|
From: <vic...@us...> - 2006-06-09 00:48:21
|
Revision: 7473 Author: victormote Date: 2006-06-08 17:48:13 -0700 (Thu, 08 Jun 2006) ViewCVS: http://svn.sourceforge.net/foray/?rev=7473&view=rev Log Message: ----------- Extract some methods for simplicity and clarity. Modified Paths: -------------- trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java Modified: trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java =================================================================== --- trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java 2006-06-08 23:53:02 UTC (rev 7472) +++ trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java 2006-06-09 00:48:13 UTC (rev 7473) @@ -292,45 +292,42 @@ return null; } + String theWord = new String(w, offset, len); + Hyphenation hyphenation = checkExceptions(theWord, remainCharCount, + pushCharCount); + if (hyphenation != null) { + return hyphenation; + } + + + // use algorithm to get hyphenation points int[] result = new int[len + 1]; int k = 0; + word[0] = '.'; // word start marker + word[len + 1] = '.'; // word end marker + word[len + 2] = 0; // null terminated + byte[] il = new byte[len + 3]; // initialized to zero + for (int i = 0; i < len + 1; i++) { + searchPatterns(word, i, il); + } - // check exception list first - String sw = new String(word, 1, len); - if (stoplist.containsKey(sw)) { - /* assume only simple hyphens (Hyphen.pre="-", - * Hyphen.post = Hyphen.no = null) */ - ArrayList hw = (ArrayList)stoplist.get(sw); - int j = 0; - for (int i = 0; i < hw.size(); i++) { - Object o = hw.get(i); - if (o instanceof String) { - j += ((String)o).length(); - if (j >= remainCharCount && j < (len - pushCharCount)) { - result[k++] = j; - } - } + // hyphenation points are located where interletter value is odd + for (int i = 0; i < len; i++) { + if (((il[i + 1] & 1) == 1) && i >= remainCharCount + && i <= (len - pushCharCount)) { + result[k++] = i; } - } else { - // use algorithm to get hyphenation points - word[0] = '.'; // word start marker - word[len + 1] = '.'; // word end marker - word[len + 2] = 0; // null terminated - byte[] il = new byte[len + 3]; // initialized to zero - for (int i = 0; i < len + 1; i++) { - searchPatterns(word, i, il); - } - - // hyphenation points are located where interletter value is odd - for (int i = 0; i < len; i++) { - if (((il[i + 1] & 1) == 1) && i >= remainCharCount - && i <= (len - pushCharCount)) { - result[k++] = i; - } - } } + return createHyphenation(theWord, result, k); + } - + /** + * @param result + * @param k + * @return The new Hyphenation instance. + */ + private Hyphenation createHyphenation(String theWord, + int[] result, int k) { if (k > 0) { // trim result array int[] res = new int[k]; @@ -338,7 +335,7 @@ /* TODO: values needs to be populated. Also both positive and * negative values should be returned. */ byte[] values = new byte[res.length]; - return new Hyphenation(new String(w, offset, len), res, values); + return new Hyphenation(theWord, res, values); } return null; } @@ -358,6 +355,32 @@ return word; } + private Hyphenation checkExceptions(String sw, int remainCharCount, + int pushCharCount) { + // check exception list first + if (stoplist.containsKey(sw)) { + int length = sw.length(); + int[] result = new int[length + 1]; + int k = 0; + /* assume only simple hyphens (Hyphen.pre="-", + * Hyphen.post = Hyphen.no = null) */ + ArrayList hw = (ArrayList)stoplist.get(sw); + int j = 0; + for (int i = 0; i < hw.size(); i++) { + Object o = hw.get(i); + if (o instanceof String) { + j += ((String)o).length(); + if (j >= remainCharCount + && j < (length - pushCharCount)) { + result[k++] = j; + } + } + } + return createHyphenation(sw, result, k); + } + return null; + } + /** * Add a character class to the tree. It is used by * {@link PatternParser PatternParser} as callback to This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site. |
|
From: <vic...@us...> - 2006-06-09 00:50:32
