X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fio%2FFeaturesFile.java;h=fcf1d707c7f17ed3b1d43c800e42e325923f7ae7;hb=9c007fa50a82b3ae0f987e6454ae17c941b63563;hp=ee6ba11a18cdd9615055a351cf8211992330eed1;hpb=26ba864a6c290121fe6cf616794d2d0bea65fb7d;p=jalview.git diff --git a/src/jalview/io/FeaturesFile.java b/src/jalview/io/FeaturesFile.java index ee6ba11..fcf1d70 100755 --- a/src/jalview/io/FeaturesFile.java +++ b/src/jalview/io/FeaturesFile.java @@ -20,18 +20,22 @@ */ package jalview.io; +import jalview.analysis.AlignmentUtils; import jalview.analysis.SequenceIdMatcher; import jalview.api.AlignViewportI; +import jalview.api.FeatureColourI; +import jalview.api.FeaturesSourceI; import jalview.datamodel.AlignedCodonFrame; import jalview.datamodel.Alignment; import jalview.datamodel.AlignmentI; import jalview.datamodel.SequenceDummy; import jalview.datamodel.SequenceFeature; import jalview.datamodel.SequenceI; -import jalview.schemes.AnnotationColourGradient; -import jalview.schemes.GraduatedColor; -import jalview.schemes.UserColourScheme; -import jalview.util.Format; +import jalview.io.gff.GffHelperBase; +import jalview.io.gff.GffHelperFactory; +import jalview.io.gff.GffHelperI; +import jalview.schemes.FeatureColour; +import jalview.util.ColorUtils; import jalview.util.MapList; import jalview.util.ParseHtmlBodyAndLinks; import jalview.util.StringUtils; @@ -45,7 +49,6 @@ import java.util.Iterator; import java.util.List; import java.util.Map; import java.util.Map.Entry; -import java.util.StringTokenizer; /** * Parses and writes features files, which may be in Jalview, GFF2 or GFF3 @@ -63,13 +66,11 @@ import java.util.StringTokenizer; * @author jbprocter * @author gmcarstairs */ -public class FeaturesFile extends AlignFile +public class FeaturesFile extends AlignFile implements FeaturesSourceI { - protected static final String STRAND = "STRAND"; + private static final String ID_NOT_SPECIFIED = "ID_NOT_SPECIFIED"; - protected static final String FRAME = "FRAME"; - - protected static final String ATTRIBUTES = "ATTRIBUTES"; + private static final String NOTE = "Note"; protected static final String TAB = "\t"; @@ -94,12 +95,13 @@ public class FeaturesFile extends AlignFile * Constructor which does not parse the file immediately * * @param inFile - * @param type + * @param paste * @throws IOException */ - public FeaturesFile(String inFile, String type) throws IOException + public FeaturesFile(String inFile, DataSourceType paste) + throws IOException { - super(false, inFile, type); + super(false, inFile, paste); } /** @@ -119,7 +121,8 @@ public class FeaturesFile extends AlignFile * @param type * @throws IOException */ - public FeaturesFile(boolean parseImmediately, String inFile, String type) + public FeaturesFile(boolean parseImmediately, String inFile, + DataSourceType type) throws IOException { super(parseImmediately, inFile, type); @@ -137,8 +140,8 @@ public class FeaturesFile extends AlignFile * - process html strings into plain text * @return true if features were added */ - public boolean parse(AlignmentI align, Map colours, - boolean removeHTML) + public boolean parse(AlignmentI align, + Map colours, boolean removeHTML) { return parse(align, colours, removeHTML, false); } @@ -174,8 +177,9 @@ public class FeaturesFile extends AlignFile * - 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, + boolean relaxedIdmatching) { Map gffProps = new HashMap(); /* @@ -186,7 +190,7 @@ public class FeaturesFile extends AlignFile String line = null; try { - StringTokenizer st; + String[] gffColumns; String featureGroup = null; while ((line = nextLine()) != null) @@ -201,41 +205,46 @@ public class FeaturesFile extends AlignFile continue; } - st = new StringTokenizer(line, TAB); - if (st.countTokens() == 1) + gffColumns = line.split("\\t"); // tab as regex + if (gffColumns.length == 1) { if (line.trim().equalsIgnoreCase("GFF")) { /* - * Jalview features file with appendded GFF - * assume GFF2 (though it may declare gff-version 3) + * 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) { /* * if 2 or 3 tokens, we anticipate either 'startgroup', 'endgroup' or - * a feature type colour specification; not GFF format + * a feature type colour specification */ - String ft = st.nextToken(); + String