2665N/A/**
2665N/A * @test
2665N/A * @bug 7070134
2665N/A * @summary Hotspot crashes with sigsegv from PorterStemmer
2665N/A *
2665N/A * @run shell Test7070134.sh
2665N/A */
2665N/A
2665N/A/*
2665N/A
2665N/A Porter stemmer in Java. The original paper is in
2665N/A
2665N/A Porter, 1980, An algorithm for suffix stripping, Program, Vol. 14,
2665N/A no. 3, pp 130-137,
2665N/A
3368N/A http://www.tartarus.org/~martin/PorterStemmer
3368N/A
3368N/A The software is completely free for any purpose, unless notes at the head
3368N/A of the program text indicates otherwise (which is rare). In any case,
3368N/A the notes about licensing are never more restrictive than the BSD License.
3368N/A
3368N/A In every case where the software is not written by me (Martin Porter),
3368N/A this licensing arrangement has been endorsed by the contributor, and it is
3368N/A therefore unnecessary to ask the contributor again to confirm it.
3368N/A
3368N/A I have not asked any contributors (or their employers, if they have them)
3368N/A for proofs that they have the right to distribute their software in this way.
2665N/A
2665N/A History:
2665N/A
2665N/A Release 1
2665N/A
2665N/A Bug 1 (reported by Gonzalo Parra 16/10/99) fixed as marked below.
2665N/A The words 'aed', 'eed', 'oed' leave k at 'a' for step 3, and b[k-1]
2665N/A is then out outside the bounds of b.
2665N/A
2665N/A Release 2
2665N/A
2665N/A Similarly,
2665N/A
2665N/A Bug 2 (reported by Steve Dyrdahl 22/2/00) fixed as marked below.
2665N/A 'ion' by itself leaves j = -1 in the test for 'ion' in step 5, and
2665N/A b[j] is then outside the bounds of b.
2665N/A
2665N/A Release 3
2665N/A
2665N/A Considerably revised 4/9/00 in the light of many helpful suggestions
2665N/A from Brian Goetz of Quiotix Corporation (brian@quiotix.com).
2665N/A
2665N/A Release 4
2665N/A
2665N/A*/
2665N/A
2665N/Aimport java.io.*;
2665N/A
2665N/A/**
2665N/A * Stemmer, implementing the Porter Stemming Algorithm
2665N/A *
2665N/A * The Stemmer class transforms a word into its root form. The input
2665N/A * word can be provided a character at time (by calling add()), or at once
2665N/A * by calling one of the various stem(something) methods.
2665N/A */
2665N/A
2665N/Aclass Stemmer
2665N/A{ private char[] b;
2665N/A private int i, /* offset into b */
2665N/A i_end, /* offset to end of stemmed word */
2665N/A j, k;
2665N/A private static final int INC = 50;
2665N/A /* unit of size whereby b is increased */
2665N/A public Stemmer()
2665N/A { b = new char[INC];
2665N/A i = 0;
2665N/A i_end = 0;
2665N/A }
2665N/A
2665N/A /**
2665N/A * Add a character to the word being stemmed. When you are finished
2665N/A * adding characters, you can call stem(void) to stem the word.
2665N/A */
2665N/A
2665N/A public void add(char ch)
2665N/A { if (i == b.length)
2665N/A { char[] new_b = new char[i+INC];
2665N/A for (int c = 0; c < i; c++) new_b[c] = b[c];
2665N/A b = new_b;
2665N/A }
2665N/A b[i++] = ch;
2665N/A }
2665N/A
2665N/A
2665N/A /** Adds wLen characters to the word being stemmed contained in a portion
2665N/A * of a char[] array. This is like repeated calls of add(char ch), but
2665N/A * faster.
2665N/A */
2665N/A
2665N/A public void add(char[] w, int wLen)
2665N/A { if (i+wLen >= b.length)
2665N/A { char[] new_b = new char[i+wLen+INC];
2665N/A for (int c = 0; c < i; c++) new_b[c] = b[c];
2665N/A b = new_b;
2665N/A }
2665N/A for (int c = 0; c < wLen; c++) b[i++] = w[c];
2665N/A }
2665N/A
2665N/A /**
2665N/A * After a word has been stemmed, it can be retrieved by toString(),
2665N/A * or a reference to the internal buffer can be retrieved by getResultBuffer
2665N/A * and getResultLength (which is generally more efficient.)
2665N/A */
2665N/A public String toString() { return new String(b,0,i_end); }
2665N/A
2665N/A /**
2665N/A * Returns the length of the word resulting from the stemming process.