|
Revision: 7474 Author: victormote Date: 2006-06-08 17:50:28 -0700 (Thu, 08 Jun 2006) ViewCVS: http://svn.sourceforge.net/foray/?rev=7474&view=rev Log Message: ----------- Reverse sense of test to flatten logic. Modified Paths: -------------- trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java Modified: trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java =================================================================== --- trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java 2006-06-09 00:48:13 UTC (rev 7473) +++ trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java 2006-06-09 00:50:28 UTC (rev 7474) @@ -357,28 +357,27 @@ private Hyphenation checkExceptions(String sw, int remainCharCount, int pushCharCount) { - // check exception list first - if (stoplist.containsKey(sw)) { - int length = sw.length(); - int[] result = new int[length + 1]; - int k = 0; - /* assume only simple hyphens (Hyphen.pre="-", - * Hyphen.post = Hyphen.no = null) */ - ArrayList hw = (ArrayList)stoplist.get(sw); - int j = 0; - for (int i = 0; i < hw.size(); i++) { - Object o = hw.get(i); - if (o instanceof String) { - j += ((String)o).length(); - if (j >= remainCharCount - && j < (length - pushCharCount)) { - result[k++] = j; - } + if (! stoplist.containsKey(sw)) { + return null; + } + int length = sw.length(); + int[] result = new int[length + 1]; + int k = 0; + /* assume only simple hyphens (Hyphen.pre="-", + * Hyphen.post = Hyphen.no = null) */ + ArrayList hw = (ArrayList)stoplist.get(sw); + int j = 0; + for (int i = 0; i < hw.size(); i++) { + Object o = hw.get(i); + if (o instanceof String) { + j += ((String)o).length(); + if (j >= remainCharCount + && j < (length - pushCharCount)) { + result[k++] = j; } } - return createHyphenation(sw, result, k); } - return null; + return createHyphenation(sw, result, k); } /** This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site. |
|
From: <vic...@us...> - 2006-06-09 01:11:00
|
Revision: 7475 Author: victormote Date: 2006-06-08 18:10:55 -0700 (Thu, 08 Jun 2006) ViewCVS: http://svn.sourceforge.net/foray/?rev=7475&view=rev Log Message: ----------- 1. Return all values, including the negative values. 2. Return the weights. Modified Paths: -------------- trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java Modified: trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java =================================================================== --- trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java 2006-06-09 00:50:28 UTC (rev 7474) +++ trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java 2006-06-09 01:10:55 UTC (rev 7475) @@ -302,6 +302,7 @@ // use algorithm to get hyphenation points int[] result = new int[len + 1]; + byte[] values = new byte[result.length]; int k = 0; word[0] = '.'; // word start marker word[len + 1] = '.'; // word end marker @@ -313,28 +314,58 @@ // hyphenation points are located where interletter value is odd for (int i = 0; i < len; i++) { - if (((il[i + 1] & 1) == 1) && i >= remainCharCount + int interletterValue = il[i + 1]; + if (interletterValue != 0 + && i >= remainCharCount && i <= (len - pushCharCount)) { - result[k++] = i; + k++; + result[k] = i; + byte weight = convertLiangToWeight(interletterValue); + values[k] = weight; } } - return createHyphenation(theWord, result, k); + return createHyphenation(theWord, result, values, k); } /** + * Converts the raw value from the Liang data to a weighted value. + * In the Liang system, 1) even values are negative, 2) odd values are + * positive, and the magnitude of the number conveys the magnitude of + * the direction. + * @param liangValue A value in the range 0-9 that contains the Liang + * algorithm weight. + * @return A value in the range -4 to +5 that reflects the numeric weight + * of the raw Liang value. + */ + public static byte convertLiangToWeight(int liangValue) { + switch (liangValue) { + case 0: return 