[Rlog81a] Determine correction type of words
Heuristic to determine whether a word was a typo correction or a complete
replacement by examining a correction to see if it falls within the list of
suggested words of the original.
Change-Id: Ieec4861a811e96aef0d14622e662b589ef8b4772
diff --git a/java/src/com/android/inputmethod/latin/LatinIME.java b/java/src/com/android/inputmethod/latin/LatinIME.java
index a48778a..6085cb4 100644
--- a/java/src/com/android/inputmethod/latin/LatinIME.java
+++ b/java/src/com/android/inputmethod/latin/LatinIME.java
@@ -1647,7 +1647,7 @@
mExpectingUpdateSelection = true;
mConnection.endBatchEdit();
if (ProductionFlag.IS_EXPERIMENTAL) {
- ResearchLogger.latinIME_onEndBatchInput(batchInputText, 0);
+ ResearchLogger.latinIME_onEndBatchInput(batchInputText, 0, suggestedWords);
}
// Space state must be updated before calling updateShiftState
mSpaceState = SPACE_STATE_PHANTOM;
@@ -2123,8 +2123,9 @@
Stats.onAutoCorrection(typedWord, autoCorrection, separatorString, mWordComposer);
}
if (ProductionFlag.IS_EXPERIMENTAL) {
+ final SuggestedWords suggestedWords = mSuggestionStripView.getSuggestions();
ResearchLogger.latinIme_commitCurrentAutoCorrection(typedWord, autoCorrection,
- separatorString, mWordComposer.isBatchMode());
+ separatorString, mWordComposer.isBatchMode(), suggestedWords);
}
mExpectingUpdateSelection = true;
commitChosenWord(autoCorrection, LastComposedWord.COMMIT_TYPE_DECIDED_WORD,
diff --git a/java/src/com/android/inputmethod/research/LogUnit.java b/java/src/com/android/inputmethod/research/LogUnit.java
index 715000d..0234bbc 100644
--- a/java/src/com/android/inputmethod/research/LogUnit.java
+++ b/java/src/com/android/inputmethod/research/LogUnit.java
@@ -17,6 +17,7 @@
package com.android.inputmethod.research;
import android.content.SharedPreferences;
+import android.text.TextUtils;
import android.util.JsonWriter;
import android.util.Log;
import android.view.MotionEvent;
@@ -57,16 +58,32 @@
// Assume that mTimeList is sorted in increasing order. Do not insert null values into
// mTimeList.
private final ArrayList<Long> mTimeList;
+ // Word that this LogUnit generates. Should be null if the LogUnit does not generate a genuine
+ // word (i.e. separators alone do not count as a word). Should never be empty.
private String mWord;
private boolean mMayContainDigit;
private boolean mIsPartOfMegaword;
private boolean mContainsCorrection;
+ // mCorrectionType indicates whether the word was corrected at all, and if so, whether it was
+ // to a different word or just a "typo" correction. It is considered a "typo" if the final
+ // word was listed in the suggestions available the first time the word was gestured or
+ // tapped.
