multi language support

tmarenko · Aug 27, 2022 · f1ba0a6 · f1ba0a6
1 parent dadc3cf
commit f1ba0a6
Show file tree

Hide file tree

Showing 9 changed files with 179 additions and 40 deletions.
diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
@@ -8,7 +8,7 @@ env:
   OPENCV_VERSION: 4.2.0
   LEPTONICA_VERSION: 1.74.3
   TESSERACT_VERSION: 3.05.02
-  TESSERACT_ENG_DATA: https://github.com/tesseract-ocr/tessdata/raw/3.04.00/eng.traineddata
+  TESSERACT_LANGUAGE: eng
   REPO_NAME: ${{ github.event.repository.name }}
   LEPTONICA_BUILD: ${{github.workspace}}/leptonica/build
   TESSERACT_BUILD: ${{github.workspace}}/tesseract/build
@@ -65,11 +65,11 @@ jobs:
       - name: Build ${{env.REPO_NAME}}
         run: |
           cd ${{github.workspace}}/${{env.REPO_NAME}}
-          cmake -B ${{github.workspace}}/${{env.REPO_NAME}}/build -G "NMake Makefiles" -DCMAKE_BUILD_TYPE=${{env.BUILD_TYPE}} -DCMAKE_PREFIX_PATH="${{env.LEPTONICA_BUILD}};${{env.TESSERACT_BUILD}};${{env.OPENCV_BUILD}}" -DTESSERACT_ENG_DATA=${{env.TESSERACT_ENG_DATA}}
+          cmake -B ${{github.workspace}}/${{env.REPO_NAME}}/build -G "NMake Makefiles" -DCMAKE_BUILD_TYPE=${{env.BUILD_TYPE}} -DCMAKE_PREFIX_PATH="${{env.LEPTONICA_BUILD}};${{env.TESSERACT_BUILD}};${{env.OPENCV_BUILD}}" -DTESSERACT_LANGUAGE=${{env.TESSERACT_LANGUAGE}}
           cmake --build ${{github.workspace}}/${{env.REPO_NAME}}/build --config ${{env.BUILD_TYPE}}
 
       - uses: actions/upload-artifact@master
         if: always()
         with:
-          name: ${{env.REPO_NAME}}-build
+          name: ${{env.REPO_NAME}}-${{github.ref_name}}-build
           path: ${{github.workspace}}/${{env.REPO_NAME}}/build/bin
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -1,12 +1,17 @@
 cmake_minimum_required(VERSION 3.21)
 project(GenshinImpact_PaimonShutUp)
 
+ADD_DEFINITIONS(-DUNICODE)
+ADD_DEFINITIONS(-D_UNICODE)
+
 set(CMAKE_CXX_STANDARD 14)
 set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)
 set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} /MANIFEST:NO")
 set(MAIN_MANIFEST "main.manifest")
-if(NOT TESSERACT_ENG_DATA)
-    set(TESSERACT_ENG_DATA "https://github.com/tesseract-ocr/tessdata/raw/3.04.00/eng.traineddata")
+set(CONFIG_FILE "settings.cfg")
+if(NOT TESSERACT_LANGUAGE)
+    MESSAGE(STATUS "TESSERACT_LANGUAGE is empty. Using 'eng' as default")
+    set(TESSERACT_LANGUAGE "eng")
 endif()
 
 find_package(OpenCV 4.2.0 COMPONENTS core imgproc REQUIRED)
@@ -18,16 +23,17 @@ include_directories(${OpenCV_INCLUDE_DIRS})
 include_directories(${Leptonica_INCLUDE_DIRS})
 include_directories(${Tesseract_INCLUDE_DIRS})
 
-add_executable(${PROJECT_NAME} source/main.cpp source/tesseract.cpp source/tesseract.h appicon.rc)
+add_executable(${PROJECT_NAME} source/main.cpp source/tesseract.cpp source/tesseract.h appicon.rc source/config.cpp source/config.h)
 
 MESSAGE(STATUS "LINK LIBRARIES ${OpenCV_LIBS} ${Leptonica_LIBRARIES} ${Tesseract_LIBRARIES}")
 target_link_libraries(${PROJECT_NAME} ${OpenCV_LIBS} ${Leptonica_LIBRARIES} libtesseract)
 
