app.py

import gradio as gr
import pytesseract
from PIL import Image
import requests
import re
import traceback
import os

# 配置 Tesseract OCR 的路径（Hugging Face Spaces 自动配置）
pytesseract.pytesseract.tesseract_cmd = '/usr/bin/tesseract'

# 使用环境变量获取 Hugging Face API Token
API_URL = "/static-proxy?url=https%3A%2F%2Fapi-inference.huggingface.co%2Fmodels%2FQwen%2FQwen2.5-Math-72B-Instruct%26quot%3B%3C%2Fspan%3E%3C!-- HTML_TAG_END -->
API_TOKEN = os.getenv("HF_API_TOKEN")  # 从环境变量获取 Token
HEADERS = {"Authorization": f"Bearer {API_TOKEN}"}

# OCR 识别函数
def ocr_with_tesseract(image_path):
    try:
        image = Image.open(image_path).convert("L")
        config = "--psm 6"
        text = pytesseract.image_to_string(image, config=config)
        text = re.sub(r'[^0-9a-zA-Z=+\-*/()., ]', '', text)
        return text if text else "OCR 识别失败"
    except Exception as e:
        return f"OCR 识别错误: {e}\n{traceback.format_exc()}"

# AI 解答生成函数
def generate_solution_with_qwen(question):
    prompt = f"请详细解答以下数学题目：{question}"
    payload = {"inputs": prompt}
    response = requests.post(API_URL, headers=HEADERS, json=payload)
    
    if response.status_code == 200:
        result = response.json()
        return result.get('generated_text', "解答生成失败")
    else:
        return f"API 调用失败，状态码: {response.status_code}, 响应: {response.text}"

# 主处理函数
def process(image_path):
    ocr_result = ocr_with_tesseract(image_path)
    ai_solution = generate_solution_with_qwen(ocr_result)
    return ocr_result, ai_solution

# 构建 Gradio 应用界面
def build_interface():
    with gr.Blocks() as interface:
        gr.Markdown("# 📚 高级 AI 数学解题助手")
        image_input = gr.Image(type="filepath", label="上传数学题目图片")
        ocr_output = gr.Textbox(label="OCR 识别结果")
        ai_output = gr.Markdown(label="AI 解答")
        submit_button = gr.Button("识别并解答")
        submit_button.click(fn=process, inputs=image_input, outputs=[ocr_output, ai_output])
    return interface

# 启动 Gradio 应用
interface = build_interface()
interface.launch()