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.
21 package jalview.io.gff;
23 import jalview.analysis.SequenceIdMatcher;
24 import jalview.datamodel.AlignedCodonFrame;
25 import jalview.datamodel.AlignmentI;
26 import jalview.datamodel.MappingType;
27 import jalview.datamodel.SequenceDummy;
28 import jalview.datamodel.SequenceFeature;
29 import jalview.datamodel.SequenceI;
30 import jalview.util.MapList;
31 import jalview.util.StringUtils;
33 import java.util.ArrayList;
34 import java.util.Arrays;
35 import java.util.HashMap;
36 import java.util.List;
38 import java.util.Map.Entry;
41 * Base class with common functionality for flavours of GFF handler (GFF2 or
44 public abstract class GffHelperBase implements GffHelperI
46 private static final String INVALID_GFF_ATTRIBUTE_FORMAT = "Invalid GFF attribute format: ";
48 protected static final String COMMA = ",";
50 protected static final String EQUALS = "=";
52 protected static final String NOTE = "Note";
55 * GFF columns 1-9 (zero-indexed):
57 protected static final int SEQID_COL = 0;
59 protected static final int SOURCE_COL = 1;
61 protected static final int TYPE_COL = 2;
63 protected static final int START_COL = 3;
65 protected static final int END_COL = 4;
67 protected static final int SCORE_COL = 5;
69 protected static final int STRAND_COL = 6;
71 protected static final int PHASE_COL = 7;
73 protected static final int ATTRIBUTES_COL = 8;
75 private AlignmentI lastmatchedAl = null;
77 private SequenceIdMatcher matcher = null;
80 * Constructs and returns a mapping, or null if data appear invalid
87 * type of mapping (e.g. protein to nucleotide)
90 protected MapList constructMappingFromAlign(int fromStart, int fromEnd,
91 int toStart, int toEnd, MappingType mappingType)
93 int[] from = new int[] { fromStart, fromEnd };
94 int[] to = new int[] { toStart, toEnd };
97 * Jalview always models from dna to protein, so switch values if the
98 * GFF mapping is from protein to dna
100 if (mappingType == MappingType.PeptideToNucleotide)
105 mappingType = mappingType.getInverse();
108 int fromRatio = mappingType.getFromRatio();
109 int toRatio = mappingType.getToRatio();
112 * sanity check that mapped residue counts match
113 * TODO understand why PASA generates such cases...
115 if (!trimMapping(from, to, fromRatio, toRatio))
117 System.err.println("Ignoring mapping from " + Arrays.toString(from)
118 + " to " + Arrays.toString(to) + " as counts don't match!");
123 * If a codon has an intron gap, there will be contiguous 'toRanges';
124 * this is handled for us by the MapList constructor.
125 * (It is not clear that exonerate ever generates this case)
128 return new MapList(from, to, fromRatio, toRatio);
132 * Checks that the 'from' and 'to' ranges have equivalent lengths. If not,
133 * tries to trim the end of the longer so they do. Returns true if the
134 * mappings could be made equivalent, else false. Note the range array values
135 * may be modified by this method.
143 protected static boolean trimMapping(int[] from, int[] to, int fromRatio,
146 int fromLength = Math.abs(from[1] - from[0]) + 1;
147 int toLength = Math.abs(to[1] - to[0]) + 1;
148 int fromOverlap = fromLength * toRatio - toLength * fromRatio;
149 if (fromOverlap == 0)
153 if (fromOverlap > 0 && fromOverlap % toRatio == 0)
156 * restrict from range to make them match up
157 * it's kind of arbitrary which end we truncate - here it is the end
160 "Truncating mapping from " + Arrays.toString(from) + " to ");
161 if (from[1] > from[0])
163 from[1] -= fromOverlap / toRatio;
167 from[1] += fromOverlap / toRatio;
169 System.err.println(Arrays.toString(from));
172 else if (fromOverlap < 0 && fromOverlap % fromRatio == 0)
174 fromOverlap = -fromOverlap; // > 0
176 * restrict to range to make them match up
179 "Truncating mapping to " + Arrays.toString(to) + " to ");
182 to[1] -= fromOverlap / fromRatio;
186 to[1] += fromOverlap / fromRatio;
188 System.err.println(Arrays.toString(to));
193 * Couldn't truncate to an exact match..
