比较完善的批量下载github的release文件的python脚本

一、原始版本参考：

https://www.xjoker.us/其他/批量下载Github指定仓库的所有Release/

二、增加如下功能：

1、分标签创建下一级目录

2、支持过滤不必要的标签

3、支持指定同时下载线程个数

4、支持github代理，代理格式类似：https://ghproxy.com/https://github.com/xxxx/yyyy

三、v1完整代码如下：

import sys
import requests
from requests.adapters import HTTPAdapter
from urllib3.util.retry import Retry
from urllib.parse import urlparse
import concurrent.futures
from pathlib import Path
import re
import argparse

def create_session(retries=5, backoff_factor=0.3, status_formalist=(500, 502, 504)):
    """Create and configure a requests session for automatic retries."""
    session = requests.Session()
    retry_strategy = Retry(
        total=retries,
        read=retries,
        connect=retries,
        backoff_factor=backoff_factor,
        status_forcelist=status_formalist,
    )
    adapter = HTTPAdapter(max_retries=retry_strategy)
    session.mount('http://', adapter)
    session.mount('https://', adapter)
    return session


def extract_user_repo(github_clone_url):
    """Extract username and repository name from a GitHub clone URL, supporting proxy URLs and no .git suffix."""
    parsed = urlparse(github_clone_url)

    # 检查是否为代理格式（如 ghproxy.com/https://github.com/user/repo）
    use_proxy = parsed.netloc and 'github.com' not in parsed.netloc

    # 如果没有使用代理
    path_parts = parsed.path.strip('/').split('/')
    if not use_proxy:
        user = path_parts[0]
        repo = path_parts[1].split('#')[0].split('?')[0]  # 去除 query 或 fragment
        repo = repo.rstrip('.git')  # 移除 .git（如果存在）
        return user, repo

    # 默认处理原始 GitHub 格式（如 github.com/user/repo 或 user/repo）
    path = parsed.path.strip('/').split('#')[0].split('?')[0]  # 去除 query 或 fragment
    parts = path.split('/')
    if len(parts) < 2:
        raise ValueError(f"Invalid GitHub clone URL: {github_clone_url}")

    user = parts[3]
    repo = parts[4].rstrip('.git')  # 移除 .git（如果存在）
    return user, repo

def download_asset(session, asset_url, file_path):
    """Download a single Release asset."""
    print(f"Downloading {file_path.name}...")
    try:
        response = session.get(asset_url, stream=True)
        response.raise_for_status()
        with file_path.open('wb') as file:
            for chunk in response.iter_content(chunk_size=8192):
                file.write(chunk)
        print(f"Downloaded {file_path.name}")
    except requests.RequestException as e:
        print(f"Failed to download {file_path.name}: {e}")

def download_releases(github_clone_url, filter_tag_regex=None, threads=2):
    """Download all Releases for a specified repository and organize them by tag."""
    session = create_session()
    user, repo = extract_user_repo(github_clone_url)
    print(f"Downloading releases for {user}/{repo}...")

    # API 请求始终使用官方地址
    api_url = f"https://api.github.com/repos/{user}/{repo}/releases"

    # 判断是否使用代理（即 URL 是否为代理格式）
    parsed_input = urlparse(github_clone_url)
    use_proxy = parsed_input.netloc and 'github.com' not in parsed_input.netloc

    directory = Path.cwd() / repo
    directory.mkdir(exist_ok=True)

    page = 1
    total_assets = 0
    futures = []
    with concurrent.futures.ThreadPoolExecutor(max_workers=threads) as executor:
        while True:
            params = {'per_page': 100, 'page': page}
            response = session.get(api_url, params=params)
            releases = response.json()

            if not releases:
                break

            for release in releases:
                tag_name = release['tag_name']
                if filter_tag_regex and not re.match(filter_tag_regex, tag_name):
                    print(f"Skipping tag: {tag_name} (does not match regex: {filter_tag_regex})")
                    continue

                tag_directory = directory / tag_name
                tag_directory.mkdir(exist_ok=True)

                for asset in release['assets']:
                    original_asset_url = asset['browser_download_url']

