upload code

.gitignore

@@ -0,0 +1,2 @@
+*.pyc
+cookies.txt

README.md

@@ -1,3 +1,18 @@
 # huggingface-proxy
 
-Host a proxy server on Huggingface Spaces
+_Host a proxy server on Huggingface Spaces._
+
+### Setup
+#### Space
+1. Create a new Space using `./space/Dockerfile`.
+2. Upload your rathole client config to the space under the path `./rathole-client.toml`
+3. Upload `./space/server.py` and `./space/supervisord.conf`
+4. "Factory Reboot" the space to get it to rebuild.
+
+
+#### Keep Alive
+On a local machine, run the keep alive script to prevent Huggingface from stopping your space due to "inactivity". You'll have to create a write API key on your account and save your cookies to `cookies.txt`.
+
+```
+python3 ./keepalive/keepalive.py --token hf_XXX --username XXX --cookies ./cookies.txt --headless
+```

keepalive/hf-upkeep.service

@@ -0,0 +1,15 @@
+[Unit]
+Description=Huggingface Proxy Upkeep Daemon
+After=network.target
+StartLimitIntervalSec=0
+
+[Service]
+User=server
+Group=server
+Type=simple
+ExecStart=/srv/server/huggingface-proxy/venv/bin/python3 /srv/server/huggingface-proxy/keepalive.py --token hf_XXX --username example --cookies /srv/server/huggingface-proxy/cookies.txt --headless
+Restart=always
+RestartSec=3
+
+[Install]
+WantedBy=multi-user.target

keepalive/keepalive.py

@@ -0,0 +1,81 @@
+import argparse
+import random
+import time
+from pathlib import Path
+
+from huggingface_hub import HfApi
+from selenium.webdriver.common.by import By
+from selenium.webdriver.support import expected_conditions as EC
+from selenium.webdriver.support.ui import WebDriverWait
+
+from lib.hf import list_spaces, login
+from lib.webdriver import create_driver, load_cookies, save_cookies
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--cookies", required=True, help="Path to the file containing your HF cookies")
+parser.add_argument("--token", required=True, help="Your HF API token")
+parser.add_argument("--username", required=True, help="Your HF username")
+parser.add_argument("--password", required=True, help="Your HF password")
+parser.add_argument("--headless", action='store_true', help="Run Selinium in headless mode")
+args = parser.parse_args()
+
+cookies_file = Path(args.cookies).resolve().expanduser().absolute()
+
+login(args.token)
+
+driver = create_driver(args.headless)
+
+driver.get(f'https://huggingface.co')
+if cookies_file.exists():
+    load_cookies(cookies_file)
+
+driver.get('https://huggingface.co/settings/profile')
+if driver.current_url != 'https://huggingface.co/settings/profile':
+    print('Not logged in!')
+    username = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.NAME, "username")))
+    username.send_keys(args.username)
+    password = driver.find_element(By.NAME, "password")
+    password.send_keys(args.password)
+    login_button = WebDriverWait(driver, 10).until(EC.presence_of_element_located((By.XPATH, '//button[@type="submit"]')))
+    login_button.click()
+    save_cookies(driver, cookies_file)
+    print('Manually logged in.')
+else:
+    print('Auth is correct')
+
+while True:
+    our_spaces = [x.id for x in list_spaces(args.username) if 'proxy' in x.id]
+    random.shuffle(our_spaces)
+
+    driver.get(f'https://huggingface.co')
+    time.sleep(5)
+    for space in our_spaces:
+        print(space)
+        driver.get(f'https://huggingface.co/spaces/{space}')
+        time.sleep(100)
+
+        # try:
+        if len(driver.find_elements(By.ID, 'captcha-container')):
+            print('CAPTCHA!!!')
+        else:
+            for iframe in driver.find_elements(By.TAG_NAME, 'iframe'):
+                src = iframe.get_attribute('src')
+                if 'hf.space' in src and '?__sign' in src:
+                    print('Space is online')
+                    break
+            else:
+                api = HfApi()
+                try:
+                    api.restart_space(repo_id=space)
+                    print('Restarted space')
+                except Exception as e:
+                    print('Failed to restart space!', e)
+
+        time.sleep(30)
+
+        # Avoid the captcha
+        driver.get('https://huggingface.co')
+        time.sleep(30)
+    sleep_time = 300
+    print(f'Sleeping {sleep_time} seconds...')
+    time.sleep(sleep_time)

keepalive/lib/hf.py

@@ -0,0 +1,14 @@
+from huggingface_hub import login as hf_login
+from huggingface_hub import HfApi
+
+
+def login(token: str):
+    return hf_login(token)
+
+
+def list_spaces(author: str):
+    api = HfApi()
+    result = []
+    for space in api.list_spaces(author=author):
+        result.append(space)
+    return result

