Update app.py
Browse files
app.py
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
import re
|
2 |
-
import logging
|
3 |
import os
|
|
|
4 |
from flask import Flask, request, Response
|
5 |
import requests
|
6 |
from urllib.parse import urlparse, unquote
|
@@ -9,15 +9,17 @@ from urllib.parse import urlparse, unquote
|
|
9 |
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
|
10 |
app = Flask(__name__)
|
11 |
|
12 |
-
# ---
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
|
20 |
# --- Whitelisted URL Patterns for GitHub ---
|
|
|
|
|
21 |
ALLOWED_PATTERNS = [
|
22 |
# Repositories: releases, archives, blobs, raw content
|
23 |
re.compile(r'^https://github\.com/[^/]+/[^/]+/(?:releases|archive)/.*$', re.IGNORECASE),
|
@@ -40,8 +42,8 @@ ALLOWED_PATTERNS = [
|
|
40 |
re.compile(r'^https://github\.com/[^/]+/[^/]+/?$', re.IGNORECASE),
|
41 |
]
|
42 |
|
43 |
-
# --- Custom Index Page
|
44 |
-
INDEX_PAGE_HTML =
|
45 |
<!DOCTYPE html>
|
46 |
<html lang="en">
|
47 |
<head>
|
@@ -49,25 +51,32 @@ INDEX_PAGE_HTML = f"""
|
|
49 |
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
50 |
<title>Private GitHub Proxy</title>
|
51 |
<style>
|
52 |
-
body {
|
53 |
-
.container {
|
54 |
-
.warning {
|
55 |
-
|
56 |
-
|
|
|
57 |
</style>
|
58 |
</head>
|
59 |
<body>
|
60 |
<div class="container">
|
61 |
<h1>Private GitHub Reverse Proxy</h1>
|
|
|
|
|
|
|
62 |
<div class="warning">
|
63 |
-
<strong>
|
|
|
64 |
</div>
|
65 |
<h2>How to Use</h2>
|
66 |
-
<p>To access GitHub content,
|
67 |
<p>For example, to clone a repository:</p>
|
68 |
-
<code>git clone {
|
69 |
<p>Or to view a repository page:</p>
|
70 |
-
<code>{
|
|
|
|
|
71 |
</div>
|
72 |
</body>
|
73 |
</html>
|
@@ -80,40 +89,61 @@ def is_url_allowed(url):
|
|
80 |
return True
|
81 |
return False
|
82 |
|
83 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
84 |
@app.route('/', defaults={'path': ''}, methods=['GET', 'POST', 'PUT', 'DELETE'])
|
85 |
@app.route('/<path:path>', methods=['GET', 'POST', 'PUT', 'DELETE'])
|
86 |
def proxy(path):
|
87 |
"""
|
88 |
-
|
89 |
-
|
90 |
"""
|
91 |
-
#
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
96 |
-
#
|
97 |
-
|
98 |
-
|
99 |
-
return "<h1>401 Unauthorized</h1><p>A valid secret key is required in the URL path.</p>", 401
|
100 |
-
|
101 |
-
# If the key is correct but there is no target URL, show the index page.
|
102 |
-
# This happens when accessing /<secret_key>/
|
103 |
-
if len(path_parts) == 1 or not path_parts[1]:
|
104 |
return INDEX_PAGE_HTML, 200
|
105 |
|
106 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
107 |
|
|
|
|
|
|
|
|
|
108 |
# Prepend 'https://' if the scheme is missing.
|
109 |
-
if not
|
110 |
-
target_url = 'https://' +
|
111 |
else:
|
112 |
-
target_url =
|
113 |
|
114 |
# Security check: Ensure the URL is in the whitelist.
|
115 |
if not is_url_allowed(target_url):
|
116 |
-
logging.warning(f"URL Denied!
