app.py

import os
import sys
import logging
from datetime import datetime
import torch
from transformers import AutoProcessor, Pix2StructForConditionalGeneration
from PIL import Image

# 配置日志
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s',
    handlers=[
        logging.StreamHandler(sys.stdout),
        logging.FileHandler('app.log')
    ]
)
logger = logging.getLogger(__name__)

class ChartAnalyzer:
    def __init__(self):
        try:
            logger.info("Initializing model and processor...")
            self.model = Pix2StructForConditionalGeneration.from_pretrained("google/deplot")
            self.processor = AutoProcessor.from_pretrained("google/deplot")
            logger.info("Model and processor initialized successfully")
        except Exception as e:
            logger.error(f"Error initializing model: {str(e)}")
            raise

    def process_image(self, image_path, prompt=None):
        """处理图片并生成数据表格"""
        try:
            # 验证文件存在
            if not os.path.exists(image_path):
                raise FileNotFoundError(f"Image file not found: {image_path}")

            # 打开并处理图片
            logger.info(f"Processing image: {image_path}")
            image = Image.open(image_path)
            
            # 准备输入
            if prompt is None:
                prompt = "Generate underlying data table of the figure below:"
            
            inputs = self.processor(
                images=image, 
                text=prompt,
                return_tensors="pt"
            )

            # 生成预测
            logger.info("Generating predictions...")
            with torch.no_grad():  # 提高性能并减少内存使用
                predictions = self.model.generate(
                    **inputs,
                    max_new_tokens=512,
                    num_beams=4,
                    length_penalty=1.0
                )

            # 解码预测结果
            raw_output = self.processor.decode(predictions[0], skip_special_tokens=True)
            
            # 处理结果
            split_by_newline = raw_output.split("<0x0A>")
            result_array = []
            for item in split_by_newline:
                if item.strip():  # 跳过空行
                    result_array.append([x.strip() for x in item.split("|")])

            # 保存结果
            timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
            output_file = f'results_{timestamp}.log'
            
            with open(output_file, mode='w', encoding='utf-8') as file:
                for row in result_array:
                    file.write(" | ".join(row) + "\n")
            
            logger.info(f"Results saved to {output_file}")
            return result_array

        except Exception as e:
            logger.error(f"Error processing image: {str(e)}")
            raise

def main():
    try:
        # 创建分析器实例
        analyzer = ChartAnalyzer()
        
        # 指定图片路径（在Space中使用上传的图片路径）
        image_path = '05e57f1c9acff69f1eb6fa72d4805d0.jpg'
        
        # 处理图片
        results = analyzer.process_image(image_path)
        
        # 打印结果
        print("\nAnalysis Results:")
        for row in results:
            print(" | ".join(row))
            
    except Exception as e:
        logger.error(f"Application error: {str(e)}")
        raise

if __name__ == "__main__":
    main()