-if (NOT EXISTS "${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/tessdata/eng.traineddata")
-    MESSAGE(STATUS "Downloading Tesseract-OCR data for ENG language from: ${TESSERACT_ENG_DATA}")
+if (NOT EXISTS "${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/tessdata/${TESSERACT_LANGUAGE}.traineddata")
+    set(TESSERACT_LANG_DATA "https://github.com/tesseract-ocr/tessdata/raw/3.04.00/${TESSERACT_LANGUAGE}.traineddata")
+    MESSAGE(STATUS "Downloading Tesseract-OCR data for '${TESSERACT_LANGUAGE}' language from: ${TESSERACT_LANG_DATA}")
     file(DOWNLOAD
-            ${TESSERACT_ENG_DATA}
-            ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/tessdata/eng.traineddata)
+            ${TESSERACT_LANG_DATA}
+            ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/tessdata/${TESSERACT_LANGUAGE}.traineddata)
 endif ()
 
 add_custom_command(TARGET ${PROJECT_NAME} POST_BUILD
@@ -41,6 +47,12 @@ add_custom_command(
         ${CMAKE_SOURCE_DIR}/${MAIN_MANIFEST}
         ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/${MAIN_MANIFEST})
 
+add_custom_command(
+        TARGET ${PROJECT_NAME} POST_BUILD
+        COMMAND ${CMAKE_COMMAND} -E copy
+        ${CMAKE_SOURCE_DIR}/${CONFIG_FILE}
+        ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/${CONFIG_FILE})
+
 add_custom_command(TARGET ${PROJECT_NAME} POST_BUILD
         COMMAND "mt.exe" -manifest "${CMAKE_RUNTIME_OUTPUT_DIRECTORY}\\${MAIN_MANIFEST}" -outputresource:"${CMAKE_RUNTIME_OUTPUT_DIRECTORY}\\${PROJECT_NAME}.exe"
         COMMAND ${CMAKE_COMMAND} -E remove -f "${CMAKE_RUNTIME_OUTPUT_DIRECTORY}\\${MAIN_MANIFEST}"

diff --git a/README.md b/README.md
@@ -2,13 +2,35 @@
 
 Automatically mutes Paimon when she speaks
 
+#### Watch how it works: https://youtu.be/pBpDdLcw6zk
+
 ## Installation and usage
 
 Download archive from
 [latest release](https://github.com/tmarenko/GenshinImpact_PaimonShutUp/releases)
 , extract it and run `GenshinImpact_PaimonShutUp.exe` with administrator privileges. 
 Administrator rights needed for access to Windows Audio session.
 
-## Watch how it works
+## Setup in-game language
+
+Edit `settings.cfg` file and set your in-game language in `language` variable:
+
+```buildoutcfg
+[config]
+language=eng
+```
 