ft = gffColumns[0]; if (ft.equalsIgnoreCase("startgroup")) { - featureGroup = st.nextToken(); + featureGroup = gffColumns[1]; } 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; } else { - parseFeatureColour(line, ft, st, colours); + String colscheme = gffColumns[1]; + FeatureColourI colour = FeatureColour + .parseJalviewFeatureColour(colscheme); + if (colour != null) + { + colours.put(ft, colour); + } } continue; } @@ -247,12 +256,12 @@ public class FeaturesFile extends AlignFile */ if (gffVersion == 0) { - parseJalviewFeature(line, st, align, colours, removeHTML, + parseJalviewFeature(line, gffColumns, align, colours, removeHTML, relaxedIdmatching, featureGroup); } else { - parseGffFeature(st, align, relaxedIdmatching, newseqs); + parseGff(gffColumns, align, relaxedIdmatching, newseqs); } } resetMatcher(); @@ -267,56 +276,66 @@ public class FeaturesFile extends AlignFile return false; } + /* + * experimental - add any dummy sequences with features to the alignment + * - we need them for Ensembl feature extraction - though maybe not otherwise + */ + for (SequenceI newseq : newseqs) + { + if (newseq.getFeatures().hasFeatures()) + { + align.addSequence(newseq); + } + } return true; } /** - * Try to parse a Jalview format feature specification. Returns true if - * successful or false if not. + * 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 st + * @param gffColumns * @param alignment * @param featureColours * @param removeHTML * @param relaxedIdmatching * @param featureGroup */ - protected boolean parseJalviewFeature(String line, StringTokenizer st, - AlignmentI alignment, Map featureColours, - boolean removeHTML, boolean relaxedIdmatching, String featureGroup) + protected boolean parseJalviewFeature(String line, String[] gffColumns, + AlignmentI alignment, Map featureColours, + boolean removeHTML, boolean relaxedIdMatching, String featureGroup) { /* - * Jalview: description seqid seqIndex start end type [score] + * tokens: description seqid seqIndex start end type [score] */ - String desc = st.nextToken(); - String seqId = st.nextToken(); - SequenceI seq = findName(alignment, seqId, relaxedIdmatching, null); - if (!st.hasMoreTokens()) + if (gffColumns.length < 6) { - 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. + 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); - if (!seqId.equals("ID_NOT_SPECIFIED")) + if (!ID_NOT_SPECIFIED.equals(seqId)) { - seq = findName(alignment, seqId, relaxedIdmatching, null); - st.nextToken(); + seq = findSequence(seqId, alignment, null, relaxedIdMatching); } else { seqId = null; seq = null; + String seqIndex = gffColumns[2]; try { - int idx = Integer.parseInt(st.nextToken()); + int idx = Integer.parseInt(seqIndex); seq = alignment.getSequenceAt(idx); } catch (NumberFormatException ex) { - // continue + System.err.println("Invalid sequence index: " + seqIndex); } } @@ -326,10 +345,10 @@ public class FeaturesFile extends AlignFile return false; } - int startPos = Integer.parseInt(st.nextToken()); - int endPos = Integer.parseInt(st.nextToken()); + int startPos = Integer.parseInt(gffColumns[3]); + int endPos = Integer.parseInt(gffColumns[4]); - String ft = st.nextToken(); + String ft = gffColumns[5]; if (!featureColours.containsKey(ft)) { @@ -337,21 +356,21 @@ public class FeaturesFile extends AlignFile * Perhaps an old style groups file with no colours - * synthesize a colour from the feature type */ - UserColourScheme ucs = new UserColourScheme(ft); - featureColours.put(ft, ucs.findColour('A')); + Color colour = ColorUtils.createColourFromName(ft); + featureColours.put(ft, new FeatureColour(colour)); } - SequenceFeature sf = new SequenceFeature(ft, desc, "", - startPos, endPos, featureGroup); - if (st.hasMoreTokens()) + SequenceFeature sf = new SequenceFeature(ft, desc, "", startPos, + endPos, featureGroup); + if (gffColumns.length > 6) { - float score = 0f; + float score = Float.NaN; try { - score = new Float(st.nextToken()).floatValue(); + score = new Float(gffColumns[6]).floatValue(); // update colourgradient bounds if allowed to } catch (NumberFormatException ex) { - // leave as 0 + // leave as NaN } sf.setScore(score); } @@ -369,225 +388,6 @@ public class FeaturesFile extends AlignFile } /** - * Process