aboutsummaryrefslogtreecommitdiffstats
path: root/java/src/com/android/inputmethod/latin
diff options
context:
space:
mode:
Diffstat (limited to 'java/src/com/android/inputmethod/latin')
-rw-r--r--java/src/com/android/inputmethod/latin/BinaryDictionary.java50
-rw-r--r--java/src/com/android/inputmethod/latin/utils/CollectionUtils.java15
-rw-r--r--java/src/com/android/inputmethod/latin/utils/StringUtils.java18
-rw-r--r--java/src/com/android/inputmethod/latin/utils/UserHistoryDictIOUtils.java4
-rw-r--r--java/src/com/android/inputmethod/latin/utils/WordProperty.java (renamed from java/src/com/android/inputmethod/latin/utils/UnigramProperty.java)6
5 files changed, 63 insertions, 30 deletions
diff --git a/java/src/com/android/inputmethod/latin/BinaryDictionary.java b/java/src/com/android/inputmethod/latin/BinaryDictionary.java
index e66cfca49..95823dac5 100644
--- a/java/src/com/android/inputmethod/latin/BinaryDictionary.java
+++ b/java/src/com/android/inputmethod/latin/BinaryDictionary.java
@@ -27,7 +27,7 @@ import com.android.inputmethod.latin.utils.CollectionUtils;
import com.android.inputmethod.latin.utils.JniUtils;
import com.android.inputmethod.latin.utils.LanguageModelParam;
import com.android.inputmethod.latin.utils.StringUtils;
-import com.android.inputmethod.latin.utils.UnigramProperty;
+import com.android.inputmethod.latin.utils.WordProperty;
import java.io.File;
import java.util.ArrayList;
@@ -61,18 +61,18 @@ public final class BinaryDictionary extends Dictionary {
public static final int NOT_A_VALID_TIMESTAMP = -1;
- // Format to get unigram flags from native side via getUnigramPropertyNative().
- private static final int FORMAT_UNIGRAM_PROPERTY_OUTPUT_FLAG_COUNT = 4;
- private static final int FORMAT_UNIGRAM_PROPERTY_IS_NOT_A_WORD_INDEX = 0;
- private static final int FORMAT_UNIGRAM_PROPERTY_IS_BLACKLISTED_INDEX = 1;
- private static final int FORMAT_UNIGRAM_PROPERTY_HAS_BIGRAMS_INDEX = 2;
- private static final int FORMAT_UNIGRAM_PROPERTY_HAS_SHORTCUTS_INDEX = 3;
+ // Format to get unigram flags from native side via getWordPropertyNative().
+ private static final int FORMAT_WORD_PROPERTY_OUTPUT_FLAG_COUNT = 4;
+ private static final int FORMAT_WORD_PROPERTY_IS_NOT_A_WORD_INDEX = 0;
+ private static final int FORMAT_WORD_PROPERTY_IS_BLACKLISTED_INDEX = 1;
+ private static final int FORMAT_WORD_PROPERTY_HAS_BIGRAMS_INDEX = 2;
+ private static final int FORMAT_WORD_PROPERTY_HAS_SHORTCUTS_INDEX = 3;
- // Format to get unigram historical info from native side via getUnigramPropertyNative().
- private static final int FORMAT_UNIGRAM_PROPERTY_OUTPUT_HISTORICAL_INFO_COUNT = 3;
- private static final int FORMAT_UNIGRAM_PROPERTY_TIMESTAMP_INDEX = 0;
- private static final int FORMAT_UNIGRAM_PROPERTY_LEVEL_INDEX = 1;
- private static final int FORMAT_UNIGRAM_PROPERTY_COUNT_INDEX = 2;
+ // Format to get unigram historical info from native side via getWordPropertyNative().
