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.datamodel;
23 import jalview.util.Comparison;
24 import jalview.util.Format;
25 import jalview.util.QuickSort;
26 import jalview.util.SparseCount;
29 * A class to count occurrences of residues in a profile, optimised for speed
30 * and memory footprint.
34 public class ResidueCount
37 * A data bean to hold the results of counting symbols
39 public class SymbolCounts
42 * the symbols seen (as char values), in no particular order
44 public final char[] symbols;
47 * the counts for each symbol, in the same order as the symbols
49 public final int[] values;
51 SymbolCounts(char[] s, int[] v)
58 private static final int TOUPPERCASE = 'A' - 'a';
61 * nucleotide symbols to count (including N unknown)
63 private static final String NUCS = "ACGNTU";
66 * amino acid symbols to count (including X unknown)
67 * NB we also include U so as to support counting of RNA bases
68 * in the "don't know" case of nucleotide / peptide
70 private static final String AAS = "ACDEFGHIKLMNPQRSTUVWXY";
72 private static final int GAP_COUNT = 0;
75 * fast lookup tables holding the index into our count
76 * arrays of each symbol; index 0 is reserved for gap counting
78 private static int[] NUC_INDEX = new int[26];
80 private static int[] AA_INDEX = new int[26];
83 for (int i = 0; i < NUCS.length(); i++)
85 NUC_INDEX[NUCS.charAt(i) - 'A'] = i + 1;
87 for (int i = 0; i < AAS.length(); i++)
89 AA_INDEX[AAS.charAt(i) - 'A'] = i + 1;
94 * counts array, just big enough for the nucleotide or peptide
95 * character set (plus gap counts in position 0)
97 private short[] counts;
100 * alternative array of int counts for use if any count
101 * exceeds the maximum value of short (32767)
103 private int[] intCounts;
106 * flag set if we switch from short to int counts
108 private boolean useIntCounts;
111 * general-purpose counter, only for use for characters
112 * that are not in the expected alphabet
114 private SparseCount otherData;
117 * keeps track of the maximum count value recorded
118 * (if this class ever allows decrements, would need to
119 * calculate this on request instead)
124 * if we think we are counting nucleotide, can get by with smaller
125 * array to hold counts
127 private boolean isNucleotide;
130 * Default constructor allocates arrays able to count either nucleotide or
131 * peptide bases. Use this constructor if not sure which the data is.
133 public ResidueCount()
139 * Constructor that allocates an array just big enough for the anticipated
140 * characters, plus one position to count gaps
142 public ResidueCount(boolean nucleotide)
144 isNucleotide = nucleotide;
145 int charsToCount = nucleotide ? NUCS.length() : AAS.length();
146 counts = new short[charsToCount + 1];
150 * Increments the count for the given character. The supplied character may be
151 * upper or lower case but counts are for the upper case only. Gap characters
152 * (space, ., -) are all counted together.
155 * @return the new value of the count for the character
157 public int add(final char c)
159 char u = toUpperCase(c);
161 int offset = getOffset(u);
164 * offset 0 is reserved for gap counting, so 0 here means either
165 * an unexpected character, or a gap character passed in error
169 if (Comparison.isGap(u))
175 newValue = addOtherCharacter(u);
180 newValue = increment(offset);
186 * Increment the count at the specified offset. If this would result in short
187 * overflow, promote to counting int values instead.
190 * @return the new value of the count at this offset
192 int increment(int offset)
197 newValue = intCounts[offset];
198 intCounts[offset] = ++newValue;
202 if (counts[offset] == Short.MAX_VALUE)
205 newValue = intCounts[offset];
206 intCounts[offset] = ++newValue;
210 newValue = counts[offset];
211 counts[offset] = (short) ++newValue;
214 maxCount = Math.max(maxCount, newValue);
219 * Switch from counting in short to counting in int
221 synchronized void handleOverflow()
223 intCounts = new int[counts.length];
224 for (int i = 0; i < counts.length; i++)
226 intCounts[i] = counts[i];
233 * Returns this character's offset in the count array
238 int getOffset(char c)
241 if ('A' <= c && c <= 'Z')
243 offset = isNucleotide ? NUC_INDEX[c - 'A'] : AA_INDEX[c - 'A'];
252 protected char toUpperCase(final char c)
255 if ('a' <= c && c <= 'z')
257 u = (char) (c + TOUPPERCASE);
263 * Increment count for some unanticipated character. The first time this
264 * called, a SparseCount is instantiated to hold these 'extra' counts.
267 * @return the new value of the count for the character
269 int addOtherCharacter(char c)
271 if (otherData == null)
273 otherData = new SparseCount();
275 int newValue = otherData.add(c, 1);
276 maxCount = Math.max(maxCount, newValue);
281 * Set count for some unanticipated character. The first time this called, a
282 * SparseCount is instantiated to hold these 'extra' counts.
287 void setOtherCharacter(char c, int value)
289 if (otherData == null)
291 otherData = new SparseCount();
293 otherData.put(c, value);
297 * Increment count of gap characters
299 * @return the new count of gaps
306 newValue = ++intCounts[GAP_COUNT];
310 newValue = ++counts[GAP_COUNT];
316 * Answers true if we are counting ints (only after overflow of short counts)
320 boolean isCountingInts()
326 * Sets the count for the given character. The supplied character may be upper
327 * or lower case but counts are for the upper case only.
