } catch (Exception ex) { + // should report somewhere useful for UI if necessary + warningMessage = ((warningMessage == null) ? "" : warningMessage) + + "Parsing error at\n" + line; System.out.println("Error parsing feature file: " + ex + "\n" + line); ex.printStackTrace(System.err); resetMatcher(); return false; } + /* + * experimental - add any dummy sequences with features to the alignment + * - we need them for Ensembl feature extraction - though maybe not otherwise + */ + for (SequenceI newseq : newseqs) + { + if (newseq.getSequenceFeatures() != null) + { + align.addSequence(newseq); + } + } return true; } - private AlignmentI lastmatchedAl = null; + /** + * Try to parse a Jalview format feature specification and add it as a + * sequence feature to any matching sequences in the alignment. Returns true + * if successful (a feature was added), or false if not. + * + * @param line + * @param gffColumns + * @param alignment + * @param featureColours + * @param removeHTML + * @param relaxedIdmatching + * @param featureGroup + */ + protected boolean parseJalviewFeature(String line, String[] gffColumns, + AlignmentI alignment, Map featureColours, + boolean removeHTML, boolean relaxedIdMatching, String featureGroup) + { + /* + * tokens: description seqid seqIndex start end type [score] + */ + if (gffColumns.length < 6) + { + System.err.println("Ignoring feature line '" + line + + "' with too few columns (" + gffColumns.length + ")"); + return false; + } + String desc = gffColumns[0]; + String seqId = gffColumns[1]; + SequenceI seq = findSequence(seqId, alignment, null, relaxedIdMatching); - private SequenceIdMatcher matcher = null; + if (!ID_NOT_SPECIFIED.equals(seqId)) + { + seq = findSequence(seqId, alignment, null, relaxedIdMatching); + } + else + { + seqId = null; + seq = null; + String seqIndex = gffColumns[2]; + try + { + int idx = Integer.parseInt(seqIndex); + seq = alignment.getSequenceAt(idx); + } catch (NumberFormatException ex) + { + System.err.println("Invalid sequence index: " + seqIndex); + } + } + + if (seq == null) + { + System.out.println("Sequence not found: " + line); + return false; + } + + int startPos = Integer.parseInt(gffColumns[3]); + int endPos = Integer.parseInt(gffColumns[4]); + + String ft = gffColumns[5]; + + if (!featureColours.containsKey(ft)) + { + /* + * Perhaps an old style groups file with no colours - + * synthesize a colour from the feature type + */ + UserColourScheme ucs = new UserColourScheme(ft); + featureColours.put(ft, new FeatureColour(ucs.findColour('A'))); + } + SequenceFeature sf = new SequenceFeature(ft, desc, "", startPos, + endPos, featureGroup); + if (gffColumns.length > 6) + { + float score = Float.NaN; + try + { + score = new Float(gffColumns[6]).floatValue(); + // update colourgradient bounds if allowed to + } catch (NumberFormatException ex) + { + // leave as NaN + } + sf.setScore(score); + } + + parseDescriptionHTML(sf, removeHTML); + + seq.addSequenceFeature(sf); + + while (seqId != null + && (seq = alignment.findName(seq, seqId, false)) != null) + { + seq.addSequenceFeature(new SequenceFeature(sf)); + } + return true; + } /** * clear any temporary handles used to speed up ID matching */ - private void resetMatcher() + protected void resetMatcher() { lastmatchedAl = null; matcher = null; } - private SequenceI findName(AlignmentI align, String seqId, - boolean relaxedIdMatching) + /** + * Returns a sequence matching the given id, as follows + *
    + *
  • strict matching is on exact sequence name
    • + *
  • relaxed matching allows matching on a token within the sequence name, + * or a dbxref
    • + *
  • first tries to find a match in the alignment sequences
    • + *
  • else tries to find a match in the new sequences already generated while + * parsing the features file
    • + *
  • else creates a new placeholder sequence, adds it to the new sequences + * list, and returns it
