Adds UserPhrasesLM for user phrases.

Since there is no probability information for users' custom phrases,
they should be stored in a format differs from data.txt. Using the same
format and FastLM to parse user phrases just because of laziness but it
is not the right way.

The pull request adds a new language model class to parse user phrases.
It also update the input method controller to adopt the new user phrase
format.

McBopomofo.xcodeproj/project.pbxproj

@@ -38,6 +38,7 @@
 		6AFF97F2253B299E007F1C49 /* NonModalAlertWindowController.xib in Resources */ = {isa = PBXBuildFile; fileRef = 6AFF97F0253B299E007F1C49 /* NonModalAlertWindowController.xib */; };
 		D41355D8278D74B5005E5CBD /* LanguageModelManager.mm in Sources */ = {isa = PBXBuildFile; fileRef = D41355D7278D7409005E5CBD /* LanguageModelManager.mm */; };
 		D41355DB278E6D17005E5CBD /* McBopomofoLM.cpp in Sources */ = {isa = PBXBuildFile; fileRef = D41355D9278E6D17005E5CBD /* McBopomofoLM.cpp */; };
+		D41355DE278EA3ED005E5CBD /* UserPhrasesLM.cpp in Sources */ = {isa = PBXBuildFile; fileRef = D41355DC278EA3ED005E5CBD /* UserPhrasesLM.cpp */; };
 		D427A9C125ED28CC005D43E0 /* OpenCCBridge.swift in Sources */ = {isa = PBXBuildFile; fileRef = D427A9C025ED28CC005D43E0 /* OpenCCBridge.swift */; };
 		D427F76A278C9E29004A2160 /* CandidateUI in Frameworks */ = {isa = PBXBuildFile; productRef = D427F769278C9E29004A2160 /* CandidateUI */; };
 		D427F76C278CA2B0004A2160 /* AppDelegate.swift in Sources */ = {isa = PBXBuildFile; fileRef = D427F76B278CA1BA004A2160 /* AppDelegate.swift */; };
@@ -161,6 +162,8 @@
 		D41355D7278D7409005E5CBD /* LanguageModelManager.mm */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.cpp.objcpp; path = LanguageModelManager.mm; sourceTree = "<group>"; };
 		D41355D9278E6D17005E5CBD /* McBopomofoLM.cpp */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.cpp.cpp; path = McBopomofoLM.cpp; sourceTree = "<group>"; };
 		D41355DA278E6D17005E5CBD /* McBopomofoLM.h */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.c.h; path = McBopomofoLM.h; sourceTree = "<group>"; };
+		D41355DC278EA3ED005E5CBD /* UserPhrasesLM.cpp */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.cpp.cpp; path = UserPhrasesLM.cpp; sourceTree = "<group>"; };
+		D41355DD278EA3ED005E5CBD /* UserPhrasesLM.h */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.c.h; path = UserPhrasesLM.h; sourceTree = "<group>"; };
 		D427A9BF25ED28CC005D43E0 /* McBopomofo-Bridging-Header.h */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.c.h; path = "McBopomofo-Bridging-Header.h"; sourceTree = "<group>"; };
 		D427A9C025ED28CC005D43E0 /* OpenCCBridge.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = OpenCCBridge.swift; sourceTree = "<group>"; };
 		D427F768278C9D0D004A2160 /* CandidateUI */ = {isa = PBXFileReference; lastKnownFileType = wrapper; name = CandidateUI; path = Packages/CandidateUI; sourceTree = "<group>"; };
@@ -268,10 +271,12 @@
 				6A0D4F2215FC0EB100ABF4B3 /* OpenVanilla */,
 				6A0421A615FEF3F50061ED63 /* FastLM.cpp */,
 				6A0421A715FEF3F50061ED63 /* FastLM.h */,
-				D47F7DD2278C1263002F9DD7 /* UserOverrideModel.cpp */,
+				D41355DC278EA3ED005E5CBD /* UserPhrasesLM.cpp */,
-				D47F7DD1278C1263002F9DD7 /* UserOverrideModel.h */,
+				D41355DD278EA3ED005E5CBD /* UserPhrasesLM.h */,
 				D41355D9278E6D17005E5CBD /* McBopomofoLM.cpp */,
 				D41355DA278E6D17005E5CBD /* McBopomofoLM.h */,
+				D47F7DD2278C1263002F9DD7 /* UserOverrideModel.cpp */,
+				D47F7DD1278C1263002F9DD7 /* UserOverrideModel.h */,
 			);
 			path = Engine;
 			sourceTree = "<group>";
@@ -561,6 +566,7 @@
 				D41355DB278E6D17005E5CBD /* McBopomofoLM.cpp in Sources */,
 				D47F7DD3278C1263002F9DD7 /* UserOverrideModel.cpp in Sources */,
 				6A0D4F4515FC0EB100ABF4B3 /* Mandarin.cpp in Sources */,
+				D41355DE278EA3ED005E5CBD /* UserPhrasesLM.cpp in Sources */,
 				6A0421A815FEF3F50061ED63 /* FastLM.cpp in Sources */,
 				D41355D8278D74B5005E5CBD /* LanguageModelManager.mm in Sources */,
 			);