-https://youtu.be/pBpDdLcw6zk
+Supported languages:
+* 简体中文 (`language=chi_sim`)
+* 繁體中文 (`language=chi_tra`)
+* English (`language=eng`)
+* 한국어 (`language=kor`)
+* 日本語 (`language=jpn`)
+* Español (`language=spa`)
+* Français (`language=fra`)
+* Русский язык (`language=rus`)
+* ภาษาไทย (`language=tha`)
+* Tiếng Việt (`language=vie`)
+* Deutsch (`language=deu`)
+* Bahasa Indonesia (`language=ind`)
+* Português (`language=por`)
diff --git a/settings.cfg b/settings.cfg
@@ -0,0 +1,33 @@
+[config]
+# available: chi_sim, chi_tra, eng, kor, jpn, spa, fra, rus, tha, vie, deu, ind, por
+language=eng
+
+[genshin_window_name]
+genshin_chi_sim=原神
+genshin_chi_tra=原神
+genshin_eng=Genshin Impact
+genshin_kor=원신
+genshin_jpn=原神
+genshin_spa=Genshin Impact
+genshin_fra=Genshin Impact
+genshin_rus=Genshin Impact
+genshin_tha=Genshin Impact
+genshin_vie=Genshin Impact
+genshin_deu=Genshin Impact
+genshin_ind=Genshin Impact
+genshin_por=Genshin Impact
+
+[paimon_name]
+paimon_chi_sim=派蒙
+paimon_chi_tra=派蒙
+paimon_eng=Paimon
+paimon_kor=폐이몬
+paimon_jpn=バイモン
+paimon_spa=Paimon
+paimon_fra=Paimon
+paimon_rus=Паймон
+paimon_tha=Paimon
+paimon_vie=Paimon
+paimon_deu=Paimon
+paimon_ind=Paimon
+paimon_por=Paimon
diff --git a/source/config.cpp b/source/config.cpp
@@ -0,0 +1,30 @@
+#include "config.h"
+
+
+std::map<std::string, std::string> ParseConfig(const std::string &fileName) {
+    std::map<std::string, std::string> configMap = {
+            {"language",    "eng"},
+            {"genshin_eng", "Genshin Impact"},
+            {"paimon_eng",  "Paimon"}
+    };
+    std::ifstream file(fileName);
+    if (file.is_open()) {
+        std::string line;
+        while (std::getline(file, line)) {
+            std::istringstream is_line(line);
+            std::string key;
+            if (std::getline(is_line, key, '=')) {
+                std::string value;
+                if (std::getline(is_line, value)) {
+                    configMap[key] = value;
+                }
+            }
+        }
+    }
+    return configMap;
+}
+
+std::wstring convertStringToWstring(std::string &input) {
+    std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>> converter;
+    return converter.from_bytes(input);
+}
diff --git a/source/config.h b/source/config.h
@@ -0,0 +1,12 @@
+#ifndef GENSHINIMPACT_PAIMONSHUTUP_CONFIG_H
+#define GENSHINIMPACT_PAIMONSHUTUP_CONFIG_H
+
+#include <fstream>
+#include <sstream>
+#include <map>
+#include <codecvt>
+
+std::map<std::string, std::string> ParseConfig(const std::string &fileName);
+std::wstring convertStringToWstring(std::string &input);
+
+#endif //GENSHINIMPACT_PAIMONSHUTUP_CONFIG_H
diff --git a/source/main.cpp b/source/main.cpp
@@ -8,13 +8,13 @@
 #include "psapi.h"
 #include <csignal>
 #include "tesseract.h"
+#include "config.h"
 
-#define PW_RENDERFULLCONTENT 0x00000002 // Properly capture DirectComposition genshinWindow content
 
 typedef struct GenshinWindowInfo {
-    const char *windowName = "Genshin Impact";
-    const char *windowClass = "UnityWndClass";
-    HWND hwnd = NULL;
+    std::wstring windowName = L"Genshin Impact";
+    std::wstring windowClass = L"UnityWndClass";
+    HWND hwnd = nullptr;
     int width = 0;
     int height = 0;
     bool active = false;
@@ -31,21 +31,21 @@ const cv::Scalar DEFAULT_DIALOGUE_NAME_POS = {0.4631496915663028, 0.787213101603
 const cv::Scalar OUT_DIALOGUE_NAME_POS = {0.4645968857034299, 0.7582338162046451, 0.5333868470215386, 0.8020998784944536};
 const cv::Scalar DIALOGUE_NAME_COLOR_RANGE_LOW = {0, 170, 230};   // BGR
 const cv::Scalar DIALOGUE_NAME_COLOR_RANGE_HIGH = {10, 210, 255}; // BGR
-const char * PAIMON_NAME = "Paimon";
 
 cv::Mat frame;
 HWND genshinWindow;
 GenshinWindowInfo gwi;
+std::wstring genshinExe = L"GenshinImpact.exe";
 
 
 BOOL CALLBACK EnumWindowsFunc(HWND hwnd, LPARAM lParam) {
     auto *gwiParam = (GenshinWindowInfo *) lParam;
-    TCHAR buf[1024]{};
+    WCHAR buf[1024]{};
 
