PdfProcessing/app.py

import os
import sys
import uuid
import tempfile

import pdfplumber
from flask import Flask, request, jsonify, send_file

from add_signature import find_signature_positions, add_signature_to_pdf
from add_watermark import add_watermark_to_pdf
from extract_table import extract_temp_time, extract_pdf_table_to_excel, extract_temp_by_datetime_pattern, allowed_file, \
    safe_filename, extract_temperature_data_from_pdf, extract_data_from_pdf_5
from lib import Qiniu
from werkzeug.utils import secure_filename
from flask_cors import CORS

app = Flask(__name__)
UPLOAD_FOLDER = tempfile.gettempdir()
app.config['UPLOAD_FOLDER'] = UPLOAD_FOLDER
CORS(app)


@app.route('/add_signature', methods=['POST'])
def add_signature():
    try:
        # 获取请求参数
        data = request.get_json()
        pdf_url = data.get('pdf_url')
        if not pdf_url:
            return jsonify({"error": "缺少pdf_url参数"}), 400

        # 下载PDF文件
        local_pdf = Qiniu.download_file(pdf_url)

        # 查找签名位置（使用默认公章图片）
        signature_img = './static/报告专用章.png'  # 确保项目目录下有这个文件
        positions = find_signature_positions(local_pdf)
        if not positions:
            os.remove(local_pdf)
            return jsonify({"error": "未找到签名位置"}), 400

        # 生成带公章的PDF
        signature_pdf = os.path.join("./temp", f"signature_{os.path.basename(local_pdf)}")
        add_signature_to_pdf(local_pdf, signature_pdf, signature_img, positions)

        # 上传到七牛云
        file_key = f"UpImage/{uuid.uuid4()}.pdf"
        new_pdf_url = Qiniu.upload_to_qiniu(signature_pdf, file_key)

        # 清理临时文件
        os.remove(local_pdf)
        os.remove(signature_pdf)

        return jsonify({
            "success": True,
            "signature_pdf_url": new_pdf_url
        })

    except Exception as e:
        return jsonify({
            "error": str(e),
            "success": False,
        }), 500


@app.route('/add_watermark', methods=['POST'])
def add_watermark():
    try:
        # 获取请求参数
        data = request.get_json()
        pdf_url = data.get('pdf_url')
        if not pdf_url:
            return jsonify({"error": "缺少pdf_url参数"}), 400

        # 下载PDF文件
        local_pdf = Qiniu.download_file(pdf_url)

        # 水印文件
        pdf_file_mark = './static/watermark.pdf'  # 确保项目目录下有这个文件

        # 生成带公章的PDF
        watermark_pdf = os.path.join("./temp", f"watermark_{os.path.basename(local_pdf)}")
        add_watermark_to_pdf(local_pdf, pdf_file_mark, watermark_pdf)

        # 上传到七牛云
        file_key = f"UpImage/{uuid.uuid4()}.pdf"
        new_pdf_url = Qiniu.upload_to_qiniu(watermark_pdf, file_key)

        # 清理临时文件
        os.remove(local_pdf)
        os.remove(watermark_pdf)

        return jsonify({
            "success": True,
            "watermark_pdf_url": new_pdf_url
        })

    except Exception as e:
        return jsonify({
            "error": str(e),
            "success": False,
        }), 500


@app.route('/extract_table', methods=['POST'])
def extract_table():
    if 'file' not in request.files:
        return jsonify({'error': 'No file part'}), 400

    file = request.files['file']

    if file.filename == '':
        return jsonify({'error': 'No selected file'}), 400

    if file and allowed_file(file.filename):
        original_filename = file.filename
        filename = safe_filename(original_filename)
        # filename = secure_filename(safe_name)
        filepath = os.path.join(app.config['UPLOAD_FOLDER'], filename)
        file.save(filepath)
        df = None
        with pdfplumber.open(filepath) as pdf:
            # 获取第一页
            first_page = pdf.pages[0]
            text = first_page.extract_text()
            if text:
                if "温湿度数据报告" in text:
                    df = extract_pdf_table_to_excel(filepath)
                if "历史数据表" in text:
                    df = extract_temp_time(filepath)
                if "设备汇总报告" in text:
                    df = extract_temp_by_datetime_pattern(filepath)
                if "详细数据" in text:
                    df = extract_temperature_data_from_pdf(filepath)
                else:
                    df = extract_data_from_pdf_5(filepath)


        if df is None:
            os.remove(filepath)
            return jsonify({'error': '所有处理方法均失败'}), 400

        # 保存Excel文件
        output_filename = filename.replace('.pdf', '.xlsx')
        output_path = os.path.join(app.config['UPLOAD_FOLDER'], output_filename)
        df.to_excel(output_path, index=False, engine='openpyxl')

        # 删除上传的PDF文件
        os.remove(filepath)

        # 返回Excel文件
        try:
            return send_file(
                output_path,
                as_attachment=True,
                download_name=output_filename,
                mimetype='application/vnd.openxmlformats-officedocument.spreadsheetml.sheet'
            )
        finally:
            # 确保临时文件最终被删除
            try:
                os.remove(output_path)
            except:
                pass
    else:
        return jsonify({'error': 'Invalid file type'}), 400


if __name__ == '__main__':
    print("项目地址：", os.path.dirname(__file__))
    if len(sys.argv) != 2:
        print("请填写端口号")
        sys.exit()
    # app.debug = True  # 设置调试模式，生产模式的时候要关掉debug
    # app.config['JSON_AS_ASCII'] = False
    app.run(host='0.0.0.0', port=6500, debug=True)