1.0.3

.yarnclean

@@ -0,0 +1,45 @@
+# test directories
+__tests__
+test
+tests
+powered-test
+
+# asset directories
+docs
+doc
+website
+images
+#assets
+
+# examples
+example
+examples
+
+# code coverage directories
+coverage
+.nyc_output
+
+# build scripts
+Makefile
+Gulpfile.js
+Gruntfile.js
+
+# configs
+appveyor.yml
+circle.yml
+codeship-services.yml
+codeship-steps.yml
+wercker.yml
+.tern-project
+.gitattributes
+.editorconfig
+.*ignore
+.eslintrc
+.jshintrc
+.flowconfig
+.documentup.json
+.yarn-metadata.json
+.travis.yml
+
+# misc
+*.md

electron-builder.json5

@@ -12,10 +12,23 @@
   linux: {
     artifactName: "${productName}_${version}_${arch}.${ext}",
     target: ["deb"],
-    icon: "build/logo.ico",
-    desktop: {
-      Icon: "build/logo.ico",
-    },
+    icon: "build/logo.png",
+    description: "my-ocr",
+    category: "Application;AudioVideo;Graphics;",
+    extraFiles: [
+      {
+        from: "./release/ocr_server/",
+        to: "ocr_server",
+      },
+      {
+        from: "./build/logo.png",
+        to: "resources",
+      },
+    ],
+    desktop:{
+      Icon: "/opt/MY-OCR/resources/logo.png",
+      Category: "Application;AudioVideo;Graphics;"
+    }
   },
   mac: {
     target: [
@@ -26,6 +39,10 @@
     ],
     artifactName: "${productName}_${version}_${arch}.${ext}",
     target: ["dmg", "zip"],
+    extraFiles: {
+      from: "./release/ocr_server/",
+      to: "ocr_server",
+    },
   },
   win: {
     target: [
@@ -55,7 +72,7 @@
   },
   publish: {
     provider: "generic",
-    url: "http://127.0.0.1:5500/",
+    url: "http://136.33.32.127/my-ocr/",
     channel: "latest",
     updaterCacheDirName: "my-ocr-updater",
   },

electron/main/index.ts

@@ -90,7 +90,7 @@ function createWindow() {
   });
 
   if (process.platform === "darwin") {
-    app.dock.setIcon(path.join(__dirname, "logo.png"));
+    app.dock.setIcon(path.join(process.env.VITE_PUBLIC, "logo.png"));
   }
 
   // and load the index.html of the app.
@@ -116,11 +116,28 @@ function createWindow() {
       console.log(...args);
     },
   });
+
+  let lastVisualStatus = true
+
+  screenshots.on("startCapture", (_e) => {
+    if (!mainWindow) return;
+    lastVisualStatus = mainWindow.isVisible()
+    mainWindow.hide();
+  });
+
+  screenshots.on("endCapture", (_e) => {
+    if (!mainWindow) return;
+    if(lastVisualStatus){
+      mainWindow.show();
+      mainWindow.focus();
+    }
+  });
 
   screenshots.on("ok", (_e, buffer) => {
-    mainWindow?.webContents.send("SCREENSHOTS:ok", buffer);
-    mainWindow?.show();
-    mainWindow?.focus();
+    if (!mainWindow) return;
+    mainWindow.webContents.send("SCREENSHOTS:ok", buffer);
+    mainWindow.show();
+    mainWindow.focus();
   });
 
   ipcMain.on("startCapture", () => {
@@ -141,7 +158,7 @@ app
   .whenReady()
   .then(() => {
     let mainWindow = createWindow();
-    tray = new Tray(path.join(process.env.VITE_PUBLIC, "favicon.ico"));
+    tray = new Tray(path.join(process.env.VITE_PUBLIC, "logo.png"));
     tray.setToolTip("MY-OCR");
     const menu = getTrayMenu(mainWindow);
     if (isMac) {

electron/main/screenshots/index.ts

@@ -87,6 +87,13 @@ export default class Screenshots extends Events {
    * 开始截图
    */
   public async startCapture(): Promise<void> {
+    const event = new Event();
+    this.emit("startCapture", event);
+
+    await new Promise<void>((resolve) => {
+      setTimeout(() => resolve(), 100);
+    }),
+
     this.logger("startCapture");
 