                    # 使用提前判断的 use_proxy 变量
                    if use_proxy:
                        asset_url = f"{parsed_input.scheme}://{parsed_input.netloc}/{original_asset_url}"
                    else:
                        asset_url = original_asset_url

                    total_assets += 1
                    futures.append(
                        executor.submit(download_asset, session, asset_url, tag_directory / asset['name'])
                    )

            page += 1

        concurrent.futures.wait(futures)

    print(f"Total assets downloaded: {total_assets}")

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Download GitHub Release assets filtered by tag.")
    parser.add_argument("repo_url", help="GitHub repository clone URL (HTTPS or SSH)")
    parser.add_argument("-f", "--filter-tag", help="Regex pattern to filter tags (e.g., '^v\\d+\\.\\d+\\.\\d+$')")
    parser.add_argument("-t", "--threads", type=int, default=2, help="Number of threads to use for downloading (default: 2)")

    args = parser.parse_args()

    download_releases(args.repo_url, args.filter_tag, args.threads)

执行时的参数

python github_release.py https://ghproxy.com/https://github.com/xxx/yyyy --filter-tag=^v\\d+\\.\\d+\\.\\d+$ --threads=4

四、v2版本增加功能：

1、增加下载文件大小校验，下载出错会自动重试

2、支持参数传入重试次数

五、相应代码：

import sys
import requests
from requests.adapters import HTTPAdapter
from urllib3.util.retry import Retry
from urllib.parse import urlparse
import concurrent.futures
from pathlib import Path
import re
import argparse

def create_session(retries=5, backoff_factor=0.3, status_formalist=(500, 502, 504)):
    """Create and configure a requests session for automatic retries."""
    session = requests.Session()
    retry_strategy = Retry(
        total=retries,
        read=retries,
        connect=retries,
        backoff_factor=backoff_factor,
        status_forcelist=status_formalist,
    )
    adapter = HTTPAdapter(max_retries=retry_strategy)
    session.mount('http://', adapter)
    session.mount('https://', adapter)
    return session


def extract_user_repo(github_clone_url):
    """Extract username and repository name from a GitHub clone URL, supporting proxy URLs and no .git suffix."""
    parsed = urlparse(github_clone_url)

    # 检查是否为代理格式（如 ghproxy.com/https://github.com/user/repo）
    use_proxy = parsed.netloc and 'github.com' not in parsed.netloc

    # 如果没有使用代理
    path_parts = parsed.path.strip('/').split('/')
    if not use_proxy:
        user = path_parts[0]
        repo = path_parts[1].split('#')[0].split('?')[0]  # 去除 query 或 fragment
        repo = repo.rstrip('.git')  # 移除 .git（如果存在）
        return user, repo

    # 默认处理原始 GitHub 格式（如 github.com/user/repo 或 user/repo）
    path = parsed.path.strip('/').split('#')[0].split('?')[0]  # 去除 query 或 fragment
    parts = path.split('/')
    if len(parts) < 2:
        raise ValueError(f"Invalid GitHub clone URL: {github_clone_url}")

    user = parts[3]
    repo = parts[4].rstrip('.git')  # 移除 .git（如果存在）
    return user, repo

def download_asset(session, asset_url, file_path, expected_size, max_retries):
    """Download a single Release asset with retry on failure or size mismatch."""
    attempt = 0
    while True:
        attempt += 1
        if max_retries != 0 and attempt > max_retries:
            print(f"Failed to download {file_path.name} after {max_retries} attempts.")
            return False

        print(f"Attempt {attempt} for downloading: {file_path.name}")
        try:
            with file_path.open('wb') as file:
                response = session.get(asset_url, stream=True)
                response.raise_for_status()
                downloaded_size = 0
                for chunk in response.iter_content(chunk_size=8192):
                    file.write(chunk)
                    downloaded_size += len(chunk)

                # 校验文件大小
                if downloaded_size == expected_size:
                    print(f"Downloaded {file_path.name}")
                    return True
                else:
                    print(f"Size mismatch (expected {expected_size}, got {downloaded_size}), retrying...")
        except (requests.RequestException, IOError) as e:
            print(f"Attempt {attempt} failed: {e}")

        # 删除损坏文件
        if file_path.exists():
            file_path.unlink()

def download_releases(github_clone_url, filter_tag_regex=None, threads=2, max_retries=3):
    """Download all Releases for a specified repository and organize them by tag."""
    session = create_session()
    user, repo = extract_user_repo(github_clone_url)
    print(f"Downloading releases for {user}/{repo}...")

