Spaces:
Running
Running
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,183 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""
|
| 2 |
+
CroxyProxy Rotating Proxy API - HuggingFace Spaces
|
| 3 |
+
90 servers: Paris, LA, Dallas, Warsaw, Amsterdam...
|
| 4 |
+
|
| 5 |
+
GET /health - Status
|
| 6 |
+
GET /servers - List 90 servers
|
| 7 |
+
POST /proxy/fetch - Rotating proxy
|
| 8 |
+
POST /proxy/random - Random server
|
| 9 |
+
POST /proxy/batch - Multiple URLs
|
| 10 |
+
"""
|
| 11 |
+
|
| 12 |
+
import json, base64, re, random, time, threading
|
| 13 |
+
from datetime import datetime, timezone
|
| 14 |
+
from flask import Flask, request, jsonify
|
| 15 |
+
from bs4 import BeautifulSoup
|
| 16 |
+
import cloudscraper
|
| 17 |
+
from html import unescape
|
| 18 |
+
import warnings
|
| 19 |
+
warnings.filterwarnings("ignore")
|
| 20 |
+
|
| 21 |
+
BASE = "https://www.croxyproxy.com"
|
| 22 |
+
app = Flask(__name__)
|
| 23 |
+
|
| 24 |
+
class S:
|
| 25 |
+
servers = []
|
| 26 |
+
idx = 0
|
| 27 |
+
lock = threading.Lock()
|
| 28 |
+
last = None
|
| 29 |
+
stats = {"req": 0, "ok": 0, "fail": 0}
|
| 30 |
+
|
| 31 |
+
def dec(e):
|
| 32 |
+
try:
|
| 33 |
+
return json.loads(bytes.fromhex(base64.b64decode(e).decode()).decode())
|
| 34 |
+
except:
|
| 35 |
+
return None
|
| 36 |
+
|
| 37 |
+
def fetch(url, sid=None):
|
| 38 |
+
sc = cloudscraper.create_scraper(browser={"browser":"chrome","platform":"windows","desktop":True})
|
| 39 |
+
S.stats["req"] += 1
|
| 40 |
+
try:
|
| 41 |
+
# 1. GET / → csrf
|
| 42 |
+
r1 = sc.get(BASE, timeout=30)
|
| 43 |
+
if r1.status_code != 200:
|
| 44 |
+
S.stats["fail"] += 1
|
| 45 |
+
return {"success": False, "error": f"Homepage {r1.status_code}"}
|
| 46 |
+
|
| 47 |
+
s1 = BeautifulSoup(r1.text, "lxml")
|
| 48 |
+
ci = s1.find("input", {"name": "csrf"})
|
| 49 |
+
if not ci:
|
| 50 |
+
S.stats["fail"] += 1
|
| 51 |
+
return {"success": False, "error": "No CSRF"}
|
| 52 |
+
|
| 53 |
+
# 2. POST /servers → selector page
|
| 54 |
+
r2 = sc.post(f"{BASE}/servers", data={
|
| 55 |
+
"url": url, "proxyServerId": "274", "csrf": ci["value"],
|
| 56 |
+
"demo": "0", "frontOrigin": BASE,
|
| 57 |
+
}, headers={
|
| 58 |
+
"Content-Type": "application/x-www-form-urlencoded",
|
| 59 |
+
"Origin": BASE, "Referer": BASE + "/",
|
| 60 |
+
}, allow_redirects=True, timeout=30)
|
| 61 |
+
|
| 62 |
+
if r2.status_code != 200:
|
| 63 |
+
S.stats["fail"] += 1
|
| 64 |
+
return {"success": False, "error": f"Servers {r2.status_code}"}
|
| 65 |
+
|
| 66 |
+
s2 = BeautifulSoup(r2.text, "lxml")
|
| 67 |
+
sel = s2.find("script", {"id": "serverSelectorScript"})
|
| 68 |
+
if not sel:
|
| 69 |
+
S.stats["fail"] += 1
|
| 70 |
+
return {"success": False, "error": "No selector"}
|
| 71 |
+
|
| 72 |
+
# 3. Parse servers + csrf2
|
| 73 |
+
ss = [x for x in (dec(i) for i in json.loads(unescape(sel.get("data-ss","")))) if x and x.get("id")]
|
| 74 |
+
csrf2 = unescape(sel.get("data-csrf", "")).strip('"')
|
| 75 |
+
fo = unescape(sel.get("data-fo", "")).strip('"')
|
| 76 |
+
|
| 77 |
+
if not ss:
|
| 78 |
+
S.stats["fail"] += 1
|
| 79 |
+
return {"success": False, "error": "No servers"}
|
| 80 |
+
|
| 81 |
+
S.servers = ss
|
| 82 |
+
S.last = datetime.now(timezone.utc).isoformat()
|
| 83 |
+
|
| 84 |
+
# Choose server
|
| 85 |
+
ch = None
|
| 86 |
+
if sid:
|
| 87 |
+
ch = next((x for x in ss if x["id"] == sid), None)
|
| 88 |
+
if not ch:
|
| 89 |
+
with S.lock:
|
| 90 |
+
ch = ss[S.idx % len(ss)]
|
| 91 |
+
S.idx += 1
|
| 92 |
+
|
| 93 |
+
# 4. POST /requests?fso= → 302
|
| 94 |
+
r3 = sc.post(f"{BASE}/requests?fso=", data={
|
| 95 |
+
"url": url, "proxyServerId": str(ch["id"]),
|
| 96 |
+
"csrf": csrf2, "demo": "0", "frontOrigin": fo,
|
| 97 |
+