     GetClassName(hwnd, buf, 100);
-    if (!lstrcmp(buf, gwiParam->windowClass)) {
+    if (!lstrcmp(buf, gwiParam->windowClass.c_str())) {
         GetWindowText(hwnd, buf, 100);
-        if (!lstrcmp(buf, gwiParam->windowName)) {
+        if (!lstrcmp(buf, gwiParam->windowName.c_str())) {
             gwiParam->hwnd = hwnd;
             RECT windowRect;
             GetWindowRect(hwnd, &windowRect);
@@ -61,17 +61,17 @@ void FindGenshinWindow() {
     EnumWindows(EnumWindowsFunc, (LPARAM) (&gwi));
     genshinWindow = gwi.hwnd;
 
-    TCHAR buff[1024]{};
+    WCHAR buff[1024]{};
     GetWindowText(genshinWindow, buff, 100);
-    if (!lstrcmp(buff, gwi.windowName)) {
-        if (!gwi.active) {
+    if (!lstrcmp(buff, gwi.windowName.c_str())) {
+        if (gwi.hwnd && !gwi.active) {
             gwi.active = true;
             std::cout << "Ready for Paimon!" << std::endl;
         }
     } else {
         if (gwi.active) {
             gwi.active = false;
-            gwi.hwnd = NULL;
+            gwi.hwnd = nullptr;
             std::cout << "Game was closed" << std::endl;
         }
     }
@@ -130,7 +130,7 @@ std::string GetTextFromImageByRect(const cv::Mat &image, const cv::Rect& rect) {
 }
 
 
-bool IsPaimonSpeaking() {
+bool IsPaimonSpeaking(const std::string &paimonName) {
     GetFrame(gwi.width, gwi.height);
     if (frame.empty())
         return false;
@@ -144,16 +144,16 @@ bool IsPaimonSpeaking() {
                                 (int) (OUT_DIALOGUE_NAME_POS.val[3] * frame.rows) - (int) (OUT_DIALOGUE_NAME_POS.val[1] * frame.rows));
     std::string defaultDialogue = GetTextFromImageByRect(frame, cropDefault);
     std::string outDialogue = GetTextFromImageByRect(frame, cropOut);
-    return IsStringsSimilar(defaultDialogue, PAIMON_NAME) || IsStringsSimilar(outDialogue, PAIMON_NAME);
+    return IsStringsSimilar(defaultDialogue, paimonName) || IsStringsSimilar(outDialogue, paimonName);
 }
 
 
 bool IsGenshinProcess(DWORD pid) {
-    TCHAR buff[1024];
+    WCHAR buff[1024];
     HANDLE handle = OpenProcess(PROCESS_QUERY_INFORMATION | PROCESS_VM_READ, FALSE, pid);
-    if (GetProcessImageFileName(handle, reinterpret_cast<LPSTR>(buff), sizeof(buff))) {
+    if (GetProcessImageFileName(handle, buff, sizeof(buff))) {
         CloseHandle(handle);
-        return ((std::string) buff).find("GenshinImpact.exe") != std::string::npos;
+        return std::wstring(&buff[0]).find(genshinExe) != std::wstring::npos;
     }
     CloseHandle(handle);
     return false;
@@ -217,16 +217,20 @@ HRESULT SetMuteGenshin(BOOL bMute) {
 
 
 int PaimonShutUp() {
-    if (InitTesseract(NULL, "eng"))
+    std::map<std::string, std::string> configMap = ParseConfig("settings.cfg");
+    gwi.windowName = convertStringToWstring(configMap["genshin_" + configMap["language"]]);
+    DownloadTessdataFileIfNecessary(configMap["language"]);
+    if (InitTesseract(nullptr, configMap["language"].c_str()))
         return 1;
 
