Spaces:
Sleeping
Sleeping
Commit
·
bb99b47
1
Parent(s):
36fd3e9
Uploaded
Browse files- Dockerfile +18 -32
- server.py +30 -19
- worker.py +16 -55
Dockerfile
CHANGED
|
@@ -5,57 +5,43 @@ WORKDIR /app
|
|
| 5 |
|
| 6 |
ENV DEBIAN_FRONTEND=noninteractive
|
| 7 |
ENV TZ=Etc/UTC
|
| 8 |
-
ENV HOME=/
|
| 9 |
-
ENV PYTHONUSERBASE=/
|
| 10 |
|
| 11 |
-
# HuggingFace Spaces
|
| 12 |
-
ENV CHROME_BIN=/usr/bin/chromium
|
| 13 |
-
ENV CHROMEDRIVER_PATH=/usr/bin/chromedriver
|
| 14 |
-
ENV DISPLAY=:99
|
| 15 |
-
|
| 16 |
-
# System deps + Chromium + ChromeDriver + essential libs for HF Spaces
|
| 17 |
RUN apt-get update && apt-get install -y --no-install-recommends \
|
| 18 |
-
wget \
|
| 19 |
-
gnupg \
|
| 20 |
-
unzip \
|
| 21 |
-
ca-certificates \
|
| 22 |
chromium \
|
| 23 |
chromium-driver \
|
| 24 |
-
|
| 25 |
-
|
| 26 |
libasound2 \
|
| 27 |
-
libxtst6 \
|
| 28 |
libatk-bridge2.0-0 \
|
|
|
|
| 29 |
libgtk-3-0 \
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
libgbm1 \
|
| 31 |
-
|
| 32 |
-
libappindicator3-1 \
|
| 33 |
-
libasound2 \
|
| 34 |
-
libdbus-glib-1-2 \
|
| 35 |
-
libdrm2 \
|
| 36 |
libxcomposite1 \
|
| 37 |
libxdamage1 \
|
| 38 |
-
libxrandr2 \
|
| 39 |
-
xvfb \
|
| 40 |
procps \
|
| 41 |
&& rm -rf /var/lib/apt/lists/*
|
| 42 |
|
| 43 |
-
# Create
|
| 44 |
-
RUN
|
| 45 |
-
|
| 46 |
-
chown -R appuser:appuser /app /home/appuser /tmp/chrome-temp && \
|
| 47 |
-
chmod 755 /tmp/chrome-temp
|
| 48 |
|
| 49 |
-
USER
|
| 50 |
|
| 51 |
-
COPY --chown=
|
| 52 |
RUN pip install --no-cache-dir --user -r requirements.txt
|
| 53 |
|
| 54 |
-
ENV PATH="/
|
| 55 |
|
| 56 |
-
COPY --chown=
|
| 57 |
|
| 58 |
-
# Create config directory if needed
|
| 59 |
RUN mkdir -p config
|
| 60 |
|
| 61 |
EXPOSE 7860
|
|
|
|
| 5 |
|
| 6 |
ENV DEBIAN_FRONTEND=noninteractive
|
| 7 |
ENV TZ=Etc/UTC
|
| 8 |
+
ENV HOME=/tmp/hf-user
|
| 9 |
+
ENV PYTHONUSERBASE=/tmp/hf-user/.local
|
| 10 |
|
| 11 |
+
# Install Chrome and dependencies for HuggingFace Spaces
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
RUN apt-get update && apt-get install -y --no-install-recommends \
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
chromium \
|
| 14 |
chromium-driver \
|
| 15 |
+
ca-certificates \
|
| 16 |
+
fonts-liberation \
|
| 17 |
libasound2 \
|
|
|
|
| 18 |
libatk-bridge2.0-0 \
|
| 19 |
+
libdrm2 \
|
| 20 |
libgtk-3-0 \
|
| 21 |
+
libnspr4 \
|
| 22 |
+
libnss3 \
|
| 23 |
+
libxss1 \
|
| 24 |
+
libxtst6 \
|
| 25 |
libgbm1 \
|
| 26 |
+
libxrandr2 \
|
|
|
|
|
|
|
|
|
|
|
|
|
| 27 |
libxcomposite1 \
|
| 28 |
libxdamage1 \
|
|
|
|
|
|
|
| 29 |
procps \
|
| 30 |
&& rm -rf /var/lib/apt/lists/*
|
| 31 |
|
| 32 |
+
# Create user directories
|
| 33 |
+
RUN mkdir -p /tmp/hf-user/.local && \
|
| 34 |
+
chmod 755 /tmp/hf-user /tmp/hf-user/.local
|
|
|
|
|
|
|
| 35 |
|
| 36 |
+
USER 1000
|
| 37 |
|
| 38 |
+
COPY --chown=1000:1000 requirements.txt .