    • + *
+ * + * @param seqId + * @param align + * @param newseqs + * @param relaxedIdMatching + * + * @return + */ + protected SequenceI findSequence(String seqId, AlignmentI align, + List newseqs, boolean relaxedIdMatching) { + // TODO encapsulate in SequenceIdMatcher, share the matcher + // with the GffHelper (removing code duplication) SequenceI match = null; if (relaxedIdMatching) { if (lastmatchedAl != align) { - matcher = new SequenceIdMatcher( - (lastmatchedAl = align).getSequencesArray()); + lastmatchedAl = align; + matcher = new SequenceIdMatcher(align.getSequencesArray()); + if (newseqs != null) + { + matcher.addAll(newseqs); + } } match = matcher.findIdMatch(seqId); } else { match = align.findName(seqId, true); + if (match == null && newseqs != null) + { + for (SequenceI m : newseqs) + { + if (seqId.equals(m.getName())) + { + return m; + } + } + } + + } + if (match == null && newseqs != null) + { + match = new SequenceDummy(seqId); + if (relaxedIdMatching) + { + matcher.addAll(Arrays.asList(new SequenceI[] { match })); + } + // add dummy sequence to the newseqs list + newseqs.add(match); } return match; } @@ -657,7 +466,7 @@ public class FeaturesFile extends AlignFile { return; } - jalview.util.ParseHtmlBodyAndLinks parsed = new jalview.util.ParseHtmlBodyAndLinks( + ParseHtmlBodyAndLinks parsed = new ParseHtmlBodyAndLinks( sf.getDescription(), removeHTML, newline); sf.description = (removeHTML) ? parsed.getNonHtmlContent() @@ -672,21 +481,22 @@ public class FeaturesFile extends AlignFile /** * generate a features file for seqs includes non-pos features by default. * - * @param seqs + * @param sequences * source of sequence features * @param visible * hash of feature types and colours * @return features file contents */ - public String printJalviewFormat(SequenceI[] seqs, Map visible) + public String printJalviewFormat(SequenceI[] sequences, + Map visible) { - return printJalviewFormat(seqs, visible, true, true); + return printJalviewFormat(sequences, visible, true, true); } /** * generate a features file for seqs with colours from visible (if any) * - * @param seqs + * @param sequences * source of features * @param visible * hash of Colours for each feature type @@ -697,11 +507,11 @@ public class FeaturesFile extends AlignFile * of group or type) * @return features file contents */ - public String printJalviewFormat(SequenceI[] seqs, Map visible, - boolean visOnly, boolean nonpos) + public String printJalviewFormat(SequenceI[] sequences, + Map visible, boolean visOnly, + boolean nonpos) { - StringBuffer out = new StringBuffer(); - SequenceFeature[] next; + StringBuilder out = new StringBuilder(256); boolean featuresGen = false; if (visOnly && !nonpos && (visible == null || visible.size() < 1)) { @@ -714,83 +524,40 @@ public class FeaturesFile extends AlignFile // write feature colours only if we're given them and we are generating // viewed features // TODO: decide if feature links should also be written here ? - Iterator en = visible.keySet().iterator(); - String type, color; + Iterator en = visible.keySet().iterator(); while (en.hasNext()) { - type = en.next().toString(); - - if (visible.get(type) instanceof GraduatedColor) - { - GraduatedColor gc = (GraduatedColor) visible.get(type); - color = (gc.isColourByLabel() ? "label|" : "") - + Format.getHexString(gc.getMinColor()) + "|" - + Format.getHexString(gc.getMaxColor()) - + (gc.isAutoScale() ? "|" : "|abso|") + gc.getMin() + "|" - + gc.getMax() + "|"; - if (gc.getThreshType() != AnnotationColourGradient.NO_THRESHOLD) - { - if (gc.getThreshType() == AnnotationColourGradient.BELOW_THRESHOLD) - { - color += "below"; - } - else - { - if (gc.getThreshType() != AnnotationColourGradient.ABOVE_THRESHOLD) - { - System.err.println("WARNING: Unsupported threshold type (" - + gc.getThreshType() + ") : Assuming 