0; + case 1: return 1; + case 2: return -1; + case 3: return 2; + case 4: return -2; + case 5: return 3; + case 6: return -3; + case 7: return 4; + case 8: return -4; + case 9: return 5; + default: return 0; + } + } + + /** * @param result * @param k * @return The new Hyphenation instance. */ - private Hyphenation createHyphenation(String theWord, - int[] result, int k) { + private Hyphenation createHyphenation(String theWord, int[] result, + byte[] values, int k) { if (k > 0) { // trim result array int[] res = new int[k]; System.arraycopy(result, 0, res, 0, k); /* TODO: values needs to be populated. Also both positive and * negative values should be returned. */ - byte[] values = new byte[res.length]; return new Hyphenation(theWord, res, values); } return null; @@ -377,7 +408,12 @@ } } } - return createHyphenation(sw, result, k); + /* Assume that all found points have a value of 1. */ + byte[] values = new byte[result.length]; + for (int i = 0; i < values.length; i++) { + values[i] = 1; + } + return createHyphenation(sw, result, values, k); } /** This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site. |
|
From: <vic...@us...> - 2006-06-09 01:18:07
|
Revision: 7476 Author: victormote Date: 2006-06-08 18:18:03 -0700 (Thu, 08 Jun 2006) ViewCVS: http://svn.sourceforge.net/foray/?rev=7476&view=rev Log Message: ----------- Extract method for clarity and consistency. Modified Paths: -------------- trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java Modified: trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java =================================================================== --- trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java 2006-06-09 01:10:55 UTC (rev 7475) +++ trunk/foray/foray-hyphen-r/src/java/org/foray/hyphenR/HyphenationTree.java 2006-06-09 01:18:03 UTC (rev 7476) @@ -298,33 +298,9 @@ if (hyphenation != null) { return hyphenation; } - - - // use algorithm to get hyphenation points - int[] result = new int[len + 1]; - byte[] values = new byte[result.length]; - int k = 0; - word[0] = '.'; // word start marker - word[len + 1] = '.'; // word end marker - word[len + 2] = 0; // null terminated - byte[] il = new byte[len + 3]; // initialized to zero - for (int i = 0; i < len + 1; i++) { - searchPatterns(word, i, il); - } - - // hyphenation points are located where interletter value is odd - for (int i = 0; i < len; i++) { - int interletterValue = il[i + 1]; - if (interletterValue != 0 - && i >= remainCharCount - && i <= (len - pushCharCount)) { - k++; - result[k] = i; - byte weight = convertLiangToWeight(interletterValue); - values[k] = weight; - } - } - return createHyphenation(theWord, result, values, k); + hyphenation = checkAlgorithm(theWord, remainCharCount, pushCharCount, + word); + return hyphenation; } /** @@ -416,6 +392,36 @@ return createHyphenation(sw, result, values, k); } + private Hyphenation checkAlgorithm(String theWord, int remainCharCount, + int pushCharCount, char[] word) { + int len = theWord.length(); + // use algorithm to get hyphenation points + int[] result = new int[len + 1]; + byte[] values = new byte[result.length]; + int k = 0; + word[0] = '.'; // word start marker + word[len + 1] = '.'; // word end marker + word[len + 2] = 0; // null terminated + byte[] il = new byte[len + 3]; // initialized to zero + for (int i = 0; i < len + 1; i++) { + searchPatterns(word, i, il); + } + + /* We report all of the non-zero values. */ + for (int i = 0; i < len; i++) { + int interletterValue = il[i + 1]; + if (interletterValue != 0 + && i >= remainCharCount + && i <= (len - pushCharCount)) { + k++; + result[k] = i; + byte weight = convertLiangToWeight(interletterValue); + values[k] = weight; + } + } + return createHyphenation(theWord, result, values, k); + } + /** * Add a character class to the tree. It is used by * {@link PatternParser PatternParser} as callback to This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site. |
