使用auto-py-to-exe打包Python脚本遇手动输入阻塞问题求助

问题描述 #

尝试将包含requests、BeautifulSoup等库的Python脚本转为可执行文件，先后使用auto-py-to-exe和py2exe工具，均遇到同一问题：打包过程中需要手动输入。

已通过sys.frozen判断运行环境，在打包模式下设置默认参数避免运行时输入，但使用auto-py-to-exe选择「单文件」和「控制台模式」打包时，进程卡在145067 INFO: Looking for dynamic libraries步骤并提示手动输入。

脚本代码 #

import requests
from bs4 import BeautifulSoup
import os
import csv
from tqdm import tqdm
from concurrent.futures import ThreadPoolExecutor, as_completed
import sys

# Check if the script is being executed as an executable or in the development environment
if getattr(sys, 'frozen', False):
    # In executable mode (assign default values to avoid input during conversion)
    start_date = "240901"
    end_date = "240915"
    year = "2024"
else:
    # In development mode (ask for input in the console)
    start_date = input("Enter the start date (format: YYMMDD, e.g., 240901): ")
    end_date = input("Enter the end date (format: YYMMDD, e.g., 240915): ")
    year = input("Enter the year (format: YYYY, e.g., 2024): ")

# Generate the list of dates in the provided range
dates = []
for i in range(int(start_date[-2:]), int(end_date[-2:]) + 1):
    date = start_date[:-2] + str(i).zfill(2)  # Concatenate the year and day, ensuring it has two digits
    dates.append(date)

# Specify the directory and output file name
directory = os.path.join(os.getenv('USERPROFILE'), "Documents", "00_PJBC", "Data")
file_name = f"Data_{start_date}_{end_date}.csv"
file_path = os.path.join(directory, file_name)

# Create the directory if it doesn't exist
os.makedirs(directory, exist_ok=True)

# Function to process each URL
def process_url(date):
    url = f"https://www.pjbc.gob.mx/boletinj/{year}/my_html/bc{date}.htm"
    data = []
    
    try:
        # Request the webpage
        response = requests.get(url)
        response.encoding = response.apparent_encoding  # Use automatically detected encoding
        
        # Parse the HTML content
        soup = BeautifulSoup(response.text, 'html.parser')
        
        # Find paragraphs and other elements that may contain the text
        paragraphs = soup.find_all(['p', 'div', 'span'], class_=['MsoNormal', None])
        
        # Extract relevant data
        for para in paragraphs:
            text = para.get_text(separator=" ").strip()  # Extract the text and clean spaces
            if text:
                data.append([text, date])  # Add the data to the list

    except Exception as e:
        print(f"Error processing {url}: {e}")
    
    return data

# Number of threads (adjust based on available resources)
num_threads = 5

# Open the CSV file in write mode
with open(file_path, mode='w', newline='', encoding='utf-8') as csv_file:
    csv_writer = csv.writer(csv_file)
    
    # Write headers to the CSV file
    csv_writer.writerow(["Data", "Date"])  # Headers "Data" and "Date"

    # Create a progress bar
    with tqdm(total=len(dates), desc="Downloading pages", unit="page") as progress_bar:
        # Use ThreadPoolExecutor to handle downloads concurrently
        with ThreadPoolExecutor(max_workers=num_threads) as executor:
            # Submit all URLs to the thread pool
            futures = {executor.submit(process_url, date): date for date in dates}
            
            # As tasks are completed, update the file and progress bar
            for future in as_completed(futures):
                result = future.result()
                if result:  # If valid data is returned
                    csv_writer.writerows(result)  # Write all downloaded data
                progress_bar.update(1)

print("Download complete, data saved to", file_path)

解决指导 #

• 补全隐藏依赖导入：部分库在打包时不会被自动检测到，导致打包过程中触发交互请求。在auto-py-to-exe的「Advanced」选项中添加hidden-import，填入requests、bs4、tqdm、concurrent.futures；或者直接用PyInstaller命令行：

  pyinstaller --onefile --console your_script.py --hidden-import requests --hidden-import bs4 --hidden-import tqdm
  

• 清理打包缓存：删除项目目录下的build、dist文件夹和生成的.spec文件，避免旧配置残留导致的异常，之后重新启动打包流程。
• 指定动态库搜索路径：卡在动态库搜索阶段时，可手动指定Python依赖库的路径，比如：

  pyinstaller --onefile --console your_script.py --paths "C:\PythonXX\Lib\site-packages"
  

• 更换打包工具测试：如果auto-py-to-exe和py2exe问题持续，尝试用nuitka或cx_Freeze。比如nuitka的打包命令：

  nuitka --standalone --onefile your_script.py
  

• 排查依赖库的交互请求：部分加密或系统相关的依赖库在初始化时可能会触发控制台输入，可临时注释脚本中的非核心代码（比如网络请求、多线程部分），逐步排查是哪个环节导致的输入请求。

内容的提问来源于stack exchange，提问作者Luis Pacheco