|
| 39 |
RUN pip install --no-cache-dir --user -r requirements.txt
|
| 40 |
|
| 41 |
+
ENV PATH="/tmp/hf-user/.local/bin:${PATH}"
|
| 42 |
|
| 43 |
+
COPY --chown=1000:1000 . .
|
| 44 |
|
|
|
|
| 45 |
RUN mkdir -p config
|
| 46 |
|
| 47 |
EXPOSE 7860
|
server.py
CHANGED
|
@@ -24,13 +24,14 @@ load_dotenv()
|
|
| 24 |
|
| 25 |
app = Flask(__name__)
|
| 26 |
app.config['SECRET_KEY'] = 'secret-key-for-hillside-automation'
|
| 27 |
-
|
| 28 |
-
CORS
|
| 29 |
-
|
|
|
|
|
|
|
| 30 |
|
| 31 |
bot_instance = None
|
| 32 |
session_data = {}
|
| 33 |
-
automation_thread = None
|
| 34 |
|
| 35 |
class EmailService:
|
| 36 |
def __init__(self):
|
|
@@ -38,7 +39,8 @@ class EmailService:
|
|
| 38 |
self.password = os.getenv('EMAIL_PASSWORD')
|
| 39 |
self.smtp_server = "smtp.gmail.com"; self.smtp_port = 587
|
| 40 |
if not self.sender_email or not self.password:
|
| 41 |
-
print("[Email] WARNING: Email credentials not found in
|
|
|
|
| 42 |
def send_report(self, recipients, subject, body, attachments=None):
|
| 43 |
if not self.sender_email or not self.password: return False
|
| 44 |
try:
|
|
@@ -52,7 +54,8 @@ class EmailService:
|
|
| 52 |
with smtplib.SMTP(self.smtp_server, self.smtp_port) as server:
|
| 53 |
server.starttls(); server.login(self.sender_email, self.password); server.send_message(msg)
|
| 54 |
print(f"Email sent successfully to {', '.join(recipients)}"); return True
|
| 55 |
-
except Exception as e:
|
|
|
|
| 56 |
|
| 57 |
class GoogleDriveService:
|
| 58 |
def __init__(self):
|
|
@@ -62,12 +65,15 @@ class GoogleDriveService:
|
|
| 62 |
from google.oauth2 import service_account
|
| 63 |
from googleapiclient.discovery import build
|
| 64 |
base64_creds = os.getenv('GDRIVE_SA_KEY_BASE64')
|
| 65 |
-
if not base64_creds or not self.folder_id:
|
|
|
|
| 66 |
creds_json = base64.b64decode(base64_creds).decode('utf-8'); creds_dict = json.loads(creds_json)
|
| 67 |
self.creds = service_account.Credentials.from_service_account_info(creds_dict, scopes=['https://www.googleapis.com/auth/drive'])
|
| 68 |
self.service = build('drive', 'v3', credentials=self.creds)
|
| 69 |
-
print("[G-Drive] Service initialized securely from
|
| 70 |
-
except Exception as e:
|
|
|
|
|
|
|
| 71 |
def upload_file(self, filename, file_content):
|
| 72 |
if not self.service: return False
|
| 73 |
try:
|
|
@@ -77,15 +83,18 @@ class GoogleDriveService:
|
|
| 77 |
self.service.files().create(body=file_metadata, media_body=media, fields='id').execute()
|
| 78 |
print(f"[G-Drive] File '{filename}' uploaded successfully.")
