QingLong/AI/WxMini/Utils/ImageUtil.py

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
    :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
    """
    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=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
                    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}")


async def recognize_content(client, pool, person_id, image_url):
    """
    识别图片中的内容，流式输出
    """
    completion = await client.chat.completions.create(
        model="qwen-vl-plus",
        messages=[{"role": "user", "content": [
            {"type": "text", "text": "这是什么"},
            {"type": "image_url", "image_url": {"url": image_url}}
        ]}],
        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
                    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, 2)
    except Exception as e:
        print(f"记录到数据库时出错：{e}")