'above'"); - } - color += "above"; - } - // add the value - color += "|" + gc.getThresh(); - } - else - { - color += "none"; - } - } - else if (visible.get(type) instanceof java.awt.Color) - { - color = Format.getHexString((java.awt.Color) visible.get(type)); - } - else - { - // legacy support for integer objects containing colour triplet values - color = Format.getHexString(new java.awt.Color(Integer - .parseInt(visible.get(type).toString()))); - } - out.append(type); - out.append("\t"); - out.append(color); - out.append(newline); + String featureType = en.next().toString(); + FeatureColourI colour = visible.get(featureType); + out.append(colour.toJalviewFormat(featureType)).append(newline); } } + // Work out which groups are both present and visible - Vector groups = new Vector(); + List groups = new ArrayList(); int groupIndex = 0; boolean isnonpos = false; - for (int i = 0; i < seqs.length; i++) + SequenceFeature[] features; + for (int i = 0; i < sequences.length; i++) { - next = seqs[i].getSequenceFeatures(); - if (next != null) + features = sequences[i].getSequenceFeatures(); + if (features != null) { - for (int j = 0; j < next.length; j++) + for (int j = 0; j < features.length; j++) { - isnonpos = next[j].begin == 0 && next[j].end == 0; + isnonpos = features[j].begin == 0 && features[j].end == 0; if ((!nonpos && isnonpos) || (!isnonpos && visOnly && !visible - .containsKey(next[j].type))) + .containsKey(features[j].type))) { continue; } - if (next[j].featureGroup != null - && !groups.contains(next[j].featureGroup)) + if (features[j].featureGroup != null + && !groups.contains(features[j].featureGroup)) { - groups.addElement(next[j].featureGroup); + groups.add(features[j].featureGroup); } } } @@ -799,12 +566,11 @@ public class FeaturesFile extends AlignFile String group = null; do { - if (groups.size() > 0 && groupIndex < groups.size()) { - group = groups.elementAt(groupIndex).toString(); + group = groups.get(groupIndex); out.append(newline); - out.append("STARTGROUP\t"); + out.append("STARTGROUP").append(TAB); out.append(group); out.append(newline); } @@ -813,17 +579,18 @@ public class FeaturesFile extends AlignFile group = null; } - for (int i = 0; i < seqs.length; i++) + for (int i = 0; i < sequences.length; i++) { - next = seqs[i].getSequenceFeatures(); - if (next != null) + features = sequences[i].getSequenceFeatures(); + if (features != null) { - for (int j = 0; j < next.length; j++) + for (SequenceFeature sequenceFeature : features) { - isnonpos = next[j].begin == 0 && next[j].end == 0; + isnonpos = sequenceFeature.begin == 0 + && sequenceFeature.end == 0; if ((!nonpos && isnonpos) || (!isnonpos && visOnly && !visible - .containsKey(next[j].type))) + .containsKey(sequenceFeature.type))) { // skip if feature is nonpos and we ignore them or if we only // output visible and it isn't non-pos and it's not visible @@ -831,65 +598,66 @@ public class FeaturesFile extends AlignFile } if (group != null - && (next[j].featureGroup == null || !next[j].featureGroup + && (sequenceFeature.featureGroup == null || !sequenceFeature.featureGroup .equals(group))) { continue; } - if (group == null && next[j].featureGroup != null) + if (group == null && sequenceFeature.featureGroup != null) { continue; } // we have features to output featuresGen = true; - if (next[j].description == null - || next[j].description.equals("")) + if (sequenceFeature.description == null + || sequenceFeature.description.equals("")) { - out.append(next[j].type + "\t"); + out.append(sequenceFeature.type).append(TAB); } else { - if (next[j].links != null - && next[j].getDescription().indexOf("") == -1) + if (sequenceFeature.links != null + && sequenceFeature.getDescription().indexOf("") == -1) { out.append(""); } - out.append(next[j].description + " "); - if (next[j].links != null) + out.append(sequenceFeature.description); + if (sequenceFeature.links != null) { - for (int l = 0; l < next[j].links.size(); l++) + for (int l = 0; l < sequenceFeature.links.size(); l++) { - String label = next[j].links.elementAt(l).toString(); + String label = sequenceFeature.links.elementAt(l); String href = label.substring(label.indexOf("|") + 1); label = label.substring(0, label.indexOf("|")); - if (next[j].description.indexOf(href) == -1) + if (sequenceFeature.description.indexOf(href) == -1) { - out.append("" + label + ""); + out.append(" " + label + + ""); } } - if (next[j].getDescription().indexOf("") == -1) + if (sequenceFeature.getDescription().indexOf("") == -1) { out.append(""); } } - out.append("\t"); + out.append(TAB); } - out.append(seqs[i].getName()); + out.append(sequences[i].getName()); out.append("\t-1\t"); - out.append(next[j].begin); - out.append("\t"); - out.append(next[j].end); - out.append("\t"); - out.append(next[j].type); - if (next[j].score != Float.NaN) + out.append(sequenceFeature.begin); + out.append(TAB); + out.append(sequenceFeature.end); + out.append(TAB); + out.append(sequenceFeature.type); + if (!Float.isNaN(sequenceFeature.score)) { - out.append("\t"); - out.append(next[j].score); + out.append(TAB); + out.append(sequenceFeature.score); } out.append(newline); } @@ -898,7 +666,7 @@ public class FeaturesFile extends AlignFile if (group != null) { - out.append("ENDGROUP\t"); + out.append("ENDGROUP").append(TAB); out.append(group); out.append(newline); groupIndex++; @@ -919,110 +687,493 @@ public class FeaturesFile extends AlignFile } /** - * generate a gff file for sequence features includes non-pos features by - * default. + * Parse method that is called when a GFF file is dragged to the desktop + */ + @Override + public void parse() + { + AlignViewportI av = getViewport(); + if (av != null) + { + if (av.getAlignment() != null) + { + dataset = av.getAlignment().getDataset(); + } + if (dataset == null) + { + // working in the applet context ? + dataset = av.getAlignment(); + } + } + else + { + dataset = new Alignment(new SequenceI[] {}); + } + + boolean parseResult = parse(dataset, null, false, true); + if (!parseResult) + { + // pass error up somehow + } + if (av != null) + { + // update viewport with the dataset data ? + } + else + { + setSeqs(dataset.getSequencesArray()); + } + } + + /** + * Implementation of unused abstract method * - * @param seqs + * @return error message + */ + @Override + public String print() + { + return "Use printGffFormat() or printJalviewFormat()"; + } + + /** + * Returns features output in GFF2 format, including hidden and non-positional + * features + * + * @param sequences + * the sequences whose features are to be output * @param visible + * a map whose keys are the type names of visible features * @return */ - public String printGFFFormat(SequenceI[] seqs, Map visible) + public String printGffFormat(SequenceI[] sequences, + Map visible) { - return printGFFFormat(seqs, visible, true, true); + return printGffFormat(sequences, visible, true, true); } - public String printGFFFormat(SequenceI[] seqs, Map visible, - boolean visOnly, boolean nonpos) + /** + * Returns features output in GFF2 format + * + * @param sequences + * the sequences whose features are to be output + * @param visible + * a map whose keys are the type names of visible features + * @param outputVisibleOnly + * @param includeNonPositionalFeatures + * @return + */ + public String printGffFormat(SequenceI[] sequences, + Map visible, boolean outputVisibleOnly, + boolean includeNonPositionalFeatures) { - StringBuffer out = new StringBuffer(); - SequenceFeature[] next; + StringBuilder out = new StringBuilder(256); + int version = gffVersion == 0 ? 