|
117 |
return "<h1>403 Forbidden</h1><p>Request blocked by proxy security policy.</p>", 403
|
118 |
|
119 |
try:
|
@@ -129,19 +159,27 @@ def proxy(path):
|
|
129 |
headers['Host'] = target_host
|
130 |
|
131 |
try:
|
|
|
|
|
|
|
|
|
132 |
resp = requests.request(
|
133 |
method=request.method,
|
134 |
url=target_url,
|
135 |
headers=headers,
|
136 |
data=request.get_data(),
|
137 |
cookies=request.cookies,
|
138 |
-
allow_redirects=False,
|
139 |
stream=True,
|
140 |
-
timeout=30
|
141 |
)
|
|
|
|
|
142 |
excluded_headers = ['content-encoding', 'content-length', 'transfer-encoding', 'connection']
|
143 |
response_headers = [(name, value) for (name, value) in resp.raw.headers.items()
|
144 |
if name.lower() not in excluded_headers]
|
|
|
|
|
145 |
return Response(resp.iter_content(chunk_size=8192), resp.status_code, response_headers)
|
146 |
|
147 |
except requests.exceptions.RequestException as e:
|
@@ -149,4 +187,4 @@ def proxy(path):
|
|
149 |
return "An error occurred while proxying the request.", 502
|
150 |
|
151 |
if __name__ == '__main__':
|
152 |
-
app.run(host='0.0.0.0', port=7860)
|
|
|
1 |
import re
|
|
|
2 |
import os
|
3 |
+
import logging
|
4 |
from flask import Flask, request, Response
|
5 |
import requests
|
6 |
from urllib.parse import urlparse, unquote
|
|
|
9 |
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
|
10 |
app = Flask(__name__)
|
11 |
|
12 |
+
# --- Load Secret Key from Environment ---
|
13 |
+
PROXY_SECRET_KEY = os.environ.get('PROXY_SECRET_KEY')
|
14 |
+
if not PROXY_SECRET_KEY:
|
15 |
+
logging.error("PROXY_SECRET_KEY environment variable is not set!")
|
16 |
+
exit(1)
|
17 |
+
|
18 |
+
logging.info(f"Proxy service initialized with secret key authentication")
|
19 |
|
20 |
# --- Whitelisted URL Patterns for GitHub ---
|
21 |
+
# This list defines all URL patterns that are permitted to be proxied.
|
22 |
+
# It's a security measure to prevent the proxy from being used to access unintended domains.
|
23 |
ALLOWED_PATTERNS = [
|
24 |
# Repositories: releases, archives, blobs, raw content
|
25 |
re.compile(r'^https://github\.com/[^/]+/[^/]+/(?:releases|archive)/.*$', re.IGNORECASE),
|
|
|
42 |
re.compile(r'^https://github\.com/[^/]+/[^/]+/?$', re.IGNORECASE),
|
43 |
]
|
44 |
|
45 |
+
# --- Custom Index Page ---
|
46 |
+
INDEX_PAGE_HTML = """
|
47 |
<!DOCTYPE html>
|
48 |
<html lang="en">
|
49 |
<head>
|
|
|
51 |
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
52 |
<title>Private GitHub Proxy</title>
|
53 |
<style>
|
54 |
+
body { font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif; line-height: 1.6; color: #333; max-width: 800px; margin: 40px auto; padding: 20px; }
|
55 |
+
.container { border: 1px solid #ddd; border-radius: 8px; padding: 20px 40px; background-color: #f9f9f9; }
|
56 |
+
.warning { color: #856404; background-color: #fff3cd; border: 1px solid #ffeeba; padding: 15px; border-radius: 4px; margin-bottom: 20px; }
|
57 |
+
.security { color: #721c24; background-color: #f8d7da; border: 1px solid #f5c6cb; padding: 15px; border-radius: 4px; margin-bottom: 20px; }
|
58 |
+
h1, h2 { border-bottom: 1px solid #eaecef; padding-bottom: 0.3em; }
|
59 |
+
code { background-color: #eef; padding: 2px 4px; border-radius: 3px; }
|
60 |
</style>
|
61 |
</head>
|
62 |
<body>
|
63 |
<div class="container">
|
64 |
<h1>Private GitHub Reverse Proxy</h1>
|
65 |
+
<div class="security">
|
66 |
+
<strong>Security Notice:</strong> This is a private proxy service that requires authentication.
|
67 |
+
</div>
|
68 |
<div class="warning">
|
69 |
+
<strong>Warning:</strong> You are accessing GitHub content through a reverse proxy.
|
70 |
+
All content served from this page is provided directly by GitHub.
|
71 |
</div>
|
72 |
<h2>How to Use</h2>
|
73 |
+
<p>To access GitHub content, you need to include the secret key in the URL path:</p>
|
74 |
<p>For example, to clone a repository:</p>
|
75 |
+
<code>git clone {YOUR_PROXY_URL}/<SECRET_KEY>/https://github.com/owner/repo.git</code>
|
76 |
<p>Or to view a repository page:</p>
|
77 |
+
<code>{YOUR_PROXY_URL}/<SECRET_KEY>/https://github.com/owner/repo</code>
|
78 |
+
<h2>Authentication Required</h2>
|
79 |
+
<p>This proxy requires a valid secret key to access. Contact the administrator for access credentials.</p>
|
80 |
</div>
|
81 |
</body>
|
82 |
</html>
|
|
|
89 |
return True
|
90 |
return False
|
91 |
|
92 |
+
def extract_secret_and_url(path):
|
93 |
+
"""Extract secret key and target URL from the path."""