a feature type colour specification - * - * @param line - * the current input line (for error messages only) - * @param featureType - * the first token on the line - * @param st - * holds remaining tokens on the line - * @param colours - * map to which to add derived colour specification - */ - protected void parseFeatureColour(String line, String featureType, - StringTokenizer st, Map colours) - { - Object colour = null; - String colscheme = st.nextToken(); - if (colscheme.indexOf("|") > -1 - || colscheme.trim().equalsIgnoreCase("label")) - { - colour = parseGraduatedColourScheme(line, colscheme); - } - else - { - UserColourScheme ucs = new UserColourScheme(colscheme); - colour = ucs.findColour('A'); - } - if (colour != null) - { - colours.put(featureType, colour); - } - } - - /** - * Parse a Jalview graduated colour descriptor - * - * @param line - * @param colourDescriptor - * @return - */ - protected GraduatedColor parseGraduatedColourScheme(String line, - String colourDescriptor) - { - // Parse '|' separated graduated colourscheme fields: - // [label|][mincolour|maxcolour|[absolute|]minvalue|maxvalue|thresholdtype|thresholdvalue] - // can either provide 'label' only, first is optional, next two - // colors are required (but may be - // left blank), next is optional, nxt two min/max are required. - // first is either 'label' - // first/second and third are both hexadecimal or word equivalent - // colour. - // next two are values parsed as floats. - // fifth is either 'above','below', or 'none'. - // sixth is a float value and only required when fifth is either - // 'above' or 'below'. - StringTokenizer gcol = new StringTokenizer(colourDescriptor, "|", true); - // set defaults - float min = Float.MIN_VALUE, max = Float.MAX_VALUE; - boolean labelCol = false; - // Parse spec line - String mincol = gcol.nextToken(); - if (mincol == "|") - { - System.err - .println("Expected either 'label' or a colour specification in the line: " - + line); - return null; - } - String maxcol = null; - if (mincol.toLowerCase().indexOf("label") == 0) - { - labelCol = true; - mincol = (gcol.hasMoreTokens() ? gcol.nextToken() : null); // skip '|' - mincol = (gcol.hasMoreTokens() ? gcol.nextToken() : null); - } - String abso = null, minval, maxval; - if (mincol != null) - { - // at least four more tokens - if (mincol.equals("|")) - { - mincol = ""; - } - else - { - gcol.nextToken(); // skip next '|' - } - // continue parsing rest of line - maxcol = gcol.nextToken(); - if (maxcol.equals("|")) - { - maxcol = ""; - } - else - { - gcol.nextToken(); // skip next '|' - } - abso = gcol.nextToken(); - gcol.nextToken(); // skip next '|' - if (abso.toLowerCase().indexOf("abso") != 0) - { - minval = abso; - abso = null; - } - else - { - minval = gcol.nextToken(); - gcol.nextToken(); // skip next '|' - } - maxval = gcol.nextToken(); - if (gcol.hasMoreTokens()) - { - gcol.nextToken(); // skip next '|' - } - try - { - if (minval.length() > 0) - { - min = Float.valueOf(minval); - } - } catch (Exception e) - { - System.err - .println("Couldn't parse the minimum value for graduated colour for type (" - + colourDescriptor - + ") - did you misspell 'auto' for the optional automatic colour switch ?"); - e.printStackTrace(); - } - try - { - if (maxval.length() > 0) - { - max = Float.valueOf(maxval); - } - } catch (Exception e) - { - System.err - .println("Couldn't parse the maximum value for graduated colour for type (" - + colourDescriptor + ")"); - e.printStackTrace(); - } - } - else - { - // add in some dummy min/max colours for the label-only - // colourscheme. - mincol = "FFFFFF"; - maxcol = "000000"; - } - - GraduatedColor colour = null; - try - { - colour = new GraduatedColor( - new UserColourScheme(mincol).findColour('A'), - new UserColourScheme(maxcol).findColour('A'), min, max); - } catch (Exception e) - { - System.err.println("Couldn't parse the graduated colour scheme (" - + colourDescriptor + ")"); - e.printStackTrace(); - } - if (colour != null) - { - colour.setColourByLabel(labelCol); - colour.setAutoScaled(abso == null); - // add in any additional parameters - String ttype = null, tval = null; - if (gcol.hasMoreTokens()) - { - // threshold type and possibly a threshold value - ttype = gcol.nextToken(); - if (ttype.toLowerCase().startsWith("below")) - { - colour.setThreshType(AnnotationColourGradient.BELOW_THRESHOLD); - } - else if (ttype.toLowerCase().startsWith("above")) - { - colour.setThreshType(AnnotationColourGradient.ABOVE_THRESHOLD); - } - else - { - colour.setThreshType(AnnotationColourGradient.NO_THRESHOLD); - if (!ttype.toLowerCase().startsWith("no")) - { - System.err.println("Ignoring unrecognised threshold type : " - + ttype); - } - } - } - if (colour.getThreshType() != AnnotationColourGradient.NO_THRESHOLD) - { - try - { - gcol.nextToken(); - tval = gcol.nextToken(); - colour.setThresh(new Float(tval).floatValue()); - } catch (Exception e) - { - System.err.println("Couldn't parse threshold value as a float: (" - + tval + ")"); - e.printStackTrace(); - } - } - // parse the thresh-is-min token ? - if (gcol.hasMoreTokens()) - { - System.err - .println("Ignoring additional tokens in parameters in graduated colour specification\n"); - while (gcol.hasMoreTokens()) - { - System.err.println("|" + gcol.nextToken()); - } - System.err.println("\n"); - } - } - return colour; - } - - /** * clear any temporary handles used to speed up ID matching */ protected void resetMatcher() @@ -599,24 +399,28 @@ public class FeaturesFile extends AlignFile /** * Returns a sequence matching the given id, as follows *
    - *
  • matching is on exact sequence name, or on a token within the sequence - * name, or a dbxref, if relaxed matching is selected
    • + *
  • 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 + *
  • 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 align * @param seqId - * @param relaxedIdMatching + * @param align * @param newseqs + * @param relaxedIdMatching + * * @return */ - protected SequenceI findName(AlignmentI align, String seqId, - boolean relaxedIdMatching, List newseqs) + 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) { @@ -687,7 +491,7 @@ public class FeaturesFile extends AlignFile * @return features file contents */ public String printJalviewFormat(SequenceI[] sequences, - Map visible) + Map visible) { return printJalviewFormat(sequences, visible, true, true); } @@ -707,7 +511,8 @@ public class FeaturesFile extends AlignFile * @return features file contents */ public String printJalviewFormat(SequenceI[] sequences, - Map visible, boolean visOnly, boolean nonpos) + Map visible, boolean visOnly, + boolean nonpos) { StringBuilder out = new StringBuilder(256); boolean featuresGen = false; @@ -723,58 +528,14 @@ public class FeaturesFile extends AlignFile // viewed features // TODO: decide if feature links should also be written here ? Iterator en = visible.keySet().iterator(); - String featureType, color; while (en.hasNext()) { - featureType = en.next().toString(); - - if (visible.get(featureType) instanceof GraduatedColor) - { - GraduatedColor gc = (GraduatedColor) visible.get(featureType); - color = (gc.isColourByLabel() ? "label|" : "") - + Format.getHexString(gc.getMinColor()) + "|" - + Format.getHexString(gc.getMaxColor()) - + (gc.isAutoScale() ? "|" : "|abso|") + gc.getMin() + "|" - + gc.getMax() + "|"; - if (gc.getThreshType() != AnnotationColourGradient.NO_THRESHOLD) - { - if (gc.getThreshType() == AnnotationColourGradient.BELOW_THRESHOLD) - { - color += "below"; - } - else - { - if (gc.getThreshType() != AnnotationColourGradient.ABOVE_THRESHOLD) - { - System.err.println("WARNING: Unsupported threshold type (" - + gc.getThreshType() + ") : Assuming 'above'"); - } - color += "above"; - } - // add the value - color += "|" + gc.getThresh(); - } - else - { - color += "none"; - } - } - else if (visible.get(featureType) instanceof Color) - { - color = Format.getHexString((Color) visible.get(featureType)); - } - else - { - // legacy support for integer objects containing colour triplet values - color = Format.getHexString(new Color(Integer.parseInt(visible - .get(featureType).toString()))); - } - out.append(featureType); - out.append(TAB); - out.append(color); - out.append(newline); + 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 List groups = new ArrayList(); int groupIndex = 0; @@ -826,12 +587,13 @@ public class FeaturesFile extends AlignFile features = sequences[i].getSequenceFeatures(); if (features != null) { - for (int j = 