|
| 79 |
return True
|
| 80 |
-
except Exception as e:
|
|
|
|
| 81 |
|
| 82 |
email_service = EmailService()
|
| 83 |
drive_service = GoogleDriveService()
|
| 84 |
|
| 85 |
def get_email_list():
|
| 86 |
try:
|
| 87 |
-
with open('config/emails.conf', 'r') as f:
|
| 88 |
-
|
|
|
|
|
|
|
| 89 |
|
| 90 |
def run_automation_process(session_id):
|
| 91 |
global bot_instance
|
|
@@ -94,7 +103,7 @@ def run_automation_process(session_id):
|
|
| 94 |
data = session_data.get(session_id, {}); csv_content = data.get('csv_content')
|
| 95 |
df = pd.read_csv(io.StringIO(csv_content));
|
| 96 |
if 'Status' not in df.columns: df.insert(1, 'Status', '')
|
| 97 |
-
patient_list = df[
|
| 98 |
socketio.emit('initial_stats', {'total': len(patient_list)})
|
| 99 |
results = bot_instance.process_patient_list(patient_list)
|
| 100 |
is_terminated = bot_instance.termination_event.is_set()
|
|
@@ -128,12 +137,13 @@ def generate_and_send_reports(session_id, results, is_crash_report=False, is_ter
|
|
| 128 |
|
| 129 |
@app.route('/')
|
| 130 |
def status_page():
|
| 131 |
-
APP_STATUS_HTML = """<!DOCTYPE html><html lang="en"><head><title>API
|
| 132 |
return Response(APP_STATUS_HTML)
|
| 133 |
|
| 134 |
@socketio.on('connect')
|
| 135 |
def handle_connect():
|
| 136 |
-
print(f'Frontend connected
|
|
|
|
| 137 |
|
| 138 |
@socketio.on('initialize_session')
|
| 139 |
def handle_init(data):
|
|
@@ -162,7 +172,8 @@ def handle_otp(data):
|
|
| 162 |
if is_success:
|
| 163 |
emit('login_successful')
|
| 164 |
session_id = 'user_session'
|
| 165 |
-
|
|
|
|
| 166 |
else: emit('error', {'message': f'OTP failed: {error_message}'})
|
| 167 |
|
| 168 |
@socketio.on('terminate_process')
|
|
@@ -171,8 +182,8 @@ def handle_terminate():
|
|
| 171 |
|
| 172 |
if __name__ == '__main__':
|
| 173 |
print("====================================================================")
|
| 174 |
-
print(" Hillside Automation Backend -
|
| 175 |
-
print("
|
| 176 |
-
print("
|
| 177 |
print("====================================================================")
|
| 178 |
socketio.run(app, host='0.0.0.0', port=int(os.getenv('PORT', 7860)))
|
|
|
|
| 24 |
|
| 25 |
app = Flask(__name__)
|
| 26 |
app.config['SECRET_KEY'] = 'secret-key-for-hillside-automation'
|
| 27 |
+
|
| 28 |
+
# Configure CORS for both local development and your Netlify frontend
|
| 29 |
+
FRONTEND_ORIGIN = os.getenv('FRONTEND_URL', 'https://quantbot.netlify.app')
|
| 30 |
+
CORS(app, resources={r"/*": {"origins": [FRONTEND_ORIGIN, "http://localhost:3000", "http://127.0.0.1:5500"]}})
|
| 31 |
+
socketio = SocketIO(app, cors_allowed_origins=[FRONTEND_ORIGIN, "http://localhost:3000", "http://127.0.0.1:5500"], async_mode='eventlet')
|
| 32 |
|
| 33 |
bot_instance = None
|
| 34 |
session_data = {}
|
|
|
|
| 35 |
|
| 36 |
class EmailService:
|
| 37 |
def __init__(self):
|
|
|
|
| 39 |
self.password = os.getenv('EMAIL_PASSWORD')
|
| 40 |
self.smtp_server = "smtp.gmail.com"; self.smtp_port = 587
|
| 41 |
if not self.sender_email or not self.password:
|
| 42 |
+
print("[Email] WARNING: Email credentials not found in environment variables.")
