1 files changed, 17 insertions, 3 deletions
diff --git a/java/src/com/android/inputmethod/research/MainLogBuffer.java b/java/src/com/android/inputmethod/research/MainLogBuffer.java
index 745768d35..0185e5fc0 100644
--- a/java/src/com/android/inputmethod/research/MainLogBuffer.java
+++ b/java/src/com/android/inputmethod/research/MainLogBuffer.java
@@ -16,16 +16,23 @@
 
 package com.android.inputmethod.research;
 
+import android.util.Log;
+
 import com.android.inputmethod.latin.Dictionary;
 import com.android.inputmethod.latin.Suggest;
+import com.android.inputmethod.latin.define.ProductionFlag;
 
 import java.util.Random;
 
 public class MainLogBuffer extends LogBuffer {
+    private static final String TAG = MainLogBuffer.class.getSimpleName();
+    private static final boolean DEBUG = false && ProductionFlag.IS_EXPERIMENTAL_DEBUG;
+
     // The size of the n-grams logged.  E.g. N_GRAM_SIZE = 2 means to sample bigrams.
     private static final int N_GRAM_SIZE = 2;
     // The number of words between n-grams to omit from the log.
-    private static final int DEFAULT_NUMBER_OF_WORDS_BETWEEN_SAMPLES = 18;
+    private static final int DEFAULT_NUMBER_OF_WORDS_BETWEEN_SAMPLES =
+            ProductionFlag.IS_EXPERIMENTAL_DEBUG ? 2 : 18;
 
     private final ResearchLog mResearchLog;
     private Suggest mSuggest;
@@ -61,6 +68,9 @@ public class MainLogBuffer extends LogBuffer {
                 mWordsUntilSafeToSample--;
             }
         }
+        if (DEBUG) {
+            Log.d(TAG, "shiftedIn " + (newLogUnit.hasWord() ? newLogUnit.getWord() : ""));
+        }
     }
 
     public void resetWordCounter() {
@@ -104,12 +114,16 @@ public class MainLogBuffer extends LogBuffer {
             final String word = logUnit.getWord();
             if (word == null) {
                 // Digits outside words are a privacy threat.
-                if (logUnit.hasDigit()) {
+                if (logUnit.mayContainDigit()) {
                     return false;
                 }
             } else {
                 // Words not in the dictionary are a privacy threat.
-                if (!(dictionary.isValidWord(word))) {
+                if (ResearchLogger.hasLetters(word) && !(dictionary.isValidWord(word))) {
+                    if (DEBUG) {
+                        Log.d(TAG, "NOT SAFE!: hasLetters: " + ResearchLogger.hasLetters(word)
+                                + ", isValid: " + (dictionary.isValidWord(word)));
+                    }
                     return false;
                 }
             }