    # API 请求始终使用官方地址
    api_url = f"https://api.github.com/repos/{user}/{repo}/releases"

    # 判断是否使用代理（即 URL 是否为代理格式）
    parsed_input = urlparse(github_clone_url)
    use_proxy = parsed_input.netloc and 'github.com' not in parsed_input.netloc

    directory = Path.cwd() / repo
    directory.mkdir(exist_ok=True)

    page = 1
    total_assets = 0
    futures = []
    with concurrent.futures.ThreadPoolExecutor(max_workers=threads) as executor:
        while True:
            params = {'per_page': 100, 'page': page}
            response = session.get(api_url, params=params)
            releases = response.json()

            if not releases:
                break

            for release in releases:
                tag_name = release['tag_name']
                if filter_tag_regex and not re.match(filter_tag_regex, tag_name):
                    print(f"Skipping tag: {tag_name} (does not match regex: {filter_tag_regex})")
                    continue

                tag_directory = directory / tag_name
                tag_directory.mkdir(exist_ok=True)

                for asset in release['assets']:
                    original_asset_url = asset['browser_download_url']
                    expected_size = asset['size']

                    # 使用提前判断的 use_proxy 变量
                    if use_proxy:
                        asset_url = f"{parsed_input.scheme}://{parsed_input.netloc}/{original_asset_url}"
                    else:
                        asset_url = original_asset_url

                    total_assets += 1
                    futures.append(
                        executor.submit(download_asset, session, asset_url, tag_directory / asset['name'], expected_size, max_retries)
                    )

            page += 1

        concurrent.futures.wait(futures)

    print(f"Total assets downloaded: {total_assets}")

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Download GitHub Release assets filtered by tag.")
    parser.add_argument("repo_url", help="GitHub repository clone URL (HTTPS or SSH)")
    parser.add_argument("-f", "--filter-tag", help="Regex pattern to filter tags (e.g., '^v\\d+\\.\\d+\\.\\d+$')")
    parser.add_argument("-t", "--threads", type=int, default=2, help="Number of threads to use for downloading (default: 2)")
    parser.add_argument("-r", "--retries", type=int, default=3, help="Maximum number of retries for each file (0 means infinite).")

    args = parser.parse_args()

    download_releases(args.repo_url, args.filter_tag, args.threads, args.retries)

六、v3版本增加功能

1、支持文件名以正则表达式规则进行过滤，指定了过滤规则时，不匹配的文件会被直接跳过

v3相应代码如下：

import sys
import requests
from requests.adapters import HTTPAdapter
from urllib3.util.retry import Retry
from urllib.parse import urlparse
import concurrent.futures
from pathlib import Path
import re
import argparse

def create_session(retries=5, backoff_factor=0.3, status_formalist=(500, 502, 504)):
    """Create and configure a requests session for automatic retries."""
    session = requests.Session()
    retry_strategy = Retry(
        total=retries,
        read=retries,
        connect=retries,
        backoff_factor=backoff_factor,
        status_forcelist=status_formalist,
    )
    adapter = HTTPAdapter(max_retries=retry_strategy)
    session.mount('http://', adapter)
    session.mount('https://', adapter)
    return session


def extract_user_repo(github_clone_url):
    """Extract username and repository name from a GitHub clone URL, supporting proxy URLs and no .git suffix."""
    parsed = urlparse(github_clone_url)

    # 检查是否为代理格式（如 ghproxy.com/https://github.com/user/repo）
    use_proxy = parsed.netloc and 'github.com' not in parsed.netloc

    # 如果没有使用代理
    path_parts = parsed.path.strip('/').split('/')
    if not use_proxy:
        user = path_parts[0]
        repo = path_parts[1].split('#')[0].split('?')[0]  # 去除 query 或 fragment
        repo = repo.rstrip('.git')  # 移除 .git（如果存在）
        return user, repo