|
| 43 |
+
|
| 44 |
def send_report(self, recipients, subject, body, attachments=None):
|
| 45 |
if not self.sender_email or not self.password: return False
|
| 46 |
try:
|
|
|
|
| 54 |
with smtplib.SMTP(self.smtp_server, self.smtp_port) as server:
|
| 55 |
server.starttls(); server.login(self.sender_email, self.password); server.send_message(msg)
|
| 56 |
print(f"Email sent successfully to {', '.join(recipients)}"); return True
|
| 57 |
+
except Exception as e:
|
| 58 |
+
print(f"Failed to send email: {e}"); return False
|
| 59 |
|
| 60 |
class GoogleDriveService:
|
| 61 |
def __init__(self):
|
|
|
|
| 65 |
from google.oauth2 import service_account
|
| 66 |
from googleapiclient.discovery import build
|
| 67 |
base64_creds = os.getenv('GDRIVE_SA_KEY_BASE64')
|
| 68 |
+
if not base64_creds or not self.folder_id:
|
| 69 |
+
raise ValueError("Google Drive secrets not found.")
|
| 70 |
creds_json = base64.b64decode(base64_creds).decode('utf-8'); creds_dict = json.loads(creds_json)
|
| 71 |
self.creds = service_account.Credentials.from_service_account_info(creds_dict, scopes=['https://www.googleapis.com/auth/drive'])
|
| 72 |
self.service = build('drive', 'v3', credentials=self.creds)
|
| 73 |
+
print("[G-Drive] Service initialized securely from environment variables.")
|
| 74 |
+
except Exception as e:
|
| 75 |
+
print(f"[G-Drive] CRITICAL ERROR: Could not initialize Google Drive service: {e}")
|
| 76 |
+
|
| 77 |
def upload_file(self, filename, file_content):
|
| 78 |
if not self.service: return False
|
| 79 |
try:
|
|
|
|
| 83 |
self.service.files().create(body=file_metadata, media_body=media, fields='id').execute()
|
| 84 |
print(f"[G-Drive] File '{filename}' uploaded successfully.")
|
| 85 |
return True
|
| 86 |
+
except Exception as e:
|
| 87 |
+
print(f"[G-Drive] ERROR: File upload failed: {e}"); return False
|
| 88 |
|
| 89 |
email_service = EmailService()
|
| 90 |
drive_service = GoogleDriveService()
|
| 91 |
|
| 92 |
def get_email_list():
|
| 93 |
try:
|
| 94 |
+
with open('config/emails.conf', 'r') as f:
|
| 95 |
+
return [line.strip() for line in f if line.strip()]
|
| 96 |
+
except FileNotFoundError:
|
| 97 |
+
return []
|
| 98 |
|
| 99 |
def run_automation_process(session_id):
|
| 100 |
global bot_instance
|
|
|
|
| 103 |
data = session_data.get(session_id, {}); csv_content = data.get('csv_content')
|
| 104 |
df = pd.read_csv(io.StringIO(csv_content));
|
| 105 |
if 'Status' not in df.columns: df.insert(1, 'Status', '')
|
| 106 |
+
patient_list = df[(df['Status'] != 'Done') & (df['Status'] != 'Bad')]['Name'].tolist()
|
| 107 |
socketio.emit('initial_stats', {'total': len(patient_list)})
|
| 108 |
results = bot_instance.process_patient_list(patient_list)
|
| 109 |
is_terminated = bot_instance.termination_event.is_set()
|
|
|
|
| 137 |
|
| 138 |
@app.route('/')
|
| 139 |
def status_page():
|
| 140 |
+
