Merge "Quit auto-correct explicit accented letters to base letters."

5 files changed, 30 insertions, 8 deletions

diff --git a/native/jni/src/suggest/core/dictionary/dictionary.h b/native/jni/src/suggest/core/dictionary/dictionary.h
index a5e986d15..b1774371b 100644
--- a/native/jni/src/suggest/core/dictionary/dictionary.h
+++ b/native/jni/src/suggest/core/dictionary/dictionary.h
@@ -60,6 +60,7 @@ class Dictionary {
     static const int KIND_FLAG_POSSIBLY_OFFENSIVE = 0x80000000;
     static const int KIND_FLAG_EXACT_MATCH = 0x40000000;
     static const int KIND_FLAG_EXACT_MATCH_WITH_INTENTIONAL_OMISSION = 0x20000000;
+    static const int KIND_FLAG_APPROPRIATE_FOR_AUTOCORRECTION = 0x10000000;
 
     Dictionary(JNIEnv *env, DictionaryStructureWithBufferPolicy::StructurePolicyPtr
             dictionaryStructureWithBufferPolicy);
diff --git a/native/jni/src/suggest/core/dictionary/error_type_utils.cpp b/native/jni/src/suggest/core/dictionary/error_type_utils.cpp
index 8f07ce275..61093e174 100644
--- a/native/jni/src/suggest/core/dictionary/error_type_utils.cpp
+++ b/native/jni/src/suggest/core/dictionary/error_type_utils.cpp
@@ -21,13 +21,14 @@ namespace latinime {
 const ErrorTypeUtils::ErrorType ErrorTypeUtils::NOT_AN_ERROR = 0x0;
 const ErrorTypeUtils::ErrorType ErrorTypeUtils::MATCH_WITH_WRONG_CASE = 0x1;
 const ErrorTypeUtils::ErrorType ErrorTypeUtils::MATCH_WITH_MISSING_ACCENT = 0x2;
-const ErrorTypeUtils::ErrorType ErrorTypeUtils::MATCH_WITH_WRONG_ACCENT = 0x4;
-const ErrorTypeUtils::ErrorType ErrorTypeUtils::MATCH_WITH_DIGRAPH = 0x8;
-const ErrorTypeUtils::ErrorType ErrorTypeUtils::INTENTIONAL_OMISSION = 0x10;
-const ErrorTypeUtils::ErrorType ErrorTypeUtils::EDIT_CORRECTION = 0x20;
-const ErrorTypeUtils::ErrorType ErrorTypeUtils::PROXIMITY_CORRECTION = 0x40;
-const ErrorTypeUtils::ErrorType ErrorTypeUtils::COMPLETION = 0x80;
-const ErrorTypeUtils::ErrorType ErrorTypeUtils::NEW_WORD = 0x100;
+const ErrorTypeUtils::ErrorType ErrorTypeUtils::MATCH_WITH_MISSING_EXPLICIT_ACCENT = 0x4;
+const ErrorTypeUtils::ErrorType ErrorTypeUtils::MATCH_WITH_WRONG_ACCENT = 0x8;
+const ErrorTypeUtils::ErrorType ErrorTypeUtils::MATCH_WITH_DIGRAPH = 0x10;
+const ErrorTypeUtils::ErrorType ErrorTypeUtils::INTENTIONAL_OMISSION = 0x20;
+const ErrorTypeUtils::ErrorType ErrorTypeUtils::EDIT_CORRECTION = 0x40;
+const ErrorTypeUtils::ErrorType ErrorTypeUtils::PROXIMITY_CORRECTION = 0x80;
+const ErrorTypeUtils::ErrorType ErrorTypeUtils::COMPLETION = 0x100;
+const ErrorTypeUtils::ErrorType ErrorTypeUtils::NEW_WORD = 0x200;
 
 const ErrorTypeUtils::ErrorType ErrorTypeUtils::ERRORS_TREATED_AS_AN_EXACT_MATCH =
         NOT_AN_ERROR | MATCH_WITH_WRONG_CASE | MATCH_WITH_MISSING_ACCENT | MATCH_WITH_DIGRAPH;
diff --git a/native/jni/src/suggest/core/dictionary/error_type_utils.h b/native/jni/src/suggest/core/dictionary/error_type_utils.h
index e92c509fa..75111ba75 100644
--- a/native/jni/src/suggest/core/dictionary/error_type_utils.h
+++ b/native/jni/src/suggest/core/dictionary/error_type_utils.h
@@ -32,6 +32,7 @@ class ErrorTypeUtils {
     static const ErrorType NOT_AN_ERROR;
     static const ErrorType MATCH_WITH_WRONG_CASE;
     static const ErrorType MATCH_WITH_MISSING_ACCENT;
+    static const ErrorType MATCH_WITH_MISSING_EXPLICIT_ACCENT;
     static const ErrorType MATCH_WITH_WRONG_ACCENT;
     static const ErrorType MATCH_WITH_DIGRAPH;
     // Treat error as an intentional omission when the CorrectionType is omission and the node can
@@ -61,6 +62,10 @@ class ErrorTypeUtils {
                 & ~ERRORS_TREATED_AS_AN_EXACT_MATCH_WITH_INTENTIONAL_OMISSION) == 0;
     }
 