199 * Returns a sequence matching the given id, as follows
201 * <li>strict matching is on exact sequence name</li>
202 * <li>relaxed matching allows matching on a token within the sequence name,
204 * <li>first tries to find a match in the alignment sequences</li>
205 * <li>else tries to find a match in the new sequences already generated while
206 * parsing the features file</li>
207 * <li>else creates a new placeholder sequence, adds it to the new sequences
208 * list, and returns it</li>
214 * @param relaxedIdMatching
218 protected SequenceI findSequence(String seqId, AlignmentI align,
219 List<SequenceI> newseqs, boolean relaxedIdMatching)
225 SequenceI match = null;
226 if (relaxedIdMatching)
228 if (lastmatchedAl != align)
230 lastmatchedAl = align;
231 matcher = new SequenceIdMatcher(align.getSequencesArray());
234 matcher.addAll(newseqs);
237 match = matcher.findIdMatch(seqId);
241 match = align.findName(seqId, true);
242 if (match == null && newseqs != null)
244 for (SequenceI m : newseqs)
246 if (seqId.equals(m.getName()))
254 if (match == null && newseqs != null)
256 match = new SequenceDummy(seqId);
257 if (relaxedIdMatching)
259 matcher.addAll(Arrays.asList(new SequenceI[] { match }));
261 // add dummy sequence to the newseqs list
268 * Parses the input line to a map of name / value(s) pairs. For example the
272 * Notes=Fe-S;Method=manual curation, prediction; source = Pfam; Notes = Metal
275 * if parsed with delimiter=";" and separators {' ', '='} <br>
276 * would return a map with { Notes={Fe=S, Metal}, Method={manual curation,
277 * prediction}, source={Pfam}} <br>
279 * This method supports parsing of either GFF2 format (which uses space ' ' as
280 * the name/value delimiter, and allows multiple occurrences of the same
281 * name), or GFF3 format (which uses '=' as the name/value delimiter, and
282 * strictly does not allow repeat occurrences of the same name - but does
283 * allow a comma-separated list of values).
285 * Returns a (possibly empty) map of lists of values by attribute name.
288 * @param namesDelimiter
289 * the major delimiter between name-value pairs
290 * @param nameValueSeparator
291 * separator used between name and value
292 * @param valuesDelimiter
293 * delimits a list of more than one value
296 public static Map<String, List<String>> parseNameValuePairs(String text,
297 String namesDelimiter, char nameValueSeparator,
298 String valuesDelimiter)
300 Map<String, List<String>> map = new HashMap<>();
301 if (text == null || text.trim().length() == 0)
307 * split by major delimiter (; for GFF3)
309 for (String nameValuePair : text.trim().split(namesDelimiter))
311 nameValuePair = nameValuePair.trim();
312 if (nameValuePair.length() == 0)
318 * find name/value separator (= for GFF3)
320 int sepPos = nameValuePair.indexOf(nameValueSeparator);
323 // no name=value found
327 String name = nameValuePair.substring(0, sepPos).trim();
328 String values = nameValuePair.substring(sepPos + 1).trim();
329 if (values.isEmpty())
334 List<String> vals = map.get(name);
337 vals = new ArrayList<>();
342 * if 'values' contains more name/value separators, parse as a map
343 * (nested sub-attribute values)
345 if (values.indexOf(nameValueSeparator) != -1)
351 for (String val : values.split(valuesDelimiter))
362 * Constructs a SequenceFeature from the GFF column data. Subclasses may wish
363 * to call this method then adjust the SequenceFeature depending on the
364 * particular usage of different tools that generate GFF.
