How to Batch Download Images with Python: From XPath Extraction to Automated Saving

1. Introduction

The previous article covered the theory of image crawling; this guide completes the process by showing how to batch‑download images from a website using Python.

2. Image URL Parsing

1. Open the target page, right‑click an image and inspect its element to locate the src attribute.

2. Isolate the src value and note its parent <ul> (or higher) tag.

3. Use an XPath selector to extract the src value, prepend the appropriate https:// prefix, and obtain the full image URL.

4. Run the XPath query in Python to retrieve the URLs, as illustrated in the screenshots.

3. Downloading Images

1. Create a filename (or folder) to store the downloaded images, e.g., a directory named 天堂爬的图片.

2. Ensure the folder exists in the same directory as your script; otherwise, the script will raise an error.

3. Use a with open(filename, "wb") as f block to write binary image data to disk. Example code:

with open(filename, "wb") as f:
    f.write(html)

4. The snippet explains the meaning of the file mode "wb" (write binary) and the variable f.

5. Install and import fake_useragent to generate random request headers, reducing the chance of being blocked:

from fake_useragent import UserAgent
ua = UserAgent()
print(ua.ie)       # random IE version
print(ua.firefox)  # random Firefox version
print(ua.chrome)   # random Chrome version
print(ua.random)   # random user‑agent string

6. Initialize the UserAgent object and select a random header (e.g., ua.random) for each request, optionally choosing from a pool of 50 agents.

7. Execute the download loop; the terminal output shows successful retrieval of each image.

8. After the script finishes, all images are saved locally in high resolution.

4. Conclusion

The article demonstrates a practical workflow using Python's requests , lxml , and fake_useragent libraries to parse webpage structures, extract image URLs, and perform batch downloads efficiently. Readers are encouraged to try the method while respecting server load.