2 : gffVersion; + out.append(String.format("%s %d\n", GFF_VERSION, version)); String source; boolean isnonpos; - for (int i = 0; i < seqs.length; i++) + for (SequenceI seq : sequences) { - if (seqs[i].getSequenceFeatures() != null) + SequenceFeature[] features = seq.getSequenceFeatures(); + if (features != null) { - next = seqs[i].getSequenceFeatures(); - for (int j = 0; j < next.length; j++) + for (SequenceFeature sf : features) { - isnonpos = next[j].begin == 0 && next[j].end == 0; - if ((!nonpos && isnonpos) - || (!isnonpos && visOnly && !visible - .containsKey(next[j].type))) + isnonpos = sf.begin == 0 && sf.end == 0; + if (!includeNonPositionalFeatures && isnonpos) + { + /* + * ignore non-positional features if not wanted + */ + continue; + } + // TODO why the test !isnonpos here? + // what about not visible non-positional features? + if (!isnonpos && outputVisibleOnly + && !visible.containsKey(sf.type)) { + /* + * ignore not visible features if not wanted + */ continue; } - source = next[j].featureGroup; + source = sf.featureGroup; if (source == null) { - source = next[j].getDescription(); + source = sf.getDescription(); } - out.append(seqs[i].getName()); - out.append("\t"); + out.append(seq.getName()); + out.append(TAB); out.append(source); - out.append("\t"); - out.append(next[j].type); - out.append("\t"); - out.append(next[j].begin); - out.append("\t"); - out.append(next[j].end); - out.append("\t"); - out.append(next[j].score); - out.append("\t"); - - if (next[j].getValue("STRAND") != null) - { - out.append(next[j].getValue("STRAND")); - out.append("\t"); - } - else + out.append(TAB); + out.append(sf.type); + out.append(TAB); + out.append(sf.begin); + out.append(TAB); + out.append(sf.end); + out.append(TAB); + out.append(sf.score); + out.append(TAB); + + int strand = sf.getStrand(); + out.append(strand == 1 ? "+" : (strand == -1 ? "-" : ".")); + out.append(TAB); + + String phase = sf.getPhase(); + out.append(phase == null ? "." : phase); + + // miscellaneous key-values (GFF column 9) + String attributes = sf.getAttributes(); + if (attributes != null) { - out.append(".\t"); + out.append(TAB).append(attributes); } - if (next[j].getValue("FRAME") != null) - { - out.append(next[j].getValue("FRAME")); - } - else - { - out.append("."); - } - // TODO: verify/check GFF - should there be a /t here before attribute - // output ? + out.append(newline); + } + } + } - if (next[j].getValue("ATTRIBUTES") != null) - { - out.append(next[j].getValue("ATTRIBUTES")); - } + return out.toString(); + } - out.append(newline); + /** + * Returns a mapping given list of one or more Align descriptors (exonerate + * format) + * + * @param alignedRegions + * a list of "Align fromStart toStart fromCount" + * @param mapIsFromCdna + * if true, 'from' is dna, else 'from' is protein + * @param strand + * either 1 (forward) or -1 (reverse) + * @return + * @throws IOException + */ + protected MapList constructCodonMappingFromAlign( + List alignedRegions, boolean mapIsFromCdna, int strand) + throws IOException + { + if (strand == 0) + { + throw new IOException( + "Invalid strand for a codon mapping (cannot be 0)"); + } + int regions = alignedRegions.size(); + // arrays to hold [start, end] for each aligned region + int[] fromRanges = new int[regions * 2]; // from dna + int[] toRanges = new int[regions * 2]; // to protein + int fromRangesIndex = 0; + int toRangesIndex = 0; + + for (String range : alignedRegions) + { + /* + * Align mapFromStart mapToStart mapFromCount + * e.g. if mapIsFromCdna + * Align 11270 143 120 + * means: + * 120 bases from pos 11270 align to pos 143 in peptide + * if !mapIsFromCdna this would instead be + * Align 143 11270 40 + */ + String[] tokens = range.split(" "); + if (tokens.length != 3) + { + throw new IOException("Wrong number of fields for Align"); + } + int fromStart = 0; + int toStart = 0; + int fromCount = 0; + try + { + fromStart = Integer.parseInt(tokens[0]); + toStart = Integer.parseInt(tokens[1]); + fromCount = Integer.parseInt(tokens[2]); + } catch (NumberFormatException nfe) + { + throw