+    static bool isMissingExplicitAccent(const ErrorType errorType) {
+        return (errorType & MATCH_WITH_MISSING_EXPLICIT_ACCENT) != 0;
+    }
+
     static bool isEditCorrectionError(const ErrorType errorType) {
         return (errorType & EDIT_CORRECTION) != 0;
     }
diff --git a/native/jni/src/suggest/core/result/suggestions_output_utils.cpp b/native/jni/src/suggest/core/result/suggestions_output_utils.cpp
index 74db95953..1aff72952 100644
--- a/native/jni/src/suggest/core/result/suggestions_output_utils.cpp
+++ b/native/jni/src/suggest/core/result/suggestions_output_utils.cpp
@@ -144,11 +144,16 @@ const int SuggestionsOutputUtils::MIN_LEN_FOR_MULTI_WORD_AUTOCORRECT = 16;
     const bool isExactMatchWithIntentionalOmission =
             ErrorTypeUtils::isExactMatchWithIntentionalOmission(
                     terminalDicNode->getContainedErrorTypes());
+    // TODO: Decide whether the word should be auto-corrected or not here.
+    const bool isAppropriateForAutoCorrection = !ErrorTypeUtils::isMissingExplicitAccent(
+            terminalDicNode->getContainedErrorTypes());
     const int outputTypeFlags =
             (wordAttributes.isPossiblyOffensive() ? Dictionary::KIND_FLAG_POSSIBLY_OFFENSIVE : 0)
             | ((isExactMatch && boostExactMatches) ? Dictionary::KIND_FLAG_EXACT_MATCH : 0)
             | (isExactMatchWithIntentionalOmission ?
-                    Dictionary::KIND_FLAG_EXACT_MATCH_WITH_INTENTIONAL_OMISSION : 0);
+                    Dictionary::KIND_FLAG_EXACT_MATCH_WITH_INTENTIONAL_OMISSION : 0)
+            | (isAppropriateForAutoCorrection ?
+                    Dictionary::KIND_FLAG_APPROPRIATE_FOR_AUTOCORRECTION : 0);
     // Entries that are blacklisted or do not represent a word should not be output.
     const bool isValidWord = !(wordAttributes.isBlacklisted() || wordAttributes.isNotAWord());
 
diff --git a/native/jni/src/suggest/policyimpl/typing/typing_weighting.cpp b/native/jni/src/suggest/policyimpl/typing/typing_weighting.cpp
index db7a39efb..a0e54115d 100644
--- a/native/jni/src/suggest/policyimpl/typing/typing_weighting.cpp
+++ b/native/jni/src/suggest/policyimpl/typing/typing_weighting.cpp
@@ -17,6 +17,7 @@
 #include "suggest/policyimpl/typing/typing_weighting.h"
 
 #include "suggest/core/dicnode/dic_node.h"
+#include "suggest/core/layout/proximity_info.h"
 #include "suggest/policyimpl/typing/scoring_params.h"
 
 namespace latinime {
@@ -39,6 +40,8 @@ ErrorTypeUtils::ErrorType TypingWeighting::getErrorType(const CorrectionType cor
                 const int primaryCodePoint = pInfoState->getPrimaryCodePointAt(
                         dicNode->getInputIndex(0));
                 const int nodeCodePoint = dicNode->getNodeCodePoint();
+                const int keyIndex = traverseSession->getProximityInfo()->getKeyIndexOf(
+                        primaryCodePoint);
                 // TODO: Check whether the input code point is on the keyboard.
                 if (primaryCodePoint == nodeCodePoint) {
                     // Node code point is same as original code point on the keyboard.
@@ -53,6 +56,9 @@ ErrorTypeUtils::ErrorType TypingWeighting::getErrorType(const CorrectionType cor
                 } else if (CharUtils::toBaseCodePoint(primaryCodePoint)
                         == CharUtils::toBaseCodePoint(nodeCodePoint)) {
                     // Base code points are the same but the code point is intentionally input.
+                    if (keyIndex == NOT_AN_INDEX) {
+                        return ErrorTypeUtils::MATCH_WITH_MISSING_EXPLICIT_ACCENT;
+                    }
                     return ErrorTypeUtils::MATCH_WITH_WRONG_ACCENT;
                 } else if (CharUtils::toLowerCase(primaryCodePoint)
                         == CharUtils::toBaseLowerCase(nodeCodePoint)) {
@@ -61,6 +67,10 @@ ErrorTypeUtils::ErrorType TypingWeighting::getErrorType(const CorrectionType cor
                     return ErrorTypeUtils::MATCH_WITH_MISSING_ACCENT
                             | ErrorTypeUtils::MATCH_WITH_WRONG_CASE;
                 } else {
+                    if (keyIndex == NOT_AN_INDEX) {
+                        return ErrorTypeUtils::MATCH_WITH_MISSING_EXPLICIT_ACCENT
+                                | ErrorTypeUtils::MATCH_WITH_WRONG_CASE;
+                    }
                     // Base code points are the same and the cases are different.
                     return ErrorTypeUtils::MATCH_WITH_WRONG_ACCENT
                             | ErrorTypeUtils::MATCH_WITH_WRONG_CASE;