'commit'

AI/WxMini/Start.py

@@ -595,16 +595,10 @@ async def generate_upload_params(current_user: dict = Depends(get_current_user))
     return response_data
 
 
-@app.get("/aichat/process_image")
-async def process_image(image_url: str, current_user: dict = Depends(get_current_user)):
+@app.get("/aichat/recognize_content")
+async def web_recognize_content(image_url: str, current_user: dict = Depends(get_current_user)):
     logger.info(f"current_user:{current_user['login_name']}")
-    """
-    处理图片，自动判断调用哪个功能
-    :param image_url: 图片 URL
-    :return: 流式输出结果
-    """
     try:
-        print("检测到图片主要是物体/场景，开始识别内容：")
         return StreamingResponse(
             recognize_content(client, app.state.mysql_pool, current_user['person_id'], image_url),
             media_type="text/plain")
@@ -612,6 +606,28 @@ async def process_image(image_url: str, current_user: dict = Depends(get_current
         raise HTTPException(status_code=500, detail=str(e))
 
 
+@app.get("/aichat/recognize_text")
+async def web_recognize_text(image_url: str, current_user: dict = Depends(get_current_user)):
+    logger.info(f"current_user:{current_user['login_name']}")
+    try:
+        return StreamingResponse(
+            recognize_text(client, app.state.mysql_pool, current_user['person_id'], image_url),
+            media_type="text/plain")
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+
+@app.get("/aichat/recognize_math")
+async def web_recognize_math(image_url: str, current_user: dict = Depends(get_current_user)):
+    logger.info(f"current_user:{current_user['login_name']}")
+    try:
+        return StreamingResponse(
+            recognize_math(app.state.mysql_pool, current_user['person_id'], image_url),
+            media_type="text/plain")
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+
 # 运行 FastAPI 应用
 if __name__ == "__main__":
     import uvicorn

AI/WxMini/Utils/ImageUtil.py

@@ -1,7 +1,11 @@
 import time
 
+from openai import OpenAI, AsyncOpenAI
+
+from WxMini.Milvus.Config.MulvusConfig import MODELSCOPE_ACCESS_TOKEN
 from WxMini.Utils.MySQLUtil import save_chat_to_mysql
 
+
 async def recognize_text(client, pool, person_id, image_url):
     """
     识别图片中的文字，流式输出，并将结果记录到数据库
@@ -37,6 +41,7 @@ async def recognize_text(client, pool, person_id, image_url):
                 if char != ' ':
                     yield char  # 流式输出字符
                     full_text += char  # 拼接字符
+                    print(char, end='')
                     time.sleep(0.1)  # 控制输出速度
 
     # 记录到数据库
@@ -65,6 +70,7 @@ async def recognize_content(client, pool, person_id, image_url):
                 if char != ' ':
                     yield char  # 流式输出字符
                     full_text += char  # 拼接字符
+                    print(char, end='')
                     time.sleep(0.1)  # 控制输出速度
 
     # 记录到数据库
@@ -72,3 +78,59 @@ async def recognize_content(client, pool, person_id, image_url):
         await save_chat_to_mysql(pool, person_id, f'{image_url}', full_text, "", 0, 2, 2, 2)
     except Exception as e:
         print(f"记录到数据库时出错：{e}")
+
+
+async def recognize_math(pool, person_id, image_url):
+    client = AsyncOpenAI(
+        api_key=MODELSCOPE_ACCESS_TOKEN,
+        base_url="https://api-inference.modelscope.cn/v1"
+    )
+    """
+    识别图片中的数学题，流式输出，并将结果记录到数据库
+    :param client: AsyncOpenAI 客户端
+    :param pool: 数据库连接池
+    :param person_id: 用户 ID
+    :param image_url: 图片 URL
+    :return: 最终拼接的字符串
+    """
+    # 提示词
+    prompt = "You are a helpful and harmless assistant. You are Qwen developed by Alibaba. You should think step-by-step."
+
+    completion = await client.chat.completions.create(
+        model="Qwen/Qwen2.5-VL-32B-Instruct",
+        messages=[
+            {
+                "role": "system",
+                "content": [
+                    {"type": "text", "text": prompt}
+                ],
+            },
+            {
+                "role": "user",
+                "content": [
+                    {
+                        "type": "image_url",
+                        "image_url": {"url": image_url}
+                    },
+                    {"type": "text", "text": "请使用中文回答：如何作答?"},
+                ],
+            }
+        ],
+        stream=True
+    )
+
+    full_text = ""  # 用于存储最终拼接的字符串
+    async for chunk in completion:
+        if chunk.choices[0].delta.content is not None:
+            for char in chunk.choices[0].delta.content:
+                if char != ' ':
+                    yield char  # 流式输出字符
+                    full_text += char  # 拼接字符
+                    print(char, end='')
+                    time.sleep(0.1)  # 控制输出速度
+
+    # 记录到数据库
+    try:
+        await save_chat_to_mysql(pool, person_id, f'{image_url}', full_text, "", 0, 2, 2, 1)
+    except Exception as e:
+        print(f"记录到数据库时出错：{e}")