Source/Engine/McBopomofoLM.h

@@ -3,6 +3,7 @@
 
 #include <stdio.h>
 #include "FastLM.h"
+#include "UserPhrasesLM.h"
 
 namespace McBopomofo {
 
@@ -23,8 +24,8 @@ public:
 
 protected:
     FastLM m_languageModel;
-    FastLM m_userPhrases;
+    UserPhrasesLM m_userPhrases;
-    FastLM m_excludedPhrases;
+    UserPhrasesLM m_excludedPhrases;
 };
 };
 

Source/Engine/UserPhrasesLM.cpp

@@ -0,0 +1,207 @@
+#include "UserPhrasesLM.h"
+#include <sys/mman.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <fstream>
+#include <unistd.h>
+
+using namespace Formosa::Gramambular;
+using namespace McBopomofo;
+
+UserPhrasesLM::UserPhrasesLM()
+    : fd(-1)
+    , data(0)
+    , length(0)
+{
+}
+
+UserPhrasesLM::~UserPhrasesLM()
+{
+    if (data) {
+        close();
+    }
+}
+
+bool UserPhrasesLM::open(const char *path)
+{
+    if (data) {
+        return false;
+    }
+
+    fd = ::open(path, O_RDONLY);
+    if (fd == -1) {
+        printf("open:: file not exist");
+        return false;
+    }
+
+    struct stat sb;
+    if (fstat(fd, &sb) == -1) {
+        printf("open:: cannot open file");
+        return false;
+    }
+
+    length = (size_t)sb.st_size;
+
+    data = mmap(NULL, length, PROT_WRITE, MAP_PRIVATE, fd, 0);
+    if (!data) {
+        ::close(fd);
+        return false;
+    }
+
+    char *head = (char *)data;
+    char *end = (char *)data + length;
+    char c;
+    Row row;
+
+start:
+    // EOF -> end
+    if (head == end) {
+        goto end;
+    }
+
+    c = *head;
+    // \s -> error
+    if (c == ' ') {
+        goto error;
+    }
+    // \n -> start
+    else if (c == '\n') {
+        head++;
+        goto start;
+    }
+
+    // \w -> record column star, state1
+    row.value = head;
+    head++;
+    // fall through to state 1
+
+state1:
+    // EOF -> error
+    if (head == end) {
+        goto error;
+    }
+
+    c = *head;
+    // \n -> error
+    if (c == '\n') {
+        goto error;
+    }
+    // \s -> state2 + zero out ending + record column start
+    else if (c == ' ') {
+        *head = 0;
+        head++;
+        row.key = head;
+        goto state2;
+    }
+
+    // \w -> state1
+    head++;
+    goto state1;
+
+state2:
+    if (head == end) {
+        *head = 0;
+        head++;
+        keyRowMap[row.key].push_back(row);
+        goto end;
+    }
+
+    c = *head;
+    // \s -> error
+    if (c == ' ' || c == '\n') {
+        *head = 0;
+        head++;
+        keyRowMap[row.key].push_back(row);
+        if (c == ' ') {
+            goto state3;
+        }
+        goto start;
+    }
+
+    // \w -> state 2
+    head++;
+    goto state2;
+
+state3:
+    if (head == end) {
+        *head = 0;
+        head++;
+        keyRowMap[row.key].push_back(row);
+        goto end;
+    }
+
+    c = *head;
+    if (c == '\n') {
+        goto start;
+    }
+
+    head++;
+    goto state3;
+
+error:
+    close();
+    return false;
+
+end:
+    static const char *space = " ";
+    Row emptyRow;
+    emptyRow.key = space;
+    emptyRow.value = space;
+    keyRowMap[space].push_back(emptyRow);
+
+    return true;
+}
+
+void UserPhrasesLM::close()
+{
+    if (data) {
+        munmap(data, length);
+        ::close(fd);
+        data = 0;
+    }
+
+    keyRowMap.clear();
+}
+
+void UserPhrasesLM::dump()
+{
+    size_t rows = 0;
+    for (map<const char *, vector<Row> >::const_iterator i = keyRowMap.begin(), e = keyRowMap.end(); i != e; ++i) {
+        const vector<Row>& r = (*i).second;
+        for (vector<Row>::const_iterator ri = r.begin(), re = r.end(); ri != re; ++ri) {
+            const Row& row = *ri;
+            cerr << row.key << " " << row.value << "\n";
+            rows++;
+        }
+    }
+}
+
+const vector<Bigram> UserPhrasesLM::bigramsForKeys(const string& preceedingKey, const string& key)
+{
+    return vector<Bigram>();
+}
+
+const vector<Unigram> UserPhrasesLM::unigramsForKey(const string& key)
+{
+    vector<Unigram> v;
+    map<const char *, vector<Row> >::const_iterator i = keyRowMap.find(key.c_str());
+
+    if (i != keyRowMap.end()) {
+        for (vector<Row>::const_iterator ri = (*i).second.begin(), re = (*i).second.end(); ri != re; ++ri) {
+            Unigram g;
+            const Row& r = *ri;
+            g.keyValue.key = r.key;
+            g.keyValue.value = r.value;
+            g.score = 0.0;
+            v.push_back(g);
+        }
+    }
+
+    return v;
+}
+
+bool UserPhrasesLM::hasUnigramsForKey(const string& key)
+{
+    return keyRowMap.find(key.c_str()) != keyRowMap.end();
+}
+