APP_STATUS_HTML = """<!DOCTYPE html><html lang="en"><head><title>Hillside Automation API - HuggingFace</title><style>body{font-family:-apple-system,BlinkMacSystemFont,Segoe UI,Roboto,Helvetica,Arial,sans-serif;display:flex;justify-content:center;align-items:center;height:100vh;margin:0;background:#f0f2f5;}.status-box{text-align:center;padding:40px 60px;background:white;border-radius:12px;box-shadow:0 8px 30px rgba(0,0,0,0.1);}h1{font-size:24px;color:#333;margin-bottom:10px;} .indicator{font-size:18px;font-weight:600;padding:8px 16px;border-radius:20px;}.active{color:#28a745;background-color:#e9f7ea;}.info{margin-top:20px;color:#666;}</style></head><body><div class="status-box"><h1>🤗 Hillside Automation API</h1><div class="indicator active">● Running on HuggingFace Spaces</div><div class="info">Frontend: <a href="https://quantbot.netlify.app" target="_blank">quantbot.netlify.app</a></div></div></body></html>"""
|
| 141 |
return Response(APP_STATUS_HTML)
|
| 142 |
|
| 143 |
@socketio.on('connect')
|
| 144 |
def handle_connect():
|
| 145 |
+
print(f'Frontend connected from: {FRONTEND_ORIGIN}')
|
| 146 |
+
emit('email_list', {'emails': get_email_list()})
|
| 147 |
|
| 148 |
@socketio.on('initialize_session')
|
| 149 |
def handle_init(data):
|
|
|
|
| 172 |
if is_success:
|
| 173 |
emit('login_successful')
|
| 174 |
session_id = 'user_session'
|
| 175 |
+
# Fixed syntax for background task
|
| 176 |
+
socketio.start_background_task(run_automation_process, session_id)
|
| 177 |
else: emit('error', {'message': f'OTP failed: {error_message}'})
|
| 178 |
|
| 179 |
@socketio.on('terminate_process')
|
|
|
|
| 182 |
|
| 183 |
if __name__ == '__main__':
|
| 184 |
print("====================================================================")
|
| 185 |
+
print(" 🤗 Hillside Automation Backend - HuggingFace Spaces")
|
| 186 |
+
print(f" Frontend URL: {FRONTEND_ORIGIN}")
|
| 187 |
+
print(f" Port: {os.getenv('PORT', 7860)}")
|
| 188 |
print("====================================================================")
|
| 189 |
socketio.run(app, host='0.0.0.0', port=int(os.getenv('PORT', 7860)))
|
worker.py
CHANGED
|
@@ -1,9 +1,6 @@
|
|
| 1 |
# worker.py
|
| 2 |
import time
|
| 3 |
import threading
|
| 4 |
-
import tempfile
|
| 5 |
-
import shutil
|
| 6 |
-
import os
|
| 7 |
import subprocess
|
| 8 |
|
| 9 |
from selenium import webdriver
|
|
@@ -22,15 +19,14 @@ class QuantumBot:
|
|
| 22 |
self.driver = None
|
| 23 |
self.DEFAULT_TIMEOUT = 30
|
| 24 |
self.termination_event = threading.Event()
|
| 25 |
-
self.temp_user_dir = None
|
| 26 |
|
| 27 |
-
def
|
| 28 |
-
"""Kill any existing Chrome processes
|
| 29 |
try:
|
| 30 |
-
subprocess.run(['pkill', '-f', 'chrome'], capture_output=True, text=True)
|
| 31 |
-
subprocess.run(['pkill', '-f', 'chromium'], capture_output=True, text=True)
|
| 32 |
-
time.sleep(
|
| 33 |
-
except
|
| 34 |
pass
|
| 35 |
|
| 36 |
def initialize_driver(self):
|
|
@@ -38,53 +34,32 @@ class QuantumBot:
|
|
| 38 |
self.micro_status("Initializing headless browser...")
|
| 39 |
|
| 40 |
# Kill any existing Chrome processes first
|
| 41 |
-
self.