keepalive/lib/webdriver.py

@@ -0,0 +1,46 @@
+import http.cookiejar
+
+import undetected_chromedriver as uc
+from selenium.common import InvalidSessionIdException
+from selenium.webdriver.chrome.options import Options
+from selenium.webdriver.chrome.webdriver import WebDriver
+from webdriver_manager.chrome import ChromeDriverManager
+
+driver: WebDriver = None
+
+
+def create_driver(headless=False):
+    global driver
+    if driver:
+        try:
+            driver.close()
+        except InvalidSessionIdException:
+            pass
+
+    options = Options()
+    if headless:
+        options.add_argument('--headless')
+        options.add_argument("--window-size=1920,1200")
+    options.add_argument("--incognito")
+    driver = uc.Chrome(driver_executable_path=ChromeDriverManager().install(), options=options, use_subprocess=False)
+    return driver
+
+
+def load_cookies(file):
+    cookie_jar = http.cookiejar.MozillaCookieJar()
+    cookie_jar.load(file, ignore_discard=True, ignore_expires=True)
+    for cookie in cookie_jar:
+        driver.add_cookie({"name": cookie.name, "value": cookie.value, 'domain': cookie.domain})
+
+
+def save_cookies(driver, location):
+    with open(location, 'w') as file:
+        file.write('# Netscape HTTP Cookie File\n')
+        for cookie in driver.get_cookies():
+            file.write(f"{cookie['domain']}\t"
+                       f"{'TRUE' if cookie['domain'].startswith('.') else 'FALSE'}\t"
+                       f"{cookie['path']}\t"
+                       f"{'TRUE' if 'secure' in cookie else 'FALSE'}\t"
+                       f"{str(int(cookie['expiry'])) if 'expiry' in cookie else '0'}\t"
+                       f"{cookie['name']}\t"
+                       f"{cookie['value']}\n")

keepalive/requirements.txt

@@ -0,0 +1,5 @@
+undetected-chromedriver
+selenium==4.9.1
+huggingface_hub
+webdriver_manager
+pyvirtualdisplay

keepalive/restart-spaces.py

@@ -0,0 +1,24 @@
+import argparse
+import random
+
+from huggingface_hub import HfApi
+
+from lib.hf import list_spaces, login
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--token", required=True, help="Your HF API token")
+parser.add_argument("--username", required=True, help="Your HF username")
+args = parser.parse_args()
+
+login(args.token)
+
+our_spaces = [x.id for x in list_spaces(args.username) if 'proxy' in x.id]
+random.shuffle(our_spaces)
+
+for space in our_spaces:
+    api = HfApi()
+    try:
+        api.restart_space(repo_id=space)
+        print(space)
+    except Exception as e:
+        print('Failed to restart space!', e)

space/Dockerfile

@@ -0,0 +1,25 @@
+FROM debian:stable
+RUN apt-get update && \
+    apt-get install -y git wget python3 python3-pip unzip && \
+    rm -rf /var/cache/apt/archives /var/lib/apt/lists/*
+
+RUN mkdir -p /var/log/app
+
+RUN git clone "https://git.evulid.cc/cyberes/huggingface-proxy.git" /app/huggingface
+
+COPY ./rathole-client.toml /app/huggingface/rathole-client.toml
+COPY ./server.py /app/huggingface/space/server.py
+
+RUN wget https://github.com/rapiz1/rathole/releases/download/v0.5.0/rathole-x86_64-unknown-linux-gnu.zip -O /tmp/rathole.zip
+RUN unzip /tmp/rathole.zip -d /app
+
+RUN pip install --break-system-packages supervisor requests flask proxy.py "httpx[http2]"
+
+RUN adduser --disabled-password --gecos '' --uid 1000 newuser
+
+RUN chown -R newuser:newuser /app
+
+EXPOSE 7860
+USER newuser
+
+CMD cd /app && /usr/local/bin/supervisord -c /app/huggingface/space/supervisord.conf

space/server.py

@@ -0,0 +1,14 @@
+import requests
+from flask import Flask, Response
+
+app = Flask(__name__)
+
+
+@app.route('/')
+def home():
+    r = requests.get('https://api.ipify.org')
+    return Response(r.text, mimetype='text/plain')
+
+
+if __name__ == '__main__':
+    app.run(host="0.0.0.0", port=7860)

space/supervisord.conf

@@ -0,0 +1,30 @@
+[supervisord]
+nodaemon=true
+
+[program:rathole]
+command=/bin/bash -c '/app/rathole -c /app/huggingface/rathole-client.toml 2>&1 | tee -a /app/rathole.log'
+autostart=true
+autorestart=true
+stdout_logfile=/dev/fd/1
+stdout_logfile_maxbytes=0
+stderr_logfile=/dev/fd/2
+stderr_logfile_maxbytes=0
+
+[program:flask]
+command=/usr/bin/python3 /app/huggingface/space/server.py
+autostart=true
+autorestart=true
+stdout_logfile=/dev/fd/1
+stdout_logfile_maxbytes=0
+stderr_logfile=/dev/fd/2
+stderr_logfile_maxbytes=0
+
+[program:proxy]
+command=proxy --hostname 0.0.0.0 --port 3128 --timeout 300 --plugins proxy.plugin.CloudflareDnsResolverPlugin
+autostart=true
+autorestart=true
+stdout_logfile=/dev/fd/1
+stdout_logfile_maxbytes=0
+stderr_logfile=/dev/fd/2
+stderr_logfile_maxbytes=0
+environment=PORT=3128