更新Readme文档

ddddocr/README.md

@@ -3,7 +3,27 @@
 # 带带弟弟OCR通用验证码识别SDK免费开源版
 
 # 今天ddddocr又更新啦！
-  ## 当前版本为1.4.0
+  ## 当前版本为1.4.3
+
+本次升级的主要原因为，[dddd_trainer](https://github.com/sml2h3/dddd_trainer) 的开源进行适配，使[dddd_trainer](https://github.com/sml2h3/dddd_trainer) 训练出的模型可以直接无缝导入到ddddocr里面来使用
+
+## 使用ddddocr调用[dddd_trainer](https://github.com/sml2h3/dddd_trainer) 训练后的模型
+
+[dddd_trainer](https://github.com/sml2h3/dddd_trainer) 训练后会在models目录里导出charsets.json和onnx模型
+
+如下所示，import_onnx_path为onnx所在地址，charsets_path为onnx所在地址
+```python
+import ddddocr
+
+ocr = ddddocr.DdddOcr(det=False, ocr=False, import_onnx_path="myproject_0.984375_139_13000_2022-02-26-15-34-13.onnx", charsets_path="charsets.json")
+
+with open('888e28774f815b01e871d474e5c84ff2.jpg', 'rb') as f:
+    image_bytes = f.read()
+
+res = ocr.classification(image_bytes)
+print(res)
+
+```
 
 # 捐赠 （如果项目有帮助到您，可以选择捐赠一些费用用于ddddocr的后续版本维护，本项目长期维护）
 

ddddocr/__init__.py

@@ -5,7 +5,7 @@ warnings.filterwarnings('ignore')
 import io
 import os
 import base64
-import pathlib
+import json
 import onnxruntime
 from PIL import Image, ImageChops
 import numpy as np
@@ -29,13 +29,26 @@ class TypeError(Exception):
 
 class DdddOcr(object):
     def __init__(self, ocr: bool = True, det: bool = False, old: bool = False, use_gpu: bool = False,
-                 device_id: int = 0, show_ad=True):
+                 device_id: int = 0, show_ad=True, import_onnx_path: str = "", charsets_path: str = ""):
         if show_ad:
             print("欢迎使用ddddocr，本项目专注带动行业内卷，个人博客:wenanzhe.com")
             print("训练数据支持来源于:http://146.56.204.113:19199/preview")
             print("爬虫框架feapder可快速一键接入，快速开启爬虫之旅：https://github.com/Boris-code/feapder")
-
+        self.use_import_onnx = False
-
+        self.__word = False
+        self.__resize = []
+        self.__channel = 1
+        if import_onnx_path != "":
+            det = False
+            ocr = False
+            self.__graph_path = import_onnx_path
+            with open(charsets_path, 'r', encoding="utf-8") as f:
+                info = json.loads(f.read())
+            self.__charset = info['charset']
+            self.__word = info['word']
+            self.__resize = info['image']
+            self.__channel = info['channel']
+            self.use_import_onnx = True
         if det:
             ocr = False
             print("开启det后自动关闭ocr")
@@ -1440,7 +1453,7 @@ class DdddOcr(object):
             self.__providers = [
                 'CPUExecutionProvider',
             ]
-        if ocr or det:
+        if ocr or det or self.use_import_onnx:
             self.__ort_session = onnxruntime.InferenceSession(self.__graph_path, providers=self.__providers)
 
     def preproc(self, img, input_size, swap=(2, 0, 1)):
@@ -1581,35 +1594,53 @@ class DdddOcr(object):
             return []
         return result
 
-    def classification(self, img):
+    def classification(self, img_bytes: bytes = None, img_base64: str = None):
         if self.det:
             raise TypeError("当前识别类型为目标检测")
-        if not isinstance(img, (bytes, str, pathlib.PurePath, Image.Image)):
+        if img_bytes:
-            raise TypeError("未知图片类型")
+            image = Image.open(io.BytesIO(img_bytes))
-        if isinstance(img, bytes):
-            image = Image.open(io.BytesIO(img))
-        elif isinstance(img, Image.Image):
-            image = img.copy()
-        elif isinstance(img, str):
-            image = base64_to_image(img)
         else:
-            assert isinstance(img, pathlib.PurePath)
+            image = base64_to_image(img_base64)
-            image = Image.open(img)
+        if not self.use_import_onnx:
-        image = image.resize((int(image.size[0] * (64 / image.size[1])), 64), Image.ANTIALIAS).convert('L')
+            image = image.resize((int(image.size[0] * (64 / image.size[1])), 64), Image.ANTIALIAS).convert('L')
+        else:
+            if self.__resize[0] == -1:
+                if self.__word:
+                    image = image.resize((self.__resize[1], self.__resize[1]), Image.ANTIALIAS)
+                else:
+                    image = image.resize((int(image.size[0] * (self.__resize[1] / image.size[1])), self.__resize[1]), Image.ANTIALIAS)
+            else:
+                image = image.resize((self.__resize[0], self.__resize[1]), Image.ANTIALIAS)
+            if self.__channel == 1:
+                image = image.convert('L')
+            else:
+                image = image.convert('RGB')
         image = np.array(image).astype(np.float32)
         image = np.expand_dims(image, axis=0) / 255.
-        image = (image - 0.5) / 0.5
+        if not self.use_import_onnx:
+            image = (image - 0.5) / 0.5
+        else:
+            if self.__channel == 1:
+                image = (image - 0.456) / 0.224
+            else:
+                image = (image - np.array([0.485, 0.456, 0.406])) / np.array([0.229, 0.224, 0.225])
+
         ort_inputs = {'input1': np.array([image])}
         ort_outs = self.__ort_session.run(None, ort_inputs)
         result = []
+
         last_item = 0
-        for item in ort_outs[0][0]:
+        if self.__word:
-            if item == last_item:
+            for item in ort_outs[1]:
-                continue
-            else:
-                last_item = item
-            if item != 0:
                 result.append(self.__charset[item])
+        else:
+            for item in ort_outs[0][0]:
+                if item == last_item:
+                    continue
+                else:
+                    last_item = item
+                if item != 0:
+                    result.append(self.__charset[item])
 
         return ''.join(result)