new IOException("Invalid number in Align field: " + + nfe.getMessage()); + } + + /* + * Jalview always models from dna to protein, so adjust values if the + * GFF mapping is from protein to dna + */ + if (!mapIsFromCdna) + { + fromCount *= 3; + int temp = fromStart; + fromStart = toStart; + toStart = temp; + } + fromRanges[fromRangesIndex++] = fromStart; + fromRanges[fromRangesIndex++] = fromStart + strand * (fromCount - 1); + + /* + * If a codon has an intron gap, there will be contiguous 'toRanges'; + * this is handled for us by the MapList constructor. + * (It is not clear that exonerate ever generates this case) + */ + toRanges[toRangesIndex++] = toStart; + toRanges[toRangesIndex++] = toStart + (fromCount - 1) / 3; + } + return new MapList(fromRanges, toRanges, 3, 1); + } + + /** + * Parse a GFF format feature. This may include creating a 'dummy' sequence to + * hold the feature, or for its mapped sequence, or both, to be resolved + * either later in the GFF file (##FASTA section), or when the user loads + * additional sequences. + * + * @param gffColumns + * @param alignment + * @param relaxedIdMatching + * @param newseqs + * @return + */ + protected SequenceI parseGff(String[] gffColumns, AlignmentI alignment, + boolean relaxedIdMatching, List newseqs) + { + /* + * GFF: seqid source type start end score strand phase [attributes] + */ + if (gffColumns.length < 5) + { + System.err.println("Ignoring GFF feature line with too few columns (" + + gffColumns.length + ")"); + return null; + } + + /* + * locate referenced sequence in alignment _or_ + * as a forward or external reference (SequenceDummy) + */ + String seqId = gffColumns[0]; + SequenceI seq = findSequence(seqId, alignment, newseqs, + relaxedIdMatching); + + SequenceFeature sf = null; + GffHelperI helper = GffHelperFactory.getHelper(gffColumns); + if (helper != null) + { + try + { + sf = helper.processGff(seq, gffColumns, alignment, newseqs, + relaxedIdMatching); + if (sf != null) + { + seq.addSequenceFeature(sf); + while ((seq = alignment.findName(seq, seqId, true)) != null) + { + seq.addSequenceFeature(new SequenceFeature(sf)); + } } + } catch (IOException e) + { + System.err.println("GFF parsing failed with: " + e.getMessage()); + return null; } } - return out.toString(); + return seq; } /** - * this is only for the benefit of object polymorphism - method does nothing. + * Process the 'column 9' data of the GFF file. This is less formally defined, + * and its interpretation will vary depending on the tool that has generated + * it. + * + * @param attributes + * @param sf */ - public void parse() + protected void processGffColumnNine(String attributes, SequenceFeature sf) { - // IGNORED + sf.setAttributes(attributes); + + /* + * Parse attributes in column 9 and add them to the sequence feature's + * 'otherData' table; use Note as a best proxy for description + */ + char nameValueSeparator = gffVersion == 3 ? '=' : ' '; + // TODO check we don't break GFF2 values which include commas here + Map> nameValues = GffHelperBase + .parseNameValuePairs(attributes, ";", nameValueSeparator, ","); + for (Entry> attr : nameValues.entrySet()) + { + String values = StringUtils.listToDelimitedString(attr.getValue(), + "; "); + sf.setValue(attr.getKey(), values); + if (NOTE.equals(attr.getKey())) + { + sf.setDescription(values); + } + } } /** - * this is only for the benefit of object polymorphism - method does nothing. + * After encountering ##fasta in a GFF3 file, process the remainder of the + * file as FAST sequence data. Any placeholder sequences created during + * feature parsing are updated with the actual sequences. * - * @return error message + * @param align + * @param newseqs + * @throws IOException */ - public String print() + protected void processAsFasta(AlignmentI align, List newseqs) + throws IOException { - return "USE printGFFFormat() or