-    bool paimonWasHere = false;
+    std::string paimonName = configMap["paimon_" + configMap["language"]];
     std::cout << "Waiting for GenshinImpact.exe process." << std::endl;
+    bool paimonWasHere = false;
     while (!stop) {
         FindGenshinWindow();
         if (!gwi.active)
             continue;
-        bool isPaimonSpeaking = IsPaimonSpeaking();
+        bool isPaimonSpeaking = IsPaimonSpeaking(paimonName);
         if (isPaimonSpeaking && !paimonWasHere) {
             paimonWasHere = true;
             std::cout << "Paimon, shut up!" << std::endl;

diff --git a/source/tesseract.cpp b/source/tesseract.cpp
@@ -1,10 +1,10 @@
 #include "tesseract.h"
 
 tesseract::TessBaseAPI *tesseractApi;
-const char *asciiLetters = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ";
+const char *tesseractDownloadUrl = "https://github.com/tesseract-ocr/tessdata/raw/3.04.00/";
 
 
-int InitTesseract(const char* dataPath, const char* language) {
+int InitTesseract(const char *dataPath, const char *language) {
     tesseractApi = new tesseract::TessBaseAPI();
     if (tesseractApi->Init(dataPath, language)) {
         std::cout << "Could not initialize tesseract lib." << std::endl;
@@ -15,7 +15,7 @@ int InitTesseract(const char* dataPath, const char* language) {
     tesseractApi->SetVariable("tessedit_oem_mode", OEM_DEFAULT);
     tesseractApi->SetVariable("debug_file", "/dev/null");
 
-    std::cout << "Initialized Tesseract API." << std::endl;
+    std::cout << "Initialized Tesseract API with '" << language << "' language." << std::endl;
     return 0;
 }
 
@@ -34,9 +34,9 @@ std::string StripText(const std::string &input) {
         return input;
     auto start_it = input.begin();
     auto end_it = input.rbegin();
-    while (start_it != input.end() && std::isspace(*start_it))
+    while (start_it != input.end() && (*start_it == ' ' || *start_it == '\n'))
         ++start_it;
-    while (end_it != input.rend() && std::isspace(*end_it))
+    while (end_it != input.rend() && (*end_it == ' ' || *end_it == '\n'))
         ++end_it;
     if (start_it > end_it.base())
         return std::string();
@@ -50,7 +50,6 @@ std::string GetTextFromImage(const cv::Mat &image) {
         return std::string();
     }
     char *tesseractOutText;
-    tesseractApi->SetVariable("tessedit_char_whitelist", asciiLetters);
     tesseractApi->SetVariable("tessedit_pageseg_mode", AUTOMATIC_PAGE_SEGMENTATION);
     tesseractApi->SetImage(image.data, image.cols, image.rows, image.channels(), image.cols * image.channels());
     tesseractOutText = tesseractApi->GetUTF8Text();
@@ -78,9 +77,31 @@ unsigned int LevenshteinDistance(const std::string &s1, const std::string &s2) {
 }
 
 
-bool IsStringsSimilar(std::string s1, std::string s2, const double overlap) {
+bool IsStringsSimilar(std::string s1, std::string s2, double overlap) {
     std::transform(s1.begin(), s1.end(), s1.begin(), ::toupper);
     std::transform(s2.begin(), s2.end(), s2.begin(), ::toupper);
     double nonSimilarity = s1.length() > 0 ? (double) LevenshteinDistance(s1, s2) / s1.length() : 1;
     return nonSimilarity <= overlap;
 }
+
+
+inline bool CheckFileExists(const std::string &name) {
+    struct stat buffer{};
+    return (stat(name.c_str(), &buffer) == 0);
+}
+
+
+void DownloadTessdataFileIfNecessary(const std::string &language) {
+    std::string pathToLanguage = "tessdata/" + language + ".traineddata";
+    if (CheckFileExists(pathToLanguage)) {
+        std::cout << "Found Tesseract data for '" << language << "' language at " + pathToLanguage << std::endl;
+        return;
+    }
+    std::string downloadUrl = tesseractDownloadUrl + language + ".traineddata";
+    std::cout << "Downloading Tesseract data for '" << language << "' language from " + downloadUrl << std::endl;
+    HRESULT hr = URLDownloadToFileA(nullptr, downloadUrl.c_str(), pathToLanguage.c_str(), 0, nullptr);
+    if (hr == S_OK)
+        std::cout << "Download complete. Saved to " + pathToLanguage << std::endl;
+    else
+        std::cout << "Error during download: " + std::to_string(hr) << std::endl;
+}