}, headers={
|
| 98 |
+
"Content-Type": "application/x-www-form-urlencoded",
|
| 99 |
+
"Origin": BASE, "Referer": f"{BASE}/servers",
|
| 100 |
+
}, allow_redirects=False, timeout=30)
|
| 101 |
+
|
| 102 |
+
loc = r3.headers.get("Location") or r3.headers.get("location")
|
| 103 |
+
if not loc:
|
| 104 |
+
S.stats["fail"] += 1
|
| 105 |
+
return {"success": False, "error": f"No redirect ({r3.status_code})", "server": ch.get("name")}
|
| 106 |
+
|
| 107 |
+
# 5. GET redirect → launching page → data-r
|
| 108 |
+
r4 = sc.get(loc, timeout=30, allow_redirects=True)
|
| 109 |
+
dr = re.search(r'data-r="([^"]+)"', r4.text)
|
| 110 |
+
if not dr:
|
| 111 |
+
S.stats["fail"] += 1
|
| 112 |
+
return {"success": False, "error": "No data-r", "server": ch.get("name")}
|
| 113 |
+
|
| 114 |
+
# 6. decode data-r → GET final
|
| 115 |
+
final = base64.b64decode(dr.group(1)).decode()
|
| 116 |
+
r5 = sc.get(final, timeout=30, allow_redirects=True)
|
| 117 |
+
|
| 118 |
+
S.stats["ok"] += 1
|
| 119 |
+
return {
|
| 120 |
+
"success": True, "status": r5.status_code,
|
| 121 |
+
"headers": dict(r5.headers), "body": r5.text, "url": url,
|
| 122 |
+
"proxy": {
|
| 123 |
+
"server_id": ch["id"], "server_name": ch.get("name"),
|
| 124 |
+
"ip": ch.get("url","").replace("https://","").replace("/__cp2.php",""),
|
| 125 |
+
},
|
| 126 |
+
"servers_available": len(ss),
|
| 127 |
+
}
|
| 128 |
+
except Exception as e:
|
| 129 |
+
S.stats["fail"] += 1
|
| 130 |
+
return {"success": False, "error": str(e)}
|
| 131 |
+
|
| 132 |
+
|
| 133 |
+
@app.route("/")
|
| 134 |
+
def index():
|
| 135 |
+
return jsonify({
|
| 136 |
+
"name": "CroxyProxy Rotating Proxy API",
|
| 137 |
+
"endpoints": {
|
| 138 |
+
"GET /health": "Status + stats",
|
| 139 |
+
"GET /servers": "List all servers",
|
| 140 |
+
"POST /proxy/fetch": "Rotating proxy {url, server_id?}",
|
| 141 |
+
"POST /proxy/random": "Random server {url}",
|
| 142 |
+
"POST /proxy/batch": "Multiple URLs {urls: [...]}",
|
| 143 |
+
},
|
| 144 |
+
"example": 'curl -X POST /proxy/fetch -H "Content-Type: application/json" -d \'{"url":"https://httpbin.org/ip"}\'',
|
| 145 |
+
})
|
| 146 |
+
|
| 147 |
+
@app.route("/health")
|
| 148 |
+
def health():
|
| 149 |
+
return jsonify({"status": "ready", "servers": len(S.servers), "last_refresh": S.last, "stats": S.stats})
|
| 150 |
+
|
| 151 |
+
@app.route("/servers")
|
| 152 |
+
def servers():
|
| 153 |
+
return jsonify({"count": len(S.servers), "servers": [
|
| 154 |
+
{"id": s.get("id"), "name": s.get("name"), "ip": s.get("url","").replace("https://","").replace("/__cp2.php","")}
|
| 155 |
+
for s in S.servers
|
| 156 |
+
]})
|
| 157 |
+
|
| 158 |
+
@app.route("/proxy/fetch", methods=["POST"])
|
| 159 |
+
def pf():
|
| 160 |
+
d = request.get_json() or {}
|
| 161 |
+
if not d.get("url"): return jsonify({"error": "url required"}), 400
|
| 162 |
+
return jsonify(fetch(d["url"], d.get("server_id")))
|
| 163 |
+
|
| 164 |
+
@app.route("/proxy/random", methods=["POST"])
|
| 165 |
+
def pr():
|
| 166 |
+
d = request.get_json() or {}
|
| 167 |
+
if not d.get("url"): return jsonify({"error": "url required"}), 400
|
| 168 |
+
sid = random.choice(S.servers)["id"] if S.servers else None
|
| 169 |
+
return jsonify(fetch(d["url"], sid))
|
| 170 |
+
|
| 171 |
+
@app.route("/proxy/batch", methods=["POST"])
|
| 172 |
+
def pb():
|
| 173 |
+
d = request.get_json() or {}
|
| 174 |
+
urls = d.get("urls", [])
|
| 175 |
+
if not urls: return jsonify({"error": "urls required"}), 400
|
| 176 |
+
res = []
|
| 177 |
+
for u in urls:
|
| 178 |
+
res.append(fetch(u))
|
| 179 |
+
time.sleep(0.5)
|
| 180 |
+
return jsonify({"results": res})
|
| 181 |
+
|
| 182 |
+
if __name__ == "__main__":
|
| 183 |
+
app.run(host="0.0.0.0", port=7860)
|