+ private int mCorrectionType;
+ public static final int CORRECTIONTYPE_NO_CORRECTION = 0;
+ public static final int CORRECTIONTYPE_CORRECTION = 1;
+ public static final int CORRECTIONTYPE_DIFFERENT_WORD = 2;
+ public static final int CORRECTIONTYPE_TYPO = 3;
+
+ private SuggestedWords mSuggestedWords;
+
public LogUnit() {
mLogStatementList = new ArrayList<LogStatement>();
mValuesList = new ArrayList<Object[]>();
mTimeList = new ArrayList<Long>();
mIsPartOfMegaword = false;
+ mCorrectionType = CORRECTIONTYPE_NO_CORRECTION;
+ mSuggestedWords = null;
}
private LogUnit(final ArrayList<LogStatement> logStatementList,
@@ -77,6 +94,8 @@
mValuesList = valuesList;
mTimeList = timeList;
mIsPartOfMegaword = isPartOfMegaword;
+ mCorrectionType = CORRECTIONTYPE_NO_CORRECTION;
+ mSuggestedWords = null;
}
private static final Object[] NULL_VALUES = new Object[0];
@@ -167,6 +186,7 @@
private static final String UPTIME_KEY = "_ut";
private static final String EVENT_TYPE_KEY = "_ty";
private static final String WORD_KEY = "_wo";
+ private static final String CORRECTION_TYPE_KEY = "_corType";
private static final String LOG_UNIT_BEGIN_KEY = "logUnitStart";
private static final String LOG_UNIT_END_KEY = "logUnitEnd";
@@ -177,6 +197,7 @@
jsonWriter.name(CURRENT_TIME_KEY).value(System.currentTimeMillis());
if (canIncludePrivateData) {
jsonWriter.name(WORD_KEY).value(getWord());
+ jsonWriter.name(CORRECTION_TYPE_KEY).value(getCorrectionType());
}
jsonWriter.name(EVENT_TYPE_KEY).value(LOG_UNIT_BEGIN_KEY);
jsonWriter.endObject();
@@ -254,7 +275,33 @@
return true;
}
- public void setWord(String word) {
+ /**
+ * Mark the current logUnit as containing data to generate {@code word}.
+ *
+ * If {@code setWord()} was previously called for this LogUnit, then the method will try to
+ * determine what kind of correction it is, and update its internal state of the correctionType
+ * accordingly.
+ *
+ * @param word The word this LogUnit generates. Caller should not pass null or the empty
+ * string.
+ */
+ public void setWord(final String word) {
+ if (mWord != null) {
+ // The word was already set once, and it is now being changed. See if the new word
+ // is close to the old word. If so, then the change is probably a typo correction.
+ // If not, the user may have decided to enter a different word, so flag it.
+ if (mSuggestedWords != null) {
+ if (isInSuggestedWords(word, mSuggestedWords)) {
+ mCorrectionType = CORRECTIONTYPE_TYPO;
+ } else {
+ mCorrectionType = CORRECTIONTYPE_DIFFERENT_WORD;
+ }
+ } else {
+ // No suggested words, so it's not clear whether it's a typo or different word.
+ // Mark it as a generic correction.
+ mCorrectionType = CORRECTIONTYPE_CORRECTION;
+ }
+ }
mWord = word;
}
@@ -282,6 +329,14 @@
return mContainsCorrection;
}
+ public void setCorrectionType(final int correctionType) {
+ mCorrectionType = correctionType;
+ }
+
+ public int getCorrectionType() {
+ return mCorrectionType;
+ }
+
public boolean isEmpty() {
return mLogStatementList.isEmpty();
}
@@ -328,8 +383,43 @@
mValuesList.addAll(logUnit.mValuesList);
mTimeList.addAll(logUnit.mTimeList);
mWord = null;
+ if (logUnit.mWord != null) {
+ setWord(logUnit.mWord);
+ }
mMayContainDigit = mMayContainDigit || logUnit.mMayContainDigit;
mContainsCorrection = mContainsCorrection || logUnit.mContainsCorrection;
mIsPartOfMegaword = false;
}
+
+ public SuggestedWords getSuggestions() {
+ return mSuggestedWords;
+ }
+
+ /**
+ * Initialize the suggestions.
+ *
+ * Once set to a non-null value, the suggestions may not be changed again. This is to keep
+ * track of the list of words that are close to the user's initial effort to type the word.
+ * Only words that are close to the initial effort are considered typo corrections.