|
| 42 |
-
|
| 43 |
-
# Create a completely isolated temporary directory for this session
|
| 44 |
-
self.temp_user_dir = tempfile.mkdtemp(prefix="hf-chrome-", dir="/tmp")
|
| 45 |
-
os.chmod(self.temp_user_dir, 0o755)
|
| 46 |
|
| 47 |
options = ChromeOptions()
|
|
|
|
| 48 |
|
| 49 |
-
#
|
| 50 |
-
|
| 51 |
-
options.add_argument(f"--data-path={self.temp_user_dir}")
|
| 52 |
-
options.add_argument(f"--disk-cache-dir={self.temp_user_dir}/cache")
|
| 53 |
|
| 54 |
-
# HuggingFace Spaces
|
| 55 |
options.add_argument("--headless=new")
|
| 56 |
options.add_argument("--no-sandbox")
|
| 57 |
options.add_argument("--disable-dev-shm-usage")
|
| 58 |
options.add_argument("--disable-gpu")
|
| 59 |
-
options.add_argument("--disable-software-rasterizer")
|
| 60 |
-
options.add_argument("--disable-background-timer-throttling")
|
| 61 |
-
options.add_argument("--disable-backgrounding-occluded-windows")
|
| 62 |
-
options.add_argument("--disable-renderer-backgrounding")
|
| 63 |
-
options.add_argument("--disable-features=TranslateUI,BlinkGenPropertyTrees")
|
| 64 |
options.add_argument("--remote-debugging-port=0")
|
|
|
|
| 65 |
options.add_argument("--no-first-run")
|
| 66 |
-
options.add_argument("--no-default-browser-check")
|
| 67 |
options.add_argument("--disable-extensions")
|
| 68 |
-
options.add_argument("--disable-plugins")
|
| 69 |
-
options.add_argument("--disable-images")
|
| 70 |
-
options.add_argument("--disable-javascript")
|
| 71 |
-
options.add_argument("--window-size=1920,1080")
|
| 72 |
|
| 73 |
-
#
|
| 74 |
user_agent = "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36"
|
| 75 |
options.add_argument(f"--user-agent={user_agent}")
|
| 76 |
|
| 77 |
-
# Set binary location for HF Spaces
|
| 78 |
-
options.binary_location = "/usr/bin/chromium"
|
| 79 |
-
|
| 80 |
service = ChromeService(executable_path="/usr/bin/chromedriver")
|
| 81 |
-
|
| 82 |
-
# Set service args for additional stability
|
| 83 |
-
service.service_args = ["--verbose", "--whitelisted-ips="]
|
| 84 |
-
|
| 85 |
self.driver = webdriver.Chrome(service=service, options=options)
|
| 86 |
|
| 87 |
-
#
|
| 88 |
self.driver.execute_cdp_cmd('Page.addScriptToEvaluateOnNewDocument', {
|
| 89 |
'source': "Object.defineProperty(navigator, 'webdriver', {get: () => undefined})"
|
| 90 |
})
|
|
@@ -94,11 +69,6 @@ class QuantumBot:
|
|
| 94 |
except Exception as e:
|
| 95 |
error_message = f"Message: {str(e)}"
|
| 96 |
print(f"CRITICAL ERROR in WebDriver Initialization: {error_message}")
|
| 97 |
-
|
| 98 |
-
# Cleanup on failure
|
| 99 |
-
if self.temp_user_dir and os.path.exists(self.temp_user_dir):
|
| 100 |
-
shutil.rmtree(self.temp_user_dir, ignore_errors=True)
|
| 101 |
-
|
| 102 |
return False, error_message
|
| 103 |
|
| 104 |
def micro_status(self, message):
|
|
@@ -244,16 +214,7 @@ class QuantumBot:
|
|
| 244 |
try:
|
| 245 |
if self.driver:
|
| 246 |
self.driver.quit()
|
| 247 |
-
|
| 248 |
-
|
| 249 |
-
self._kill_existing_chrome_processes()
|
| 250 |
-
|
| 251 |
-
# Clean up temporary directory
|
| 252 |
-
if self.temp_user_dir and os.path.exists(self.temp_user_dir):
|
| 253 |
-
shutil.rmtree(self.temp_user_dir, ignore_errors=True)
|
| 254 |
-
print(f"[Bot] Cleaned up temporary Chrome profile: {self.temp_user_dir}")
|
| 255 |
-
|
| 256 |
except Exception as e:
|
| 257 |
print(f"[Bot] Error during shutdown: {e}")
|
| 258 |
-
|
| 259 |
-
print("[Bot] Chrome session closed.")