     const display = getDisplay();
@@ -102,6 +109,9 @@ export default class Screenshots extends Events {
    * 结束截图
    */
   public async endCapture(): Promise<void> {
+    const event = new Event();
+    this.emit("endCapture", event);
+
     this.logger("endCapture");
     await this.reset();
 

package.json

@@ -1,11 +1,12 @@
 {
   "name": "my-ocr",
   "productName": "MY-OCR",
-  "version": "1.0.0",
+  "version": "1.0.3",
   "private": true,
   "main": "dist-electron/main/index.js",
   "type": "module",
   "homepage": ".",
+  "author": "liumingye <[email protected]>",
   "debug": {
     "env": {
       "VITE_DEV_SERVER_URL": "http://127.0.0.1:3000/"
@@ -59,5 +60,6 @@
   },
   "resolutions": {
     "jackspeak": "2.1.1"
-  }
+  },
+  "packageManager": "[email protected]+sha512.a6b2f7906b721bba3d67d4aff083df04dad64c399707841b7acf00f6b133b7ac24255f2652fa22ae3534329dc6180534e98d17432037ff6fd140556e2bb3137e"
 }

py-service/ocr_server.py

@@ -6,16 +6,17 @@
 import cv2
 from thrift.transport import TSocket
 from thrift.transport import TTransport
-from thrift.protocol import TBinaryProtocol
+from thrift.protocol import TCompactProtocol
 from thrift.server import TServer
 from thrift_ocr_py.ocr import OcrService
 import paddleocr
 from tbpu import getParser
 import gc
 from utils.timer import Timer
+from rectify.rectify import correct_image_base64, preProcess
 
 # 内存回收时间
-gc_time = 10
+gc_time = 60
 RUN_PATH = os.path.split(os.path.realpath(__file__))[0]
 
 
@@ -51,61 +52,91 @@ def initOcr():
         # 是否使用方向分类器
         use_angle_cls=True,
         ocr_version="PP-OCRv4",
-        enable_mkldnn=True,
+        # enable_mkldnn=True,
     )
 
+def readBase64Image(b64):
+    img_data = re.sub("^data:image/.+;base64,", "", b64)
+    im = cv2.imdecode(
+        np.frombuffer(base64.b64decode(img_data), np.uint8), -1
+    )
+    if im.shape[2] == 4:
+        im = cv2.cvtColor(im, cv2.COLOR_BGRA2BGR)
+    return im
 