0; j < features.length; j++) + for (SequenceFeature sequenceFeature : features) { - isnonpos = features[j].begin == 0 && features[j].end == 0; + isnonpos = sequenceFeature.begin == 0 + && sequenceFeature.end == 0; if ((!nonpos && isnonpos) || (!isnonpos && visOnly && !visible - .containsKey(features[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 @@ -839,47 +601,48 @@ public class FeaturesFile extends AlignFile } if (group != null - && (features[j].featureGroup == null || !features[j].featureGroup + && (sequenceFeature.featureGroup == null || !sequenceFeature.featureGroup .equals(group))) { continue; } - if (group == null && features[j].featureGroup != null) + if (group == null && sequenceFeature.featureGroup != null) { continue; } // we have features to output featuresGen = true; - if (features[j].description == null - || features[j].description.equals("")) + if (sequenceFeature.description == null + || sequenceFeature.description.equals("")) { - out.append(features[j].type).append(TAB); + out.append(sequenceFeature.type).append(TAB); } else { - if (features[j].links != null - && features[j].getDescription().indexOf("") == -1) + if (sequenceFeature.links != null + && sequenceFeature.getDescription().indexOf("") == -1) { out.append(""); } - out.append(features[j].description + " "); - if (features[j].links != null) + out.append(sequenceFeature.description); + if (sequenceFeature.links != null) { - for (int l = 0; l < features[j].links.size(); l++) + for (int l = 0; l < sequenceFeature.links.size(); l++) { - String label = features[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 (features[j].description.indexOf(href) == -1) + if (sequenceFeature.description.indexOf(href) == -1) { - out.append("" + label + ""); + out.append(" " + label + + ""); } } - if (features[j].getDescription().indexOf("") == -1) + if (sequenceFeature.getDescription().indexOf("") == -1) { out.append(""); } @@ -889,15 +652,15 @@ public class FeaturesFile extends AlignFile } out.append(sequences[i].getName()); out.append("\t-1\t"); - out.append(features[j].begin); + out.append(sequenceFeature.begin); out.append(TAB); - out.append(features[j].end); + out.append(sequenceFeature.end); out.append(TAB); - out.append(features[j].type); - if (!Float.isNaN(features[j].score)) + out.append(sequenceFeature.type); + if (!Float.isNaN(sequenceFeature.score)) { out.append(TAB); - out.append(features[j].score); + out.append(sequenceFeature.score); } out.append(newline); } @@ -950,7 +713,8 @@ public class FeaturesFile extends AlignFile dataset = new Alignment(new SequenceI[] {}); } - boolean parseResult = parse(dataset, null, false, true); + Map featureColours = new HashMap(); + boolean parseResult = parse(dataset, featureColours, false, true); if (!parseResult) { // pass error up somehow @@ -971,9 +735,10 @@ public class FeaturesFile extends AlignFile * @return error message */ @Override - public String print() + public String print(SequenceI[] sqs, boolean jvsuffix) { - return "Use printGffFormat() or printJalviewFormat()"; + System.out.println("Use printGffFormat() or printJalviewFormat()"); + return null; } /** @@ -986,7 +751,8 @@ public class FeaturesFile extends AlignFile * a map whose keys are the type names of visible features * @return */ - public String printGffFormat(SequenceI[] sequences, Map visible) + public String printGffFormat(SequenceI[] sequences, + Map visible) { return printGffFormat(sequences, visible, true, true); } @@ -1002,11 +768,13 @@ public class FeaturesFile extends AlignFile * @param includeNonPositionalFeatures * @return */ - public String printGffFormat(SequenceI[] sequences, Map visible, boolean outputVisibleOnly, + public String printGffFormat(SequenceI[] sequences, + Map visible, boolean outputVisibleOnly, boolean includeNonPositionalFeatures) { StringBuilder out = new StringBuilder(256); - out.append(String.format("%s %d\n", GFF_VERSION, gffVersion)); + int version = gffVersion == 0 ? 