332 public void put(char c, int count)
334 char u = toUpperCase(c);
335 int offset = getOffset(u);
338 * offset 0 is reserved for gap counting, so 0 here means either
339 * an unexpected character, or a gap character passed in error
343 if (Comparison.isGap(u))
349 setOtherCharacter(u, count);
350 maxCount = Math.max(maxCount, count);
356 maxCount = Math.max(maxCount, count);
361 * Sets the count at the specified offset. If this would result in short
362 * overflow, promote to counting int values instead.
367 void set(int offset, int value)
371 intCounts[offset] = value;
375 if (value > Short.MAX_VALUE || value < Short.MIN_VALUE)
378 intCounts[offset] = value;
382 counts[offset] = (short) value;
388 * Returns the count for the given character, or zero if no count held
393 public int getCount(char c)
395 char u = toUpperCase(c);
396 int offset = getOffset(u);
399 if (!Comparison.isGap(u))
401 // should have called getGapCount()
402 return otherData == null ? 0 : otherData.get(u);
405 return useIntCounts ? intCounts[offset] : counts[offset];
408 public int getGapCount()
410 return useIntCounts ? intCounts[0] : counts[0];
414 * Answers true if this object wraps a counter for unexpected characters
418 boolean isUsingOtherData()
420 return otherData != null;
424 * Returns the character (or concatenated characters) for the symbol(s) with
425 * the given count in the profile. Can be used to get the modal residue by
426 * supplying the modal count value. Returns an empty string if no symbol has
427 * the given count. The symbols are in alphabetic order of standard peptide or
428 * nucleotide characters, followed by 'other' symbols if any.
432 public String getResiduesForCount(int count)
440 * find counts for the given value and append the
441 * corresponding symbol
443 StringBuilder modal = new StringBuilder();
446 for (int i = 1; i < intCounts.length; i++)
448 if (intCounts[i] == count)
450 modal.append(isNucleotide ? NUCS.charAt(i - 1) : AAS
457 for (int i = 1; i < counts.length; i++)
459 if (counts[i] == count)
461 modal.append(isNucleotide ? NUCS.charAt(i - 1) : AAS
466 if (otherData != null)
468 for (int i = 0; i < otherData.size(); i++)
470 if (otherData.valueAt(i) == count)
472 modal.append((char) otherData.keyAt(i));
476 return modal.toString();
480 * Returns the highest count for any symbol(s) in the profile (excluding gap)
484 public int getModalCount()
490 * Returns the number of distinct symbols with a non-zero count (excluding the
499 for (int i = 1; i < intCounts.length; i++)
501 if (intCounts[i] > 0)
509 for (int i = 1; i < counts.length; i++)
519 * include 'other' characters recorded (even if count is zero
520 * though that would be a strange use case)
522 if (otherData != null)
524 size += otherData.size();
531 * Returns a data bean holding those symbols that have a non-zero count
532 * (excluding the gap symbol), with their counts.
536 public SymbolCounts getSymbolCounts()
539 char[] symbols = new char[size];
540 int[] values = new int[size];
545 for (int i = 1; i < intCounts.length; i++)
547 if (intCounts[i] > 0)
549 char symbol = isNucleotide ? NUCS.charAt(i - 1) : AAS
552 values[j] = intCounts[i];
559 for (int i = 1; i < counts.length; i++)
563 char symbol = isNucleotide ? NUCS.charAt(i - 1) : AAS
566 values[j] = counts[i];
571 if (otherData != null)
573 for (int i = 0; i < otherData.size(); i++)
575 symbols[j] = (char) otherData.keyAt(i);
576 values[j] = otherData.valueAt(i);
581 return new SymbolCounts(symbols, values);
585 * Returns a tooltip string showing residues in descending order of their
586 * percentage frequency in the profile
589 * the divisor for residue counts (may or may not include gapped
591 * @param percentageDecPl
592 * the number of decimal places to show in percentages
595 public String getTooltip(int normaliseBy, int percentageDecPl)
597 SymbolCounts symbolCounts = getSymbolCounts();
598 char[] ca = symbolCounts.symbols;
599 int[] vl = symbolCounts.values;
602 * sort characters into ascending order of their counts
604 QuickSort.sort(vl, ca);
607 * traverse in reverse order (highest count first) to build tooltip
609 boolean first = true;
610 StringBuilder sb = new StringBuilder(64);
611 for (int c = ca.length - 1; c >= 0; c--)
613 final char residue = ca[c];
614 // TODO combine residues which share a percentage
615 // (see AAFrequency.completeCdnaConsensus)
616 float tval = (vl[c] * 100f) / normaliseBy;
617 sb.append(first ? "" : "; ").append(residue).append(" ");
618 Format.appendPercentage(sb, tval, percentageDecPl);
622 return sb.toString();
626 * Returns a string representation of the symbol counts, for debug purposes.
629 public String toString()
631 StringBuilder sb = new StringBuilder();
633 SymbolCounts sc = getSymbolCounts();
634 for (int i = 0; i < sc.symbols.length; i++)
636 sb.append(sc.symbols[i]).append(":").append(sc.values[i]).append(" ");
639 return sb.toString();