 class Ocr_handler:
+    # def four_point_transform(self, img_path):
+    #     res = {
+    #         "code": 100
+    #     }
+    #     im = readBase64Image(img_path)
+    #     screenCnt, ratio = preProcess(im)
+    #     print('four_point_transform', screenCnt)
+    #     res["data"] = screenCnt
+    #     return json.dumps(res)
+
+    def rectify(self, img_path):
+        res = {
+            "code": 100
+        }
+        im = readBase64Image(img_path)
+        result = correct_image_base64(im)
+        print('rectify',result[0:999])
+        res["data"] = result
+        return json.dumps(res)
+
     def ocr(self, nid, img_path, config):
         timer.cancel()
         initOcr()
         # 设置config默认值
         default_config = {"tbpu": "none"}
         config = {**default_config, **config}
-        res = {}
-        print(f"start ocr {nid} {img_path}")
+        res = {
+            "code": 100
+        }
+        # print(f"start ocr {nid} {img_path}")
+        print(f"start ocr {nid}")
         print(f"{nid} start ocr")
-        try:
-            if img_path.startswith("data:"):
-                img_data = re.sub("^data:image/.+;base64,", "", img_path)
-                im = cv2.imdecode(
-                    np.frombuffer(base64.b64decode(img_data), np.uint8), -1
-                )
-                if im.shape[2] == 4:
-                    im = cv2.cvtColor(im, cv2.COLOR_BGRA2BGR)
-                result = ppocr.ocr(im, cls=True)
-            else:
-                result = ppocr.ocr(img_path, cls=True)
-
-            if result:
-                tbs = result[0]
-                res["code"] = 100
-
-                # 处理数据
-                res["data"] = transform_data(tbs)
-
-                # 计算平均置信度
-                score, num = 0, 0
-                for r in res["data"]:
-                    score += r["score"]
-                    num += 1
-                if num > 0:
-                    score /= num
-                res["score"] = score
-
-                # 执行 tbpu
-                res["data"] = getParser(config["tbpu"]).run(res["data"])
-                # 如果忽略区域等处理将所有文本删除，则结束tbpu
-                if not res["data"]:
-                    res["code"] = 101
-                    res["data"] = ""
-            else:
-                res["code"] = 901
-                res["data"] = f"{nid} ocr error"
-        except Exception as e:
-            print(f"{nid} ocr error {e}")
-            res["code"] = 901
-            res["data"] = f"{nid} ocr error {e}"
+        # try:
+        if img_path.startswith("data:"):
+            im = readBase64Image(img_path)
+            # im = correct_image(im)
+            # content = cv2.imencode('.jpg', im)[1]
+            # base64_str = str(base64.b64encode(content))[2:-1]
+            # res['image'] =  base64_str
+            result = ppocr.ocr(im, cls=True)
+        else:
+            result = ppocr.ocr(img_path, cls=True)
+
+        tbs = result[0]
+
+        if tbs:
+            # 处理数据
+            res["data"] = transform_data(tbs)
+
+            # 计算平均置信度
+            score, num = 0, 0
+            for r in res["data"]:
+                score += r["score"]
+                num += 1
+            if num > 0:
+                score /= num
+            res["score"] = score
+
+            # 执行 tbpu
+            res["data"] = getParser(config["tbpu"]).run(res["data"])
+
+            # 如果忽略区域等处理将所有文本删除，则结束tbpu
+            if not res["data"]:
+                res["code"] = 101
+                res["data"] = ""
+        else:
+            res["code"] = 100
+            res["data"] = ""
+        # except Exception as e:
+        #     print(f"{nid} ocr error {e}")
+        #     res["code"] = 901
+        #     res["data"] = f"{nid} ocr error {e}"
         print("done")
         timer.reset()
         return json.dumps(res)
@@ -141,9 +172,9 @@ def transform_data(data):
     # 选择传输层
     tfactory = TTransport.TBufferedTransportFactory()
     # 选择传输协议
-    pfactory = TBinaryProtocol.TBinaryProtocolFactory()
+    pfactory = TCompactProtocol.TCompactProtocolFactory()
     # 创建服务端
-    server = TServer.TThreadedServer(processor, transport, tfactory, pfactory)
+    server = TServer.TThreadPoolServer(processor, transport, tfactory, pfactory)
     print(f"start server on {port}")
     server.serve()
     print("Done")

py-service/rectify/___init__.py

@@ -0,0 +1 @@
+__all__ = ['transform','rectify','imgEnhance']

py-service/rectify/imgEnhance.py

@@ -0,0 +1,35 @@
+from PIL import ImageEnhance
+import numpy as np
+import cv2
+
+class Enhancer:
+    def bright(self, image, brightness):
+        enh_bri = ImageEnhance.Brightness(image)
+        brightness = brightness
+        imageBrightend = enh_bri.enhance(brightness)
+        return imageBrightend
+
+    def color(self, image, color):
+        enh_col = ImageEnhance.Color(image)
+        color = color
+        imageColored = enh_col.enhance(color)
+        return imageColored
+
+    def contrast(self, image, contrast):
+        enh_con = ImageEnhance.Contrast(image)
+        contrast = contrast
+        image_contrasted = enh_con.enhance(contrast)
+        return image_contrasted
+
+    def sharp(self, image, sharpness):
+        enh_sha = ImageEnhance.Sharpness(image)
+        sharpness = sharpness
+        image_sharped = enh_sha.enhance(sharpness)
+        return image_sharped
+
+    def gamma(self, image, gamma):
+        # gamma_table = [np.power(x / 255.0, gamma) * 255.0 for x in range(256)]
+        # gamma_table = np.round(np.array(gamma_table)).astype(np.uint8)
+        # return cv2.LUT(image, gamma_table)
+        gamma_image = np.power(image / float(np.max(image)), gamma)
+        return gamma_image