|
94 |
+
# Remove leading slash if present
|
95 |
+
if path.startswith('/'):
|
96 |
+
path = path[1:]
|
97 |
+
|
98 |
+
# Split the path to extract secret key and URL
|
99 |
+
parts = path.split('/', 1)
|
100 |
+
if len(parts) < 2:
|
101 |
+
return None, None
|
102 |
+
|
103 |
+
secret_key = parts[0]
|
104 |
+
target_url = parts[1]
|
105 |
+
|
106 |
+
return secret_key, target_url
|
107 |
+
|
108 |
+
# --- Core Proxy Logic ---
|
109 |
@app.route('/', defaults={'path': ''}, methods=['GET', 'POST', 'PUT', 'DELETE'])
|
110 |
@app.route('/<path:path>', methods=['GET', 'POST', 'PUT', 'DELETE'])
|
111 |
def proxy(path):
|
112 |
"""
|
113 |
+
Proxies requests to GitHub after validating secret key and URL whitelist.
|
114 |
+
It streams the response back to the client.
|
115 |
"""
|
116 |
+
# The full path might be URL-encoded, so we decode it.
|
117 |
+
target_path = unquote(request.full_path)
|
118 |
+
if target_path.startswith('/'):
|
119 |
+
target_path = target_path[1:]
|
120 |
+
|
121 |
+
# For the root path, display the custom warning page.
|
122 |
+
# Handle both '' (for a request to '/') and '?' (for a request to '/?').
|
123 |
+
if not target_path or target_path == '?':
|
|
|
|
|
|
|
|
|
|
|
124 |
return INDEX_PAGE_HTML, 200
|
125 |
|
126 |
+
# Extract secret key and target URL
|
127 |
+
provided_secret, target_url_path = extract_secret_and_url(target_path)
|
128 |
+
|
129 |
+
# Validate secret key
|
130 |
+
if not provided_secret or provided_secret != PROXY_SECRET_KEY:
|
131 |
+
logging.warning(f"Access denied: Invalid or missing secret key from IP: {request.remote_addr}")
|
132 |
+
return "<h1>403 Forbidden</h1><p>Access denied: Invalid or missing authentication key.</p>", 403
|
133 |
|
134 |
+
if not target_url_path:
|
135 |
+
logging.warning(f"Access denied: No target URL provided from IP: {request.remote_addr}")
|
136 |
+
return "<h1>400 Bad Request</h1><p>No target URL provided.</p>", 400
|
137 |
+
|
138 |
# Prepend 'https://' if the scheme is missing.
|
139 |
+
if not target_url_path.startswith(('http://', 'https://')):
|
140 |
+
target_url = 'https://' + target_url_path
|
141 |
else:
|
142 |
+
target_url = target_url_path
|
143 |
|
144 |
# Security check: Ensure the URL is in the whitelist.
|
145 |
if not is_url_allowed(target_url):
|
146 |
+
logging.warning(f"URL Denied! No pattern matched: {target_url} from IP: {request.remote_addr}")
|
147 |
return "<h1>403 Forbidden</h1><p>Request blocked by proxy security policy.</p>", 403
|
148 |
|
149 |
try:
|
|
|
159 |
headers['Host'] = target_host
|
160 |
|
161 |
try:
|
162 |
+
# Log successful proxy request
|
163 |
+
logging.info(f"Proxying request to: {target_url} from IP: {request.remote_addr}")
|
164 |
+
|
165 |
+
# Stream the request to handle large files efficiently.
|
166 |
resp = requests.request(
|
167 |
method=request.method,
|
168 |
url=target_url,
|
169 |
headers=headers,
|
170 |
data=request.get_data(),
|
171 |
cookies=request.cookies,
|
172 |
+
allow_redirects=False, # Redirects are handled by the client.
|
173 |
stream=True,
|
174 |
+
timeout=30 # Timeout for the connection.
|
175 |
)
|
176 |
+
|
177 |
+
# Exclude headers that can interfere with streaming.
|
178 |
excluded_headers = ['content-encoding', 'content-length', 'transfer-encoding', 'connection']
|
179 |
response_headers = [(name, value) for (name, value) in resp.raw.headers.items()
|
180 |
if name.lower() not in excluded_headers]
|
181 |
+
|
182 |
+
# Stream the response back to the original client.
|
183 |
return Response(resp.iter_content(chunk_size=8192), resp.status_code, response_headers)
|
184 |
|
185 |
except requests.exceptions.RequestException as e:
|
|
|
187 |
return "An error occurred while proxying the request.", 502
|
188 |
|
189 |
if __name__ == '__main__':
|
190 |
+
app.run(host='0.0.0.0', port=7860)
|