    # 默认处理原始 GitHub 格式（如 github.com/user/repo 或 user/repo）
    path = parsed.path.strip('/').split('#')[0].split('?')[0]  # 去除 query 或 fragment
    parts = path.split('/')
    if len(parts) < 2:
        raise ValueError(f"Invalid GitHub clone URL: {github_clone_url}")

    user = parts[3]
    repo = parts[4].rstrip('.git')  # 移除 .git（如果存在）
    return user, repo

def download_asset(session, asset_url, file_path, expected_size, max_retries):
    """Download a single Release asset with retry on failure or size mismatch."""
    attempt = 0
    while True:
        attempt += 1
        if max_retries != 0 and attempt > max_retries:
            print(f"Failed to download {file_path.name} after {max_retries} attempts.")
            return False

        print(f"Attempt {attempt} for downloading: {file_path.name}")
        try:
            with file_path.open('wb') as file:
                response = session.get(asset_url, stream=True)
                response.raise_for_status()
                downloaded_size = 0
                for chunk in response.iter_content(chunk_size=8192):
                    file.write(chunk)
                    downloaded_size += len(chunk)

                # 校验文件大小
                if downloaded_size == expected_size:
                    print(f"Downloaded {file_path.name}")
                    return True
                else:
                    print(f"Size mismatch (expected {expected_size}, got {downloaded_size}), retrying...")
        except (requests.RequestException, IOError) as e:
            print(f"Attempt {attempt} failed: {e}")

        # 删除损坏文件
        if file_path.exists():
            file_path.unlink()

def download_releases(github_clone_url, filter_tag_regex=None, threads=2, max_retries=3, filter_name=None):
    """Download all Releases for a specified repository and organize them by tag."""
    session = create_session()
    user, repo = extract_user_repo(github_clone_url)
    print(f"Downloading releases for {user}/{repo}...")

    # API 请求始终使用官方地址
    api_url = f"https://api.github.com/repos/{user}/{repo}/releases"

    # 判断是否使用代理（即 URL 是否为代理格式）
    parsed_input = urlparse(github_clone_url)
    use_proxy = parsed_input.netloc and 'github.com' not in parsed_input.netloc

    directory = Path.cwd() / repo
    directory.mkdir(exist_ok=True)

    page = 1
    total_assets = 0
    futures = []
    with concurrent.futures.ThreadPoolExecutor(max_workers=threads) as executor:
        while True:
            params = {'per_page': 100, 'page': page}
            response = session.get(api_url, params=params)
            releases = response.json()

            if not releases:
                break

            for release in releases:
                tag_name = release['tag_name']
                # 过滤一下标签
                if filter_tag_regex and not re.match(filter_tag_regex, tag_name):
                    print(f"Skipping tag: {tag_name} (does not match regex: {filter_tag_regex})")
                    continue

                tag_directory = directory / tag_name
                tag_directory.mkdir(exist_ok=True)

                for asset in release['assets']:
                    original_asset_url = asset['browser_download_url']
                    expected_size = asset['size']
                    asset_name = asset['name']

                    # 新增：根据文件名正则过滤
                    if filter_name and not re.match(filter_name, asset_name):
                        print(f"Skipping asset: {asset_name} (does not match regex: {filter_name})")
                        continue

                    # 使用提前判断的 use_proxy 变量
                    if use_proxy:
                        asset_url = f"{parsed_input.scheme}://{parsed_input.netloc}/{original_asset_url}"
                    else:
                        asset_url = original_asset_url

                    total_assets += 1
                    futures.append(
                        executor.submit(download_asset, session, asset_url, tag_directory / asset['name'], expected_size, max_retries)
                    )

            page += 1

        concurrent.futures.wait(futures)

    print(f"Total assets downloaded: {total_assets}")

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Download GitHub Release assets filtered by tag.")
    parser.add_argument("repo_url", help="GitHub repository clone URL (HTTPS or SSH)")
    parser.add_argument("-f", "--filter-tag", help="Regex pattern to filter tags (e.g., '^v\\d+\\.\\d+\\.\\d+$')")
    parser.add_argument("-t", "--threads", type=int, default=2, help="Number of threads to use for downloading (default: 2)")
    parser.add_argument("-r", "--retries", type=int, default=3, help="Maximum number of retries for each file (0 means infinite).")
    parser.add_argument("-n", "--filter-name", help="Regex pattern to filter asset filenames (e.g., '.*\\.zip$')")

    args = parser.parse_args()

    download_releases(args.repo_url, args.filter_tag, args.threads, args.retries, args.filter_name)

执行时的命令参数示例：

python github_release.py https://ghproxy.com/https://github.com/xxx/yyyy --filter-tag=^v\\d+\\.\\d+\\.\\d+$ --threads=4 --filter-name='.*\.(7z|dmg|zip|deb|msi)$'