'commit'

AI/WxMini/Start.py

@@ -604,15 +604,10 @@ async def process_image(image_url: str, current_user: dict = Depends(get_current
     :return: 流式输出结果
     """
     try:
-        if await is_text_dominant(client, image_url):
-            logger.info("检测到图片主要是文字内容，开始识别文字：")
-            return StreamingResponse(recognize_text(client, app.state.mysql_pool, current_user['person_id'], image_url),
-                                     media_type="text/plain")
-        else:
-            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")
+        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")
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))
 

AI/WxMini/Utils/ImageUtil.py

@@ -1,42 +1,7 @@
 import time
-import re
 
 from WxMini.Utils.MySQLUtil import save_chat_to_mysql
 
-
-async def is_text_dominant(client, image_url):
-    """
-    判断图片是否主要是文字内容
-    :param image_url: 图片 URL
-    :return: True（主要是文字） / False（主要是物体/场景）
-    """
-    completion = await client.chat.completions.create(
-        model="qwen-vl-ocr",
-        messages=[
-            {
-                "role": "user",
-                "content": [
-                    {
-                        "type": "image_url",
-                        "image_url": image_url,
-                        "min_pixels": 28 * 28 * 4,
-                        "max_pixels": 28 * 28 * 1280
-                    },
-                    {"type": "text", "text": "Read all the text in the image."},
-                ]
-            }
-        ],
-        stream=False
-    )
-    text = completion.choices[0].message.content
-
-    # 判断是否只有英文和数字
-    if re.match(r'^[A-Za-z0-9\s]+$', text):
-        print("识别到的内容只有英文和数字，可能是无意义的字符，调用识别内容功能。")
-        return False
-    return True
-
-
 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