commit e07baa6fabe1be1aef68caa9a55a073d4f118334
author satok <satok@google.com> Thu Dec 09 21:55:40 2010 +0900
committer satok <satok@google.com> Fri Dec 10 00:47:37 2010 +0900
tree 473a58e44f31ccb1bb1958163fc75b18f8154d5a
parent 8db866aaa26545d698ccf5acdc4d12d65588edec

Limit the suggestions with an excessive character by filtering proximity characters

Change-Id: Iad26dad545f1a431aa0fa53f99198b27defd03a3
ug: 3269482

native/src/unigram_dictionary.cpp

diff --git a/native/src/unigram_dictionary.cpp b/native/src/unigram_dictionary.cpp
index 55d879f..46332c7 100644
--- a/native/src/unigram_dictionary.cpp
+++ b/native/src/unigram_dictionary.cpp
@@ -58,8 +58,10 @@
     // Suggestion with excessive character
     if (SUGGEST_WORDS_WITH_EXCESSIVE_CHARACTER) {
         for (int i = 0; i < codesSize; ++i) {
-            if (DEBUG_DICT) LOGI("--- Suggest excessive characters %d", i);
-            getSuggestionCandidates(codesSize, -1, i, NULL, 0);
+            if (existsAdjacentProximityChars(i, codesSize)) {
+                if (DEBUG_DICT) LOGI("--- Suggest excessive characters %d", i);
+                getSuggestionCandidates(codesSize, -1, i, NULL, 0);
+            }
         }
     }
 
@@ -331,11 +333,34 @@
     return (c == QUOTE && userTypedChar != QUOTE) || skipPos == depth;
 }
 
+inline bool UnigramDictionary::existsAdjacentProximityChars(const int inputIndex,
+        const int inputLength) {
+    if (inputIndex < 0 || inputIndex >= inputLength) return false;
+    const int currentChar = *getInputCharsAt(inputIndex);
+    const int leftIndex = inputIndex - 1;
+    if (leftIndex >= 0) {
+        int *leftChars = getInputCharsAt(leftIndex);
+        int i = 0;
+        while (leftChars[i] > 0 && i < MAX_PROXIMITY_CHARS) {
+            if (leftChars[i++] == currentChar) return true;
+        }
+    }
+    const int rightIndex = inputIndex + 1;
+    if (rightIndex < inputLength) {
+        int *rightChars = getInputCharsAt(rightIndex);
+        int i = 0;
+        while (rightChars[i] > 0 && i < MAX_PROXIMITY_CHARS) {
+            if (rightChars[i++] == currentChar) return true;
+        }
+    }
+    return false;
+}
+
 inline int UnigramDictionary::getMatchedProximityId(const int *currentChars,
         const unsigned short c, const int skipPos) {
     const unsigned short lowerC = toLowerCase(c);
     int j = 0;
-    while (currentChars[j] > 0) {
+    while (currentChars[j] > 0 && j < MAX_PROXIMITY_CHARS) {
         const bool matched = (currentChars[j] == lowerC || currentChars[j] == c);
         // If skipPos is defined, not to search proximity collections.
         // First char is what user typed.