Source/Engine/UserPhrasesLM.h

@@ -0,0 +1,51 @@
+#ifndef USERPHRASESLM_H
+#define USERPHRASESLM_H
+
+#include <stdio.h>
+
+#include <string>
+#include <map>
+#include <iostream>
+#include "LanguageModel.h"
+
+namespace McBopomofo {
+
+using namespace Formosa::Gramambular;
+
+class UserPhrasesLM : public LanguageModel
+{
+public:
+    UserPhrasesLM();
+    ~UserPhrasesLM();
+    
+    bool open(const char *path);
+    void close();
+    void dump();
+    
+    virtual const vector<Bigram> bigramsForKeys(const string& preceedingKey, const string& key);
+    virtual const vector<Unigram> unigramsForKey(const string& key);
+    virtual bool hasUnigramsForKey(const string& key);
+    
+protected:
+    struct CStringCmp
+    {
+        bool operator()(const char* s1, const char* s2) const
+        {
+            return strcmp(s1, s2) < 0;
+        }
+    };
+    
+    struct Row {
+        const char *key;
+        const char *value;
+    };
+    
+    map<const char *, vector<Row>, CStringCmp> keyRowMap;
+    int fd;
+    void *data;
+    size_t length;
+};
+
+}
+
+#endif

Source/InputMethodController.mm

@@ -610,13 +610,11 @@ NS_INLINE size_t max(size_t a, size_t b) { return a > b ? a : b; }
     [string appendString:reading];
     [string appendString:@" "];
     NSMutableArray *readingsArray = [[NSMutableArray alloc] init];
-    vector<std::string> v = _builder->readingsAtRange(begin,end);
+    vector<std::string> v = _builder->readingsAtRange(begin, end);
     for(vector<std::string>::iterator it_i=v.begin(); it_i!=v.end(); ++it_i) {
         [readingsArray addObject:[NSString stringWithUTF8String:it_i->c_str()]];
     }
     [string appendString:[readingsArray componentsJoinedByString:@"-"]];
-    [string appendString:@" "];
-    [string appendString:@"-1.0"];
     return string;
 }
 

Source/LanguageModelManager.mm

@@ -98,17 +98,42 @@ static void LTLoadLanguageModelFile(NSString *filenameWithoutExtension, McBopomo
         return NO;
     }
 
-    NSString *currentMarkedPhrase = [userPhrase stringByAppendingString:@"\n"];
+    BOOL shuoldAddLineBreakAtFront = NO;
-
     NSString *path = [self userPhrasesDataPathMcBopomofo];
-    NSFileHandle *file = [NSFileHandle fileHandleForUpdatingAtPath:path];
+
-    if (!file) {
+    if ([[NSFileManager defaultManager] fileExistsAtPath:path]) {
+        NSError *error = nil;
+        NSDictionary *attr = [[NSFileManager defaultManager] attributesOfItemAtPath:path error:&error];
+        unsigned long long fileSize = [attr fileSize];
+        if (!error && fileSize) {
+            NSFileHandle *readFile = [NSFileHandle fileHandleForReadingAtPath:path];
+            if (readFile) {
+                [readFile seekToFileOffset:fileSize - 1];
+                NSData *data = [readFile readDataToEndOfFile];
+                const void *bytes = [data bytes];
+                if (*(char *)bytes != '\n') {
+                    shuoldAddLineBreakAtFront = YES;
+                }
+                [readFile closeFile];
+            }
+        }
+    }
+
+    NSMutableString *currentMarkedPhrase = [NSMutableString string];
+    if (shuoldAddLineBreakAtFront) {
+        [currentMarkedPhrase appendString:@"\n"];
+    }
+    [currentMarkedPhrase appendString:userPhrase];
+    [currentMarkedPhrase appendString:@"\n"];
+
+    NSFileHandle *writeFile = [NSFileHandle fileHandleForUpdatingAtPath:path];
+    if (!writeFile) {
         return NO;
     }
-    [file seekToEndOfFile];
+    [writeFile seekToEndOfFile];
     NSData *data = [currentMarkedPhrase dataUsingEncoding:NSUTF8StringEncoding];
-    [file writeData:data];
+    [writeFile writeData:data];
-    [file closeFile];
+    [writeFile closeFile];
 
     [self loadUserPhrasesModel];
     return YES;

Source/McBopomofo-Bridging-Header.h

@@ -7,5 +7,4 @@
 @interface LanguageModelManager : NSObject
 + (void)loadDataModels;
 + (void)loadUserPhrasesModel;
-+ (BOOL)checkIfUserLanguageModelFilesExist;
 @end