370 protected SequenceFeature buildSequenceFeature(String[] gff,
371 Map<String, List<String>> attributes)
373 return buildSequenceFeature(gff, TYPE_COL, gff[SOURCE_COL], attributes);
383 protected SequenceFeature buildSequenceFeature(String[] gff,
384 int typeColumn, String group, Map<String, List<String>> attributes)
388 int start = Integer.parseInt(gff[START_COL]);
389 int end = Integer.parseInt(gff[END_COL]);
392 * default 'score' is 0 rather than Float.NaN - see JAL-2554
397 score = Float.parseFloat(gff[SCORE_COL]);
398 } catch (NumberFormatException nfe)
400 // e.g. '.' - leave as zero
403 SequenceFeature sf = new SequenceFeature(gff[typeColumn],
404 gff[SOURCE_COL], start, end, score, group);
406 sf.setStrand(gff[STRAND_COL]);
408 sf.setPhase(gff[PHASE_COL]);
410 if (attributes != null)
413 * Add attributes in column 9 to the sequence feature's
414 * 'otherData' table; use Note as a best proxy for description;
415 * decode any encoded comma, equals, semi-colon as per GFF3 spec
417 for (Entry<String, List<String>> attr : attributes.entrySet())
419 String key = attr.getKey();
420 List<String> values = attr.getValue();
421 if (values.size() == 1 && values.get(0).contains(EQUALS))
424 * 'value' is actually nested subattributes as x=a,y=b,z=c
426 Map<String, String> valueMap = parseAttributeMap(values.get(0));
427 sf.setValue(key, valueMap);
431 String csvValues = StringUtils.listToDelimitedString(values,
433 csvValues = StringUtils.urlDecode(csvValues, GFF_ENCODABLE);
434 sf.setValue(key, csvValues);
435 if (NOTE.equals(key))
437 sf.setDescription(csvValues);
444 } catch (NumberFormatException nfe)
446 System.err.println("Invalid number in gff: " + nfe.getMessage());
452 * Parses a (GFF3 format) list of comma-separated key=value pairs into a Map
455 * An input string like {@code a=b,c,d=e,f=g,h} is parsed to
467 protected static Map<String, String> parseAttributeMap(String s)
469 Map<String, String> map = new HashMap<>();
470 String[] fields = s.split(EQUALS);
475 boolean valid = true;
476 if (fields.length < 2)
479 * need at least A=B here
483 else if (fields[0].isEmpty() || fields[0].contains(COMMA))
486 * A,B=C is not a valid start, nor is =C
492 for (int i = 1; i < fields.length - 1; i++)
494 if (fields[i].isEmpty() || !fields[i].contains(COMMA))
497 * intermediate tokens must include value,name
506 System.err.println(INVALID_GFF_ATTRIBUTE_FORMAT + s);
511 while (i < fields.length - 1)
513 boolean lastPair = i == fields.length - 2;
514 String before = fields[i];
515 String after = fields[i + 1];
518 * if 'key' looks like a,b,c then the last token is the
521 String theKey = before.contains(COMMA)
522 ? before.substring(before.lastIndexOf(COMMA) + 1)
525 theKey = theKey.trim();
526 if (theKey.isEmpty())
528 System.err.println(INVALID_GFF_ATTRIBUTE_FORMAT + s);
534 * if 'value' looks like a,b,c then all but the last token is the value,
535 * unless this is the last field (no more = to follow), in which case
536 * all of it makes up the value
538 String theValue = after.contains(COMMA) && !lastPair
539 ? after.substring(0, after.lastIndexOf(COMMA))
541 map.put(StringUtils.urlDecode(theKey, GFF_ENCODABLE),
542 StringUtils.urlDecode(theValue, GFF_ENCODABLE));
550 * Returns any existing mapping held on the alignment between the given
551 * dataset sequences, or a new one if none found. This is a convenience method
552 * to facilitate processing multiple GFF lines that make up a single 'spliced'
553 * mapping, by extending the first mapping as the others are read.
560 protected AlignedCodonFrame getMapping(AlignmentI align,
561 SequenceI fromSeq, SequenceI toSeq)
563 AlignedCodonFrame acf = align.getMapping(fromSeq, toSeq);
566 acf = new AlignedCodonFrame();