2665N/A */
2665N/A public int getResultLength() { return i_end; }
2665N/A
2665N/A /**
2665N/A * Returns a reference to a character buffer containing the results of
2665N/A * the stemming process. You also need to consult getResultLength()
2665N/A * to determine the length of the result.
2665N/A */
2665N/A public char[] getResultBuffer() { return b; }
2665N/A
2665N/A /* cons(i) is true <=> b[i] is a consonant. */
2665N/A
2665N/A private final boolean cons(int i)
2665N/A { switch (b[i])
2665N/A { case 'a': case 'e': case 'i': case 'o': case 'u': return false;
2665N/A case 'y': return (i==0) ? true : !cons(i-1);
2665N/A default: return true;
2665N/A }
2665N/A }
2665N/A
2665N/A /* m() measures the number of consonant sequences between 0 and j. if c is
2665N/A a consonant sequence and v a vowel sequence, and <..> indicates arbitrary
2665N/A presence,
2665N/A
2665N/A <c><v> gives 0
2665N/A <c>vc<v> gives 1
2665N/A <c>vcvc<v> gives 2
2665N/A <c>vcvcvc<v> gives 3
2665N/A ....
2665N/A */
2665N/A
2665N/A private final int m()
2665N/A { int n = 0;
2665N/A int i = 0;
2665N/A while(true)
2665N/A { if (i > j) return n;
2665N/A if (! cons(i)) break; i++;
2665N/A }
2665N/A i++;
2665N/A while(true)
2665N/A { while(true)
2665N/A { if (i > j) return n;
2665N/A if (cons(i)) break;
2665N/A i++;
2665N/A }
2665N/A i++;
2665N/A n++;
2665N/A while(true)
2665N/A { if (i > j) return n;
2665N/A if (! cons(i)) break;
2665N/A i++;
2665N/A }
2665N/A i++;
2665N/A }
2665N/A }
2665N/A
2665N/A /* vowelinstem() is true <=> 0,...j contains a vowel */
2665N/A
2665N/A private final boolean vowelinstem()
2665N/A { int i; for (i = 0; i <= j; i++) if (! cons(i)) return true;
2665N/A return false;
2665N/A }
2665N/A
2665N/A /* doublec(j) is true <=> j,(j-1) contain a double consonant. */
2665N/A
2665N/A private final boolean doublec(int j)
2665N/A { if (j < 1) return false;
2665N/A if (b[j] != b[j-1]) return false;
2665N/A return cons(j);
2665N/A }
2665N/A
2665N/A /* cvc(i) is true <=> i-2,i-1,i has the form consonant - vowel - consonant
2665N/A and also if the second c is not w,x or y. this is used when trying to
2665N/A restore an e at the end of a short word. e.g.
2665N/A
2665N/A cav(e), lov(e), hop(e), crim(e), but
2665N/A snow, box, tray.
2665N/A
2665N/A */
2665N/A
2665N/A private final boolean cvc(int i)
2665N/A { if (i < 2 || !cons(i) || cons(i-1) || !cons(i-2)) return false;
2665N/A { int ch = b[i];
2665N/A if (ch == 'w' || ch == 'x' || ch == 'y') return false;
2665N/A }
2665N/A return true;
2665N/A }
2665N/A
2665N/A private final boolean ends(String s)
2665N/A { int l = s.length();
2665N/A int o = k-l+1;
2665N/A if (o < 0) return false;
2665N/A for (int i = 0; i < l; i++) if (b[o+i] != s.charAt(i)) return false;
2665N/A j = k-l;
2665N/A return true;
2665N/A }
2665N/A
2665N/A /* setto(s) sets (j+1),...k to the characters in the string s, readjusting
2665N/A k. */
2665N/A
2665N/A private final void setto(String s)
2665N/A { int l = s.length();
2665N/A int o = j+1;
2665N/A for (int i = 0; i < l; i++) b[o+i] = s.charAt(i);
2665N/A k = j+l;
2665N/A }
2665N/A
2665N/A /* r(s) is used further down. */
2665N/A
2665N/A private final void r(String s) { if (m() > 0) setto(s); }
2665N/A
2665N/A /* step1() gets rid of plurals and -ed or -ing. e.g.