2 : gffVersion; + out.append(String.format("%s %d\n", GFF_VERSION, version)); String source; boolean isnonpos; for (SequenceI seq : sequences) @@ -1034,13 +802,13 @@ public class FeaturesFile extends AlignFile */ continue; } - + source = sf.featureGroup; if (source == null) { source = sf.getDescription(); } - + out.append(seq.getName()); out.append(TAB); out.append(source); @@ -1053,239 +821,203 @@ public class FeaturesFile extends AlignFile out.append(TAB); out.append(sf.score); out.append(TAB); - - out.append(sf.getValue(STRAND, ".")); + + int strand = sf.getStrand(); + out.append(strand == 1 ? "+" : (strand == -1 ? "-" : ".")); out.append(TAB); - - out.append(sf.getValue(FRAME, ".")); - + + String phase = sf.getPhase(); + out.append(phase == null ? "." : phase); + // miscellaneous key-values (GFF column 9) - String attributes = (String) sf.getValue(ATTRIBUTES); + String attributes = sf.getAttributes(); if (attributes != null) { out.append(TAB).append(attributes); } - + out.append(newline); } } } - + return out.toString(); } /** - * Helper method to make a mapping given a set of attributes for a GFF feature + * Returns a mapping given list of one or more Align descriptors (exonerate + * format) * - * @param set - * @param attr + * @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 InvalidGFF3FieldException + * @throws IOException */ protected MapList constructCodonMappingFromAlign( - Map> set, String attr, - int strand) throws InvalidGFF3FieldException + List alignedRegions, boolean mapIsFromCdna, int strand) + throws IOException { if (strand == 0) { - throw new InvalidGFF3FieldException(attr, set, + throw new IOException( "Invalid strand for a codon mapping (cannot be 0)"); } - List fromrange = new ArrayList(); - List torange = new ArrayList(); - int lastppos = 0, lastpframe = 0; - for (String range : set.get(attr)) + 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) { - List ints = new ArrayList(); - StringTokenizer st = new StringTokenizer(range, " "); - while (st.hasMoreTokens()) - { - String num = st.nextToken(); - try - { - ints.add(new Integer(num)); - } catch (NumberFormatException nfe) - { - throw new InvalidGFF3FieldException(attr, set, - "Invalid number in field " + num); - } - } /* - * Align positionInRef positionInQuery LengthInRef - * contig_1146 exonerate:p2g:local similarity 8534 11269 3652 - . - * alignment_id 0 ; Query DDB_G0269124 Align 11270 143 120 + * Align mapFromStart mapToStart mapFromCount + * e.g. if mapIsFromCdna + * Align 11270 143 120 * means: - * 120 bases align at pos 143 in protein to 11270 on dna (-ve strand) - * and so on for additional ' ; Align x y z' groups + * 120 bases from pos 11270 align to pos 143 in peptide + * if !mapIsFromCdna this would instead be + * Align 143 11270 40 */ - if (ints.size() != 3) + String[] tokens = range.split(" "); + if (tokens.length != 3) { - throw new InvalidGFF3FieldException(attr, set, - "Invalid number of fields for this attribute (" - + ints.size() + ")"); + throw new IOException("Wrong number of fields for Align"); } - fromrange.add(ints.get(0)); - fromrange.add(ints.get(0) + strand * ints.get(2)); - // how are intron/exon boundaries that do not align in codons - // represented - if (ints.get(1).intValue() == lastppos && lastpframe > 0) + int fromStart = 0; + int toStart = 0; + int fromCount = 0; + try { - // extend existing to map - lastppos += ints.get(2) / 3; - lastpframe = ints.get(2) % 3; - torange.set(torange.size() - 1, new Integer(lastppos)); - } - else + fromStart = Integer.parseInt(tokens[0]); + toStart = Integer.parseInt(tokens[1]); + fromCount = Integer.parseInt(tokens[2]); + } catch (NumberFormatException nfe) { - // new to map range - torange.add(ints.get(1)); - lastppos = ints.get(1) + ints.get(2) / 3; - lastpframe = ints.get(2) % 3; - torange.add(new Integer(lastppos)); + throw new IOException("Invalid number in Align field: " + + nfe.getMessage()); } - } - // from and to ranges must end up being a series of start/end intervals - if (fromrange.size() % 2 == 1) - { - throw new InvalidGFF3FieldException(attr, set, - "Couldn't parse the DNA alignment range correctly"); - } - if (torange.size() % 2 == 1) - { - throw new InvalidGFF3FieldException(attr, set, - "Couldn't parse the protein alignment range correctly"); - } - // finally, build the map - int[] frommap = new int[fromrange.size()], tomap = new int[torange - .size()]; - int p = 0; - for (Integer ip : fromrange) - { - frommap[p++] = ip.intValue(); - } - p = 0; - for (Integer ip : torange) - { - tomap[p++] = ip.intValue(); - } - - return