+ private static final int FORMAT_WORD_PROPERTY_OUTPUT_HISTORICAL_INFO_COUNT = 3;
+ private static final int FORMAT_WORD_PROPERTY_TIMESTAMP_INDEX = 0;
+ private static final int FORMAT_WORD_PROPERTY_LEVEL_INDEX = 1;
+ private static final int FORMAT_WORD_PROPERTY_COUNT_INDEX = 2;
private long mNativeDict;
private final Locale mLocale;
@@ -143,7 +143,7 @@ public final class BinaryDictionary extends Dictionary {
private static native int getFormatVersionNative(long dict);
private static native int getProbabilityNative(long dict, int[] word);
private static native int getBigramProbabilityNative(long dict, int[] word0, int[] word1);
- private static native void getUnigramPropertyNative(long dict, int[] word,
+ private static native void getWordPropertyNative(long dict, int[] word,
int[] outCodePoints, boolean[] outFlags, int[] outProbability,
int[] outHistoricalInfo, ArrayList<int[]> outShortcutTargets,
ArrayList<Integer> outShortcutProbabilities);
@@ -306,28 +306,28 @@ public final class BinaryDictionary extends Dictionary {
}
@UsedForTesting
- public UnigramProperty getUnigramProperty(final String word) {
+ public WordProperty getWordProperty(final String word) {
if (TextUtils.isEmpty(word)) {
return null;
}
final int[] codePoints = StringUtils.toCodePointArray(word);
final int[] outCodePoints = new int[MAX_WORD_LENGTH];
- final boolean[] outFlags = new boolean[FORMAT_UNIGRAM_PROPERTY_OUTPUT_FLAG_COUNT];
+ final boolean[] outFlags = new boolean[FORMAT_WORD_PROPERTY_OUTPUT_FLAG_COUNT];
final int[] outProbability = new int[1];
final int[] outHistoricalInfo =
- new int[FORMAT_UNIGRAM_PROPERTY_OUTPUT_HISTORICAL_INFO_COUNT];
+ new int[FORMAT_WORD_PROPERTY_OUTPUT_HISTORICAL_INFO_COUNT];
final ArrayList<int[]> outShortcutTargets = CollectionUtils.newArrayList();
final ArrayList<Integer> outShortcutProbabilities = CollectionUtils.newArrayList();
- getUnigramPropertyNative(mNativeDict, codePoints, outCodePoints, outFlags, outProbability,
+ getWordPropertyNative(mNativeDict, codePoints, outCodePoints, outFlags, outProbability,
outHistoricalInfo, outShortcutTargets, outShortcutProbabilities);
- return new UnigramProperty(codePoints,
- outFlags[FORMAT_UNIGRAM_PROPERTY_IS_NOT_A_WORD_INDEX],
- outFlags[FORMAT_UNIGRAM_PROPERTY_IS_BLACKLISTED_INDEX],
- outFlags[FORMAT_UNIGRAM_PROPERTY_HAS_BIGRAMS_INDEX],
- outFlags[FORMAT_UNIGRAM_PROPERTY_HAS_SHORTCUTS_INDEX], outProbability[0],
- outHistoricalInfo[FORMAT_UNIGRAM_PROPERTY_TIMESTAMP_INDEX],
- outHistoricalInfo[FORMAT_UNIGRAM_PROPERTY_LEVEL_INDEX],
- outHistoricalInfo[FORMAT_UNIGRAM_PROPERTY_COUNT_INDEX],
+ return new WordProperty(codePoints,
+ outFlags[FORMAT_WORD_PROPERTY_IS_NOT_A_WORD_INDEX],
+ outFlags[FORMAT_WORD_PROPERTY_IS_BLACKLISTED_INDEX],
+ outFlags[FORMAT_WORD_PROPERTY_HAS_BIGRAMS_INDEX],
+ outFlags[FORMAT_WORD_PROPERTY_HAS_SHORTCUTS_INDEX], outProbability[0],
+ outHistoricalInfo[FORMAT_WORD_PROPERTY_TIMESTAMP_INDEX],
+ outHistoricalInfo[FORMAT_WORD_PROPERTY_LEVEL_INDEX],
+ outHistoricalInfo[FORMAT_WORD_PROPERTY_COUNT_INDEX],
outShortcutTargets, outShortcutProbabilities);
}
diff --git a/java/src/com/android/inputmethod/latin/utils/CollectionUtils.java b/java/src/com/android/inputmethod/latin/utils/CollectionUtils.java
index cc25102ce..bbfa0f091 100644
--- a/java/src/com/android/inputmethod/latin/utils/CollectionUtils.java
+++ b/java/src/com/android/inputmethod/latin/utils/CollectionUtils.java
@@ -102,4 +102,19 @@ public final class CollectionUtils {
public static <E> SparseArray<E> newSparseArray() {
return new SparseArray<E>();
}
+
+ public static <E> ArrayList<E> arrayAsList(final E[] array, final int start, final int end) {
+ if (array == null) {
+ throw new NullPointerException();
+ }
+ if (start < 0 || start > end || end > array.length) {
+ throw new IllegalArgumentException();
+ }
+
+ final ArrayList<E> list = newArrayList(end - start);
+ for (int i = start; i < end; i++) {
+ list.add(array[i]);
+ }
+ return list;
+ }
}
