新增 release-notes.py 用于爬取 Github 上的 release note 信息，使用了 ChatGPT 辅助。

python3/release-notes.py

@@ -0,0 +1,70 @@
+import argparse
+import re
+from bs4 import BeautifulSoup
+import requests
+
+
+def parse_github_repo(github_url):
+    pattern = r"https://github.com/([^/]+)/([^/]+)"
+    match = re.match(pattern, github_url)
+    if not match:
+        raise ValueError("Invalid GitHub URL")
+    repo_owner, repo_name = match.groups()
+    return repo_owner, repo_name
+
+
+def collect_release_notes(github_url):
+    releases = []
+    page = 1
+
+    # Extract the repository owner and name from the GitHub URL
+    repo_owner, repo_name = parse_github_repo(github_url)
+
+    while True:
+        print(f"Processing Page {page}")
+        # Make a GET request to the GitHub release page
+        url = f"https://github.com/{repo_owner}/{repo_name}/releases?page={page}"
+        response = requests.get(url)
+
+        # Check if the request was successful
+        if response.status_code != 200:
+            raise Exception(f"Failed to retrieve releases from the repository: {response.text}")
+
+        # Parse the HTML content
+        soup = BeautifulSoup(response.text, "html.parser")
+
+        # Find all of the release items on the page
+        release_headers = soup.select("#repo-content-turbo-frame span a")
+        release_bodies = soup.select(".Box-body > :nth-child(2)")
+        if not release_headers:
+            break
+
+        # Extract the release information
+        for header, body in zip(release_headers, release_bodies):
+            name = header.text
+            content = body.text
+            releases.append({"name": name, "body": content})
+
+        page += 1
+
+    print("done")
+    return releases
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("github_url", help="GitHub repository URL")
+    parser.add_argument("output_file", nargs="?", default="", help="Output file name")
+    args = parser.parse_args()
+
+    github_url = args.github_url
+    _, repo_name = parse_github_repo(github_url)
+    output_file = args.output_file or f"{repo_name}.md"
+
+    releases = collect_release_notes(github_url)
+
+    # Write the release notes to a file
+    with open(output_file, "w") as file:
+        for release in releases:
+            file.write("# " + release["name"] + "\n")
+            file.write(release["body"] + "\n")