new MapList(frommap, tomap, 3, 1); - } - private List findNames(AlignmentI align, List newseqs, boolean relaxedIdMatching, - List list) - { - List found = new ArrayList(); - for (String seqId : list) - { - SequenceI seq = findName(align, seqId, relaxedIdMatching, newseqs); - if (seq != null) + /* + * Jalview always models from dna to protein, so adjust values if the + * GFF mapping is from protein to dna + */ + if (!mapIsFromCdna) { - found.add(seq); + 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 found; + + return new MapList(fromRanges, toRanges, 3, 1); } /** - * Parse a GFF format feature. This may include creating a 'dummy' sequence - * for the feature or its mapped sequence + * 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 st + * @param gffColumns * @param alignment - * @param relaxedIdmatching + * @param relaxedIdMatching * @param newseqs * @return */ - protected SequenceI parseGffFeature(StringTokenizer st, AlignmentI alignment, boolean relaxedIdmatching, - List newseqs) + protected SequenceI parseGff(String[] gffColumns, AlignmentI alignment, + boolean relaxedIdMatching, List newseqs) { - SequenceI seq; /* * GFF: seqid source type start end score strand phase [attributes] */ - String seqId = st.nextToken(); - + 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 reference (SequenceDummy) + * as a forward or external reference (SequenceDummy) */ - seq = findName(alignment, seqId, relaxedIdmatching, newseqs); - - String desc = st.nextToken(); - String group = null; - if (desc.indexOf(' ') == -1) - { - // could also be a source term rather than description line - group = desc; - } - String ft = st.nextToken(); - int startPos = StringUtils.parseInt(st.nextToken()); - int endPos = StringUtils.parseInt(st.nextToken()); - // TODO: decide if non positional feature assertion for input data - // where end==0 is generally valid - if (endPos == 0) - { - // treat as non-positional feature, regardless. - startPos = 0; - } - float score = 0f; - try - { - score = new Float(st.nextToken()).floatValue(); - } catch (NumberFormatException ex) - { - // leave at 0 - } - - SequenceFeature sf = new SequenceFeature(ft, desc, startPos, - endPos, score, group); - if (st.hasMoreTokens()) - { - sf.setValue(STRAND, st.nextToken()); - } - if (st.hasMoreTokens()) - { - sf.setValue(FRAME, st.nextToken()); - } - - if (st.hasMoreTokens()) + String seqId = gffColumns[0]; + SequenceI seq = findSequence(seqId, alignment, newseqs, + relaxedIdMatching); + + SequenceFeature sf = null; + GffHelperI helper = GffHelperFactory.getHelper(gffColumns); + if (helper != null) { - String attributes = st.nextToken(); - sf.setValue(ATTRIBUTES, attributes); - - /* - * parse semi-structured attributes in column 9 and add them to the - * sequence feature's 'otherData' table; use Note as a best proxy for - * description - */ - Map> nameValues = StringUtils.parseNameValuePairs(attributes, ";", - new char[] { ' ', '=' }); - for (Entry> attr : nameValues.entrySet()) + try { - String values = StringUtils.listToDelimitedString(attr.getValue(), - "; "); - sf.setValue(attr.getKey(), values); - if ("Note".equals(attr.getKey())) + sf = helper.processGff(seq, gffColumns, alignment, newseqs, + relaxedIdMatching); + if (sf != null) { - sf.setDescription(values); + 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; } } - - if (processOrAddSeqFeature(alignment, newseqs, seq, sf, - relaxedIdmatching)) - { - // check whether we should add the sequence feature to any other - // sequences in the alignment with the same or similar - while ((seq = alignment.findName(seq, seqId, true)) != null) + + return seq; + } + + /** + * 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 + */ + protected void processGffColumnNine(String attributes, SequenceFeature sf) + { + 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())) { - seq.addSequenceFeature(new SequenceFeature(sf)); + sf.setDescription(values); } } - return seq; } /** @@ -1308,42 +1040,73 @@ public class FeaturesFile extends AlignFile } 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 + + /* + * 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 dummyseq = smatcher.findIdMatch(includedseqs.get(p)); - if (dummyseq != null) + SequenceI