diff --git a/java/src/com/android/inputmethod/latin/utils/StringUtils.java b/java/src/com/android/inputmethod/latin/utils/StringUtils.java
index b154623ae..c632a71a9 100644
--- a/java/src/com/android/inputmethod/latin/utils/StringUtils.java
+++ b/java/src/com/android/inputmethod/latin/utils/StringUtils.java
@@ -16,6 +16,8 @@
package com.android.inputmethod.latin.utils;
+import static com.android.inputmethod.latin.Constants.CODE_UNSPECIFIED;
+
import android.text.TextUtils;
import com.android.inputmethod.annotations.UsedForTesting;
@@ -471,4 +473,20 @@ public final class StringUtils {
}
return bytes;
}
+
+ public static String toUpperCaseOfStringForLocale(final String text,
+ final boolean needsToUpperCase, final Locale locale) {
+ if (text == null || !needsToUpperCase) return text;
+ return text.toUpperCase(locale);
+ }
+
+ public static int toUpperCaseOfCodeForLocale(final int code, final boolean needsToUpperCase,
+ final Locale locale) {
+ if (!Constants.isLetterCode(code) || !needsToUpperCase) return code;
+ final String text = newSingleCodePointString(code);
+ final String casedText = toUpperCaseOfStringForLocale(
+ text, needsToUpperCase, locale);
+ return codePointCount(casedText) == 1
+ ? casedText.codePointAt(0) : CODE_UNSPECIFIED;
+ }
}
diff --git a/java/src/com/android/inputmethod/latin/utils/UserHistoryDictIOUtils.java b/java/src/com/android/inputmethod/latin/utils/UserHistoryDictIOUtils.java
index db628fe18..a6ae640d3 100644
--- a/java/src/com/android/inputmethod/latin/utils/UserHistoryDictIOUtils.java
+++ b/java/src/com/android/inputmethod/latin/utils/UserHistoryDictIOUtils.java
@@ -46,7 +46,7 @@ public final class UserHistoryDictIOUtils {
private static final boolean DEBUG = false;
private static final String USES_FORGETTING_CURVE_KEY = "USES_FORGETTING_CURVE";
private static final String USES_FORGETTING_CURVE_VALUE = "1";
- private static final String LAST_UPDATED_TIME_KEY = "date";
+ private static final String DATE_KEY = "date";
public interface OnAddWordListener {
/**
@@ -76,7 +76,7 @@ public final class UserHistoryDictIOUtils {
final FormatOptions formatOptions, final HashMap<String, String> options) {
final FusionDictionary fusionDict = constructFusionDictionary(dict, bigrams, options);
fusionDict.addOptionAttribute(USES_FORGETTING_CURVE_KEY, USES_FORGETTING_CURVE_VALUE);
- fusionDict.addOptionAttribute(LAST_UPDATED_TIME_KEY,
+ fusionDict.addOptionAttribute(DATE_KEY,
String.valueOf(TimeUnit.MILLISECONDS.toSeconds(System.currentTimeMillis())));
try {
dictEncoder.writeDictionary(fusionDict, formatOptions);
diff --git a/java/src/com/android/inputmethod/latin/utils/UnigramProperty.java b/java/src/com/android/inputmethod/latin/utils/WordProperty.java
index 4feee4393..d6c0f900a 100644
--- a/java/src/com/android/inputmethod/latin/utils/UnigramProperty.java
+++ b/java/src/com/android/inputmethod/latin/utils/WordProperty.java
@@ -26,7 +26,7 @@ import java.util.ArrayList;
// This has information that belong to a unigram. This class has some detailed attributes such as
// historical information but they have to be checked only for testing purpose.
@UsedForTesting
-public class UnigramProperty {
+public class WordProperty {
public final String mCodePoints;
public final boolean mIsNotAWord;
public final boolean mIsBlacklisted;
@@ -50,8 +50,8 @@ public class UnigramProperty {
return codePoints.length;
}
- // This represents invalid unigram when the probability is BinaryDictionary.NOT_A_PROBABILITY.
- public UnigramProperty(final int[] codePoints, final boolean isNotAWord,
+ // This represents invalid word when the probability is BinaryDictionary.NOT_A_PROBABILITY.
+ public WordProperty(final int[] codePoints, final boolean isNotAWord,
final boolean isBlacklisted, final boolean hasBigram,
final boolean hasShortcuts, final int probability, final int timestamp,
final int level, final int count, final ArrayList<int[]> shortcutTargets,
generated by cgit v1.2.3-83-g751a (git 2.47.1) at 2025-05-10 00:00:53 +0000