2665N/A
2665N/A caresses -> caress
2665N/A ponies -> poni
2665N/A ties -> ti
2665N/A caress -> caress
2665N/A cats -> cat
2665N/A
2665N/A feed -> feed
2665N/A agreed -> agree
2665N/A disabled -> disable
2665N/A
2665N/A matting -> mat
2665N/A mating -> mate
2665N/A meeting -> meet
2665N/A milling -> mill
2665N/A messing -> mess
2665N/A
2665N/A meetings -> meet
2665N/A
2665N/A */
2665N/A
2665N/A private final void step1()
2665N/A { if (b[k] == 's')
2665N/A { if (ends("sses")) k -= 2; else
2665N/A if (ends("ies")) setto("i"); else
2665N/A if (b[k-1] != 's') k--;
2665N/A }
2665N/A if (ends("eed")) { if (m() > 0) k--; } else
2665N/A if ((ends("ed") || ends("ing")) && vowelinstem())
2665N/A { k = j;
2665N/A if (ends("at")) setto("ate"); else
2665N/A if (ends("bl")) setto("ble"); else
2665N/A if (ends("iz")) setto("ize"); else
2665N/A if (doublec(k))
2665N/A { k--;
2665N/A { int ch = b[k];
2665N/A if (ch == 'l' || ch == 's' || ch == 'z') k++;
2665N/A }
2665N/A }
2665N/A else if (m() == 1 && cvc(k)) setto("e");
2665N/A }
2665N/A }
2665N/A
2665N/A /* step2() turns terminal y to i when there is another vowel in the stem. */
2665N/A
2665N/A private final void step2() { if (ends("y") && vowelinstem()) b[k] = 'i'; }
2665N/A
2665N/A /* step3() maps double suffices to single ones. so -ization ( = -ize plus
2665N/A -ation) maps to -ize etc. note that the string before the suffix must give
2665N/A m() > 0. */
2665N/A
2665N/A private final void step3() { if (k == 0) return; /* For Bug 1 */ switch (b[k-1])
2665N/A {
2665N/A case 'a': if (ends("ational")) { r("ate"); break; }
2665N/A if (ends("tional")) { r("tion"); break; }
2665N/A break;
2665N/A case 'c': if (ends("enci")) { r("ence"); break; }
2665N/A if (ends("anci")) { r("ance"); break; }
2665N/A break;
2665N/A case 'e': if (ends("izer")) { r("ize"); break; }
2665N/A break;
2665N/A case 'l': if (ends("bli")) { r("ble"); break; }
2665N/A if (ends("alli")) { r("al"); break; }
2665N/A if (ends("entli")) { r("ent"); break; }
2665N/A if (ends("eli")) { r("e"); break; }
2665N/A if (ends("ousli")) { r("ous"); break; }
2665N/A break;
2665N/A case 'o': if (ends("ization")) { r("ize"); break; }
2665N/A if (ends("ation")) { r("ate"); break; }
2665N/A if (ends("ator")) { r("ate"); break; }
2665N/A break;
2665N/A case 's': if (ends("alism")) { r("al"); break; }
2665N/A if (ends("iveness")) { r("ive"); break; }
2665N/A if (ends("fulness")) { r("ful"); break; }
2665N/A if (ends("ousness")) { r("ous"); break; }
2665N/A break;
2665N/A case 't': if (ends("aliti")) { r("al"); break; }
2665N/A if (ends("iviti")) { r("ive"); break; }
2665N/A if (ends("biliti")) { r("ble"); break; }
2665N/A break;
2665N/A case 'g': if (ends("logi")) { r("log"); break; }
2665N/A } }
2665N/A
2665N/A /* step4() deals with -ic-, -full, -ness etc. similar strategy to step3. */
2665N/A
2665N/A private final void step4() { switch (b[k])
2665N/A {
2665N/A case 'e': if (ends("icate")) { r("ic"); break; }
2665N/A if (ends("ative")) { r(""); break; }
2665N/A if (ends("alize")) { r("al"); break; }
2665N/A break;
2665N/A case 'i': if (ends("iciti")) { r("ic"); break; }
2665N/A break;
2665N/A case 'l': if (ends("ical")) { r("ic"); break; }
2665N/A if (ends("ful")) { r(""); break; }
2665N/A break;
2665N/A case 's': if (ends("ness")) { r(""); break; }
2665N/A break;
2665N/A } }
2665N/A
2665N/A /* step5() takes off -ant, -ence etc., in context <c>vcvc<v>. */
2665N/A
2665N/A private final void step5()
2665N/A { if (k == 0) return; /* for Bug 1 */ switch (b[k-1])
2665N/A { case 'a': if (ends("al")) break; return;
2665N/A case 'c': if (ends("ance")) break;