includedSeq = includedseqs.get(p); + SequenceI dummyseq = smatcher.findIdMatch(includedSeq); + if (dummyseq != null && dummyseq instanceof SequenceDummy) { - // dummyseq was created so it could be annotated and referred to in - // alignments/codon mappings - - SequenceI mseq = includedseqs.get(p); - // mseq is the 'template' imported from the FASTA file which we'll use - // to coomplete dummyseq - if (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()) { - // 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. - ((SequenceDummy) dummyseq).become(mseq); - includedseqs.set(p, dummyseq); // template is no longer needed + 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 + + /* + * 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); } } @@ -1357,7 +1120,8 @@ public class FeaturesFile extends AlignFile * @param newseqs * @throws IOException */ - protected void processGffPragma(String line, Map gffProps, AlignmentI align, + protected void processGffPragma(String line, + Map gffProps, AlignmentI align, List newseqs) throws IOException { line = line.trim(); @@ -1366,11 +1130,11 @@ public class FeaturesFile extends AlignFile // 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) @@ -1385,6 +1149,10 @@ public class FeaturesFile extends AlignFile } } } + 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 @@ -1413,108 +1181,4 @@ public class FeaturesFile extends AlignFile System.err.println("Ignoring unknown pragma: " + line); } } - - /** - * Processes the 'Query' and 'Align' properties associated with a GFF - * similarity feature; these properties define the mapping of the annotated - * feature to another from which it has transferred annotation - * - * @param set - * @param seq - * @param sf - * @return - */ - public void processGffSimilarity(Map> set, SequenceI seq, - SequenceFeature sf, AlignmentI align, List newseqs, boolean relaxedIdMatching) - throws InvalidGFF3FieldException - { - int strand = sf.getStrand(); - // exonerate cdna/protein map - // look for fields - List querySeq = findNames(align, newseqs, relaxedIdMatching, - set.get("Query")); - if (querySeq == null || querySeq.size() != 1) - { - throw new InvalidGFF3FieldException("Query", set, - "Expecting exactly one sequence in Query field (got " - + set.get("Query") + ")"); - } - if (set.containsKey("Align")) - { - // process the align maps and create cdna/protein maps - // ideally, the query sequences are in the alignment, but maybe not... - - AlignedCodonFrame alco = new AlignedCodonFrame(); - MapList codonmapping = constructCodonMappingFromAlign(set, "Align", - strand); - - // add codon mapping, and hope! - alco.addMap(seq, querySeq.get(0), codonmapping); - align.addCodonFrame(alco); - } - - } - - /** - * take a sequence feature and examine its attributes to decide how it should - * be added to a sequence - * - * @param seq - * - the destination sequence constructed or discovered in the - * current context - * @param sf - * - the base feature with ATTRIBUTES property containing any - * additional attributes - * @param gFFFile - * - true if we are processing a GFF annotation file - * @return true if sf was actually added to the sequence, false if it was - * processed in another way - */ - public boolean processOrAddSeqFeature(AlignmentI align, List newseqs, - SequenceI seq, SequenceFeature sf, boolean relaxedIdMatching) - { - String attr = (String) sf.getValue(ATTRIBUTES); - boolean addFeature = true; - if (attr != null) - { - for (String attset : attr.split(TAB)) - { - Map> set = StringUtils.parseNameValuePairs( - attset, ";", new char[] { ' ', '-' }); - - if ("similarity".equals(sf.getType())) - { - try - { - processGffSimilarity(set, seq, sf, align, newseqs, - relaxedIdMatching); - addFeature = false; - } catch (InvalidGFF3FieldException ivfe) - { - System.err.println(ivfe); - } - } - } - } - if (addFeature) - { - seq.addSequenceFeature(sf); - } - return addFeature; - } - -} - -class InvalidGFF3FieldException extends Exception -{ - String field, value; - - public InvalidGFF3FieldException(String field, - Map> set, String message) - { - super(message + " (Field was " + field + " and value was " - + set.get(field).toString()); - this.field = field; - this.value = set.get(field).toString(); - } }