From 8998b84eac0b5e7ba151c235af79c8b32b83833e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E9=99=86=E5=A4=A9=E7=9D=BF?= <trlu.develop@outlook.com>
Date: Wed, 15 Jul 2020 21:10:34 +0800
Subject: [PATCH] Support: Mathpix OCR, close #8

---
 README.md            | 18 ++++++++++++++++
 src_alfred/config.py |  8 +++++--
 src_alfred/ocr.py    | 51 ++++++++++++++++++++++++++++++++++++++------
 3 files changed, 69 insertions(+), 8 deletions(-)

diff --git a/README.md b/README.md
index 1442888..80e3f41 100755
--- a/README.md
+++ b/README.md
@@ -18,6 +18,11 @@
 
 ## 版本
 
+### 4.8
+
+- 支持 Mathpix 公式识别（因学业原因，仅支持识别后输出 Latex 格式文本，更多支持选项待后期开发）;
+- 由于可选识别方式过多，CNOCR 的触发方式修改为唯一触发词 ooc (CNOCR)。
+
 ### 近期更新
 
 - 紧急修复批量文件识别报错的问题；
@@ -26,10 +31,13 @@
 
 ## 能力
 
+### CNOCR
+
 - 离线 OCR (CNOCR)
 - 通用 OCR (百度 | 腾讯优图 | Google)
 - 二维码识别 (百度 | ZXing)
 - 表格文字识别 (百度)
+- 数学公式识别 (Mathpix)
 - 多文件识别 (百度)
 - 文本翻译 (彩云小译)
 
@@ -159,6 +167,16 @@ pip install zxing
 
 - Google OCR 为收费业务，需绑定信用卡，故本项目不带测试 Token，需要自行申请。
 
+### [Mathpix](https://mathpix.com)
+
+#### 触发
+
+- 公式识别：截图至剪贴板后使用关键词 `oom (mathpix)` 触发。
+
+#### 说明
+
+- Mathpix 为收费业务，需绑定信用卡，故本项目不带测试 Token，需要自行[申请](https://accounts.mathpix.com/ocr-api)。
+
 ### [ZXing][10]
 
 #### 触发
diff --git a/src_alfred/config.py b/src_alfred/config.py
index 09b94eb..452abfd 100755
--- a/src_alfred/config.py
+++ b/src_alfred/config.py
@@ -2,10 +2,9 @@
 @Description: Capture than OCR - Variable
 @Author: Chandler Lu
 @Date: 2020-03-09 20:32:15
-@LastEditTime: 2020-04-21 19:42:11
+@LastEditTime: 2020-07-15 20:51:39
 '''
 # -*- coding: UTF-8 -*-
-
 import os
 
 # Control
@@ -26,6 +25,8 @@
 GOOGLE_POST_REFERER = os.environ['google_post_referer']
 GOOGLE_HTTP_PROXY = os.environ['google_http_proxy']
 CAIYUN_TRANSLATE_TOKEN = os.environ['caiyun_token']
+MATHPIX_APP_ID = os.environ['mathpix_app_id']
+MATHPIX_APP_KEY = os.environ['mathpix_app_key']
 
 # Key - Quicker
 
@@ -37,6 +38,8 @@
 # GOOGLE_POST_REFERER = ''
 # GOOGLE_HTTP_PROXY = ''
 # CAIYUN_TRANSLATE_TOKEN = '3975l6lr5pcbvidl6jl2'
+# MATHPIX_APP_ID = ''
+# MATHPIX_APP_KEY = ''
 
 # API
 BAIDU_GET_TOKEN_URL = 'https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials&client_id=' + \
@@ -47,3 +50,4 @@
 TENCENT_YOUTU_OCR_API = 'https://api.ai.qq.com/fcgi-bin/ocr/ocr_generalocr'
 GOOGLE_OCR_API = 'https://vision.googleapis.com/v1/images:annotate'
 CAIYUN_TRANSLATE_API = 'http://api.interpreter.caiyunai.com/v1/translator'
+MATHPIX_API = 'https://api.mathpix.com/v3/text'
diff --git a/src_alfred/ocr.py b/src_alfred/ocr.py
index 55c1b2e..3a04a76 100755
--- a/src_alfred/ocr.py
+++ b/src_alfred/ocr.py
@@ -1,9 +1,9 @@
 '''
-@Description: Capture than OCR - macOS - Online OCR
-@version: 4.7
+@Description: Capture than OCR - Online OCR
+@version: 4.8
 @Author: Chandler Lu
 @Date: 2019-11-26 23:52:36
-@LastEditTime: 2020-04-24 17:25:21
+@LastEditTime: 2020-07-15 20:52:29
 '''
 # -*- coding: UTF-8 -*-
 import sys
@@ -282,6 +282,42 @@ def google_ocr(pic_path):
         declare_network_error()
 
 
+'''
+Mathpix OCR
+'''
+
+
+def mathpix_ocr(pic_path):
+    ''' https://docs.mathpix.com/?python#request-parameters
+    file_path = 'limit.jpg'
+    image_uri = "data:image/jpg;base64," + b64encode(open(pic_path, "rb").read()).decode()
+    r = requests.post("https://api.mathpix.com/v3/text",
+                    data=json.dumps({'src': image_uri}),
+                    headers={"app_id": "YOUR_APP_ID", "app_key": "YOUR_APP_KEY",
+                            "Content-type": "application/json"})
+    print(json.dumps(json.loads(r.text), indent=4, sort_keys=True))
+    '''
+    try:
+        response = requests.post(
+            url=c.MATHPIX_API,
+            headers={
+                'app_id': c.MATHPIX_APP_ID,
+                'app_key': c.MATHPIX_APP_KEY,
+                'Content-type': 'application/json',
+            },
+            data=json.dumps({
+                'src': 'data:image/jpg;base64,' + convert_image_base64(pic_path)
+            })
+        )
+        if (response.status_code == 200):
+            response_json = response.json()['latex_styled']
+            print(response_json, end='')
+        else:
+            print('Request failed!', end='')
+    except requests.exceptions.ConnectionError:
+        declare_network_error()
+
+
 def multi_file_ocr():
     names = [name for name in os.listdir(temp_path)
              if re.search(r'.png|.jpg|.jpeg|.bmp', name, re.IGNORECASE) and os.path.getsize(temp_path + '/' + name) <= 4194304]
@@ -459,7 +495,7 @@ def remove_pic(pic_path):
 
 
 if __name__ == "__main__":
-    if (0 <= ocr_select <= 6):
+    if (ocr_select != 99):
         try:
             os.path.getsize(pic_path)
         except FileNotFoundError:
@@ -472,7 +508,8 @@ def remove_pic(pic_path):
     4: tencent
     5: google
     6: zxing
-    7: file
+    7: mathpix
+    99: file
     '''
     if (ocr_select == 0):
         cnocr_ocr(pic_path)
@@ -489,8 +526,10 @@ def remove_pic(pic_path):
     elif (ocr_select == 6):
         barcode_decode(pic_path)
     elif (ocr_select == 7):
+        mathpix_ocr(pic_path)
+    elif (ocr_select == 99):
         multi_file_ocr()
-    if (ocr_select != 7):
+    if (ocr_select != 99):
         remove_pic(pic_path)
 
 '''