printJalviewFormat()"; + try + { + mark(); + } catch (IOException q) + { + } + FastaFile parser = new FastaFile(this); + List includedseqs = parser.getSeqs(); + + SequenceIdMatcher smatcher = new SequenceIdMatcher(newseqs); + + /* + * iterate over includedseqs, and replacing matching ones with newseqs + * sequences. Generic iterator not used here because we modify + * includedseqs as we go + */ + for (int p = 0, pSize = includedseqs.size(); p < pSize; p++) + { + // search for any dummy seqs that this sequence can be used to update + SequenceI includedSeq = includedseqs.get(p); + SequenceI dummyseq = smatcher.findIdMatch(includedSeq); + if (dummyseq != null && dummyseq instanceof SequenceDummy) + { + // probably have the pattern wrong + // idea is that a flyweight proxy for a sequence ID can be created for + // 1. stable reference creation + // 2. addition of annotation + // 3. future replacement by a real sequence + // current pattern is to create SequenceDummy objects - a convenience + // constructor for a Sequence. + // problem is that when promoted to a real sequence, all references + // need to be updated somehow. We avoid that by keeping the same object. + ((SequenceDummy) dummyseq).become(includedSeq); + dummyseq.createDatasetSequence(); + + /* + * Update mappings so they are now to the dataset sequence + */ + for (AlignedCodonFrame mapping : align.getCodonFrames()) + { + mapping.updateToDataset(dummyseq); + } + + /* + * replace parsed sequence with the realised forward reference + */ + includedseqs.set(p, dummyseq); + + /* + * and remove from the newseqs list + */ + newseqs.remove(dummyseq); + } + } + + /* + * finally add sequences to the dataset + */ + for (SequenceI seq : includedseqs) + { + // experimental: mapping-based 'alignment' to query sequence + AlignmentUtils.alignSequenceAs(seq, align, + String.valueOf(align.getGapCharacter()), false, true); + + // rename sequences if GFF handler requested this + // TODO a more elegant way e.g. gffHelper.postProcess(newseqs) ? + SequenceFeature[] sfs = seq.getSequenceFeatures(); + if (sfs != null) + { + String newName = (String) sfs[0].getValue(GffHelperI.RENAME_TOKEN); + if (newName != null) + { + seq.setName(newName); + } + } + align.addSequence(seq); + } } + /** + * Process a ## directive + * + * @param line + * @param gffProps + * @param align + * @param newseqs + * @throws IOException + */ + protected void processGffPragma(String line, + Map gffProps, AlignmentI align, + List newseqs) throws IOException + { + line = line.trim(); + if ("###".equals(line)) + { + // close off any open 'forward references' + return; + } + + String[] tokens = line.substring(2).split(" "); + String pragma = tokens[0]; + String value = tokens.length == 1 ? null : tokens[1]; + + if ("gff-version".equalsIgnoreCase(pragma)) + { + if (value != null) + { + try + { + // value may be e.g. "3.1.2" + gffVersion = Integer.parseInt(value.split("\\.")[0]); + } catch (NumberFormatException e) + { + // ignore + } + } + } + else if ("sequence-region".equalsIgnoreCase(pragma)) + { + // could capture if wanted here + } + else if ("feature-ontology".equalsIgnoreCase(pragma)) + { + // should resolve against the specified feature ontology URI + } + else if ("attribute-ontology".equalsIgnoreCase(pragma)) + { + // URI of attribute ontology - not currently used in GFF3 + } + else if ("source-ontology".equalsIgnoreCase(pragma)) + { + // URI of source ontology - not currently used in GFF3 + } + else if ("species-build".equalsIgnoreCase(pragma)) + { + // save URI of specific NCBI taxon version of annotations + gffProps.put("species-build", value); + } + else if ("fasta".equalsIgnoreCase(pragma)) + { + // process the rest of the file as a fasta file and replace any dummy + // sequence IDs + processAsFasta(align, newseqs); + } + else + { + System.err.println("Ignoring unknown pragma: " + line); + } + } }