+ */
+ public void initializeSuggestions(final SuggestedWords suggestedWords) {
+ if (mSuggestedWords == null) {
+ mSuggestedWords = suggestedWords;
+ }
+ }
+
+ private static boolean isInSuggestedWords(final String queryWord,
+ final SuggestedWords suggestedWords) {
+ if (TextUtils.isEmpty(queryWord)) {
+ return false;
+ }
+ final int size = suggestedWords.size();
+ for (int i = 0; i < size; i++) {
+ final SuggestedWordInfo wordInfo = suggestedWords.getInfo(i);
+ if (queryWord.equals(wordInfo.mWord)) {
+ return true;
+ }
+ }
+ return false;
+ }
}
diff --git a/java/src/com/android/inputmethod/research/ResearchLogger.java b/java/src/com/android/inputmethod/research/ResearchLogger.java
index 0a24af6..29bc708 100644
--- a/java/src/com/android/inputmethod/research/ResearchLogger.java
+++ b/java/src/com/android/inputmethod/research/ResearchLogger.java
@@ -745,6 +745,10 @@
mCurrentLogUnit.setContainsCorrection();
}
+ private void setCurrentLogUnitCorrectionType(final int correctionType) {
+ mCurrentLogUnit.setCorrectionType(correctionType);
+ }
+
/* package for test */ void commitCurrentLogUnit() {
if (DEBUG) {
Log.d(TAG, "commitCurrentLogUnit" + (mCurrentLogUnit.hasWord() ?
@@ -1194,13 +1198,17 @@
"suggestion", "x", "y");
public static void latinIME_pickSuggestionManually(final String replacedWord,
final int index, final String suggestion, final boolean isBatchMode) {
- final String scrubbedWord = scrubDigitsFromString(suggestion);
final ResearchLogger researchLogger = getInstance();
+ if (!replacedWord.equals(suggestion.toString())) {
+ // The user choose something other than what was already there.
+ researchLogger.setCurrentLogUnitContainsCorrection();
+ researchLogger.setCurrentLogUnitCorrectionType(LogUnit.CORRECTIONTYPE_TYPO);
+ }
+ final String scrubbedWord = scrubDigitsFromString(suggestion);
researchLogger.enqueueEvent(LOGSTATEMENT_LATINIME_PICKSUGGESTIONMANUALLY,
scrubDigitsFromString(replacedWord), index,
suggestion == null ? null : scrubbedWord, Constants.SUGGESTION_STRIP_COORDINATE,
Constants.SUGGESTION_STRIP_COORDINATE);
- researchLogger.setCurrentLogUnitContainsCorrection();
researchLogger.commitCurrentLogUnitAsWord(scrubbedWord, Long.MAX_VALUE, isBatchMode);
researchLogger.mStatistics.recordManualSuggestion(SystemClock.uptimeMillis());
}
@@ -1490,10 +1498,12 @@
new LogStatement("LatinIMECommitCurrentAutoCorrection", true, true, "typedWord",
"autoCorrection", "separatorString");
public static void latinIme_commitCurrentAutoCorrection(final String typedWord,
- final String autoCorrection, final String separatorString, final boolean isBatchMode) {
+ final String autoCorrection, final String separatorString, final boolean isBatchMode,
+ final SuggestedWords suggestedWords) {
final String scrubbedTypedWord = scrubDigitsFromString(typedWord);
final String scrubbedAutoCorrection = scrubDigitsFromString(autoCorrection);
final ResearchLogger researchLogger = getInstance();
+ researchLogger.mCurrentLogUnit.initializeSuggestions(suggestedWords);
researchLogger.commitCurrentLogUnitAsWord(scrubbedAutoCorrection, Long.MAX_VALUE,
isBatchMode);
@@ -1691,10 +1701,11 @@
new LogStatement("LatinIMEOnEndBatchInput", true, false, "enteredText",
"enteredWordPos");
public static void latinIME_onEndBatchInput(final CharSequence enteredText,
- final int enteredWordPos) {
+ final int enteredWordPos, final SuggestedWords suggestedWords) {
final ResearchLogger researchLogger = getInstance();
researchLogger.enqueueEvent(LOGSTATEMENT_LATINIME_ONENDBATCHINPUT, enteredText,
enteredWordPos);
+ researchLogger.mCurrentLogUnit.initializeSuggestions(suggestedWords);
researchLogger.mStatistics.recordGestureInput(enteredText.length(),
SystemClock.uptimeMillis());
}