|
|
|
|
| 1 |
# worker.py
|
| 2 |
import time
|
| 3 |
import threading
|
|
|
|
|
|
|
|
|
|
| 4 |
import subprocess
|
| 5 |
|
| 6 |
from selenium import webdriver
|
|
|
|
| 19 |
self.driver = None
|
| 20 |
self.DEFAULT_TIMEOUT = 30
|
| 21 |
self.termination_event = threading.Event()
|
|
|
|
| 22 |
|
| 23 |
+
def _kill_chrome_processes(self):
|
| 24 |
+
"""Kill any existing Chrome processes to prevent conflicts"""
|
| 25 |
try:
|
| 26 |
+
subprocess.run(['pkill', '-f', 'chrome'], capture_output=True, text=True, timeout=5)
|
| 27 |
+
subprocess.run(['pkill', '-f', 'chromium'], capture_output=True, text=True, timeout=5)
|
| 28 |
+
time.sleep(1)
|
| 29 |
+
except:
|
| 30 |
pass
|
| 31 |
|
| 32 |
def initialize_driver(self):
|
|
|
|
| 34 |
self.micro_status("Initializing headless browser...")
|
| 35 |
|
| 36 |
# Kill any existing Chrome processes first
|
| 37 |
+
self._kill_chrome_processes()
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
|
| 39 |
options = ChromeOptions()
|
| 40 |
+
options.binary_location = "/usr/bin/chromium"
|
| 41 |
|
| 42 |
+
# DON'T set --user-data-dir - let Chrome handle it automatically
|
| 43 |
+
# This prevents the "already in use" error
|
|
|
|
|
|
|
| 44 |
|
| 45 |
+
# Minimal required flags for HuggingFace Spaces
|
| 46 |
options.add_argument("--headless=new")
|
| 47 |
options.add_argument("--no-sandbox")
|
| 48 |
options.add_argument("--disable-dev-shm-usage")
|
| 49 |
options.add_argument("--disable-gpu")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 50 |
options.add_argument("--remote-debugging-port=0")
|
| 51 |
+
options.add_argument("--window-size=1920,1080")
|
| 52 |
options.add_argument("--no-first-run")
|
|
|
|
| 53 |
options.add_argument("--disable-extensions")
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
|
| 55 |
+
# User agent for better compatibility
|
| 56 |
user_agent = "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36"
|
| 57 |
options.add_argument(f"--user-agent={user_agent}")
|
| 58 |
|
|
|
|
|
|
|
|
|
|
| 59 |
service = ChromeService(executable_path="/usr/bin/chromedriver")
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
self.driver = webdriver.Chrome(service=service, options=options)
|
| 61 |
|
| 62 |
+
# Anti-detection
|
| 63 |
self.driver.execute_cdp_cmd('Page.addScriptToEvaluateOnNewDocument', {
|
| 64 |
'source': "Object.defineProperty(navigator, 'webdriver', {get: () => undefined})"
|
| 65 |
})
|
|
|
|
| 69 |
except Exception as e:
|
| 70 |
error_message = f"Message: {str(e)}"
|
| 71 |
print(f"CRITICAL ERROR in WebDriver Initialization: {error_message}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 72 |
return False, error_message
|
| 73 |
|
| 74 |
def micro_status(self, message):
|
|
|
|
| 214 |
try:
|
| 215 |
if self.driver:
|
| 216 |
self.driver.quit()
|
| 217 |
+
self._kill_chrome_processes()
|
| 218 |
+
print("[Bot] Chrome session closed and cleaned up.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 219 |
except Exception as e:
|
| 220 |
print(f"[Bot] Error during shutdown: {e}")
|
|
|
|
|
|