2665N/A if (ends("ence")) break; return;
2665N/A case 'e': if (ends("er")) break; return;
2665N/A case 'i': if (ends("ic")) break; return;
2665N/A case 'l': if (ends("able")) break;
2665N/A if (ends("ible")) break; return;
2665N/A case 'n': if (ends("ant")) break;
2665N/A if (ends("ement")) break;
2665N/A if (ends("ment")) break;
2665N/A /* element etc. not stripped before the m */
2665N/A if (ends("ent")) break; return;
2665N/A case 'o': if (ends("ion") && j >= 0 && (b[j] == 's' || b[j] == 't')) break;
2665N/A /* j >= 0 fixes Bug 2 */
2665N/A if (ends("ou")) break; return;
2665N/A /* takes care of -ous */
2665N/A case 's': if (ends("ism")) break; return;
2665N/A case 't': if (ends("ate")) break;
2665N/A if (ends("iti")) break; return;
2665N/A case 'u': if (ends("ous")) break; return;
2665N/A case 'v': if (ends("ive")) break; return;
2665N/A case 'z': if (ends("ize")) break; return;
2665N/A default: return;
2665N/A }
2665N/A if (m() > 1) k = j;
2665N/A }
2665N/A
2665N/A /* step6() removes a final -e if m() > 1. */
2665N/A
2665N/A private final void step6()
2665N/A { j = k;
2665N/A if (b[k] == 'e')
2665N/A { int a = m();
2665N/A if (a > 1 || a == 1 && !cvc(k-1)) k--;
2665N/A }
2665N/A if (b[k] == 'l' && doublec(k) && m() > 1) k--;
2665N/A }
2665N/A
2665N/A /** Stem the word placed into the Stemmer buffer through calls to add().
2665N/A * Returns true if the stemming process resulted in a word different
2665N/A * from the input. You can retrieve the result with
2665N/A * getResultLength()/getResultBuffer() or toString().
2665N/A */
2665N/A public void stem()
2665N/A { k = i - 1;
2665N/A if (k > 1) { step1(); step2(); step3(); step4(); step5(); step6(); }
2665N/A i_end = k+1; i = 0;
2665N/A }
2665N/A
2665N/A /** Test program for demonstrating the Stemmer. It reads text from a
2665N/A * a list of files, stems each word, and writes the result to standard
2665N/A * output. Note that the word stemmed is expected to be in lower case:
2665N/A * forcing lower case must be done outside the Stemmer class.
2665N/A * Usage: Stemmer file-name file-name ...
2665N/A */
2665N/A public static void main(String[] args)
2665N/A {
2665N/A char[] w = new char[501];
2665N/A Stemmer s = new Stemmer();
2665N/A for (int i = 0; i < args.length; i++)
2665N/A try
2665N/A {
2665N/A FileInputStream in = new FileInputStream(args[i]);
2665N/A
2665N/A try
2665N/A { while(true)
2665N/A
2665N/A { int ch = in.read();
2665N/A if (Character.isLetter((char) ch))
2665N/A {
2665N/A int j = 0;
2665N/A while(true)
2665N/A { ch = Character.toLowerCase((char) ch);
2665N/A w[j] = (char) ch;
2665N/A if (j < 500) j++;
2665N/A ch = in.read();
2665N/A if (!Character.isLetter((char) ch))
2665N/A {
2665N/A /* to test add(char ch) */
2665N/A for (int c = 0; c < j; c++) s.add(w[c]);
2665N/A
2665N/A /* or, to test add(char[] w, int j) */
2665N/A /* s.add(w, j); */
2665N/A
2665N/A s.stem();
2665N/A { String u;
2665N/A
2665N/A /* and now, to test toString() : */
2665N/A u = s.toString();
2665N/A
2665N/A /* to test getResultBuffer(), getResultLength() : */
2665N/A /* u = new String(s.getResultBuffer(), 0, s.getResultLength()); */
2665N/A
2665N/A System.out.print(u);
2665N/A }
2665N/A break;
2665N/A }
2665N/A }
2665N/A }
2665N/A if (ch < 0) break;
2665N/A System.out.print((char)ch);
2665N/A }
2665N/A }
2665N/A catch (IOException e)
2665N/A { System.out.println("error reading " + args[i]);
2665N/A break;
2665N/A }
2665N/A }
2665N/A catch (FileNotFoundException e)
2665N/A { System.out.println("file " + args[i] + " not found");
2665N/A break;
2665N/A }
2665N/A }
2665N/A}