arena-annotation-progress / count_ip_data.py
Terry Zhuo
update
6da81f8
raw
history blame
11.9 kB
import os
import logging
from datetime import datetime, timedelta
from urllib.parse import unquote
import json
from collections import defaultdict
import smbclient
import shutil
import re
import argparse
# List of IP addresses we care about
WHITELIST_IPS = [
"199.111.212.5",
"175.159.122.63",
"109.245.193.97",
"158.195.18.232",
"2607:fea8:4f40:4b00:e5b9:9806:6b69:233b",
"66.254.231.49",
"129.74.154.194",
"175.196.44.217",
"2601:600:8d00:9510:1d77:b610:9358:f443",
"74.90.222.68",
"2a02:169:3e9:0:6ce8:e76f:faed:c830",
"70.50.179.57",
"2a02:842a:24:5a01:8cd6:5b22:1189:6035",
"2408:8418:6390:7603:40b:555f:774:a05d"
]
logging.basicConfig(level=logging.WARNING)
log = logging.getLogger(__name__)
def get_ip_from_jsonl(file_path):
"""Extract IP from the first line of a JSONL file"""
try:
with smbclient.open_file(file_path, mode='r') as f:
first_line = f.readline()
data = json.loads(first_line)
return data.get('ip')
except Exception as e:
log.error(f"Error reading file {file_path}: {e}")
return None
def get_chat_session_id(file_path):
"""Extract chat_session_id based on the file location:
- For files under conv_logs: extract from filename
- For files under sandbox_logs: read from file content
"""
try:
if 'conv_logs' in file_path:
# Extract from filename for conv_logs
# Handle Windows UNC path format
filename = file_path.split('\\')[-1] # Get the last component of the path
match = re.match(r'conv-log-([a-f0-9]+)\.json', filename)
if match:
return match.group(1)
elif 'sandbox_logs' in file_path:
# Read from file content for sandbox_logs
with smbclient.open_file(file_path, mode='r') as f:
data = json.loads(f.read())
return data['sandbox_state'].get('chat_session_id')
return None
except Exception as e:
log.error(f"Error getting chat_session_id from {file_path}: {e}")
return None
def get_sandbox_session_ids(server, share, date_str):
"""Get all chat_session_ids from sandbox logs for a given date"""
sandbox_folder = f"\\\\{server}\\{share}\\{date_str}\\sandbox_logs"
session_ids = set()
if not smbclient.path.exists(sandbox_folder):
return session_ids
try:
for file_info in smbclient.scandir(sandbox_folder):
if file_info.name.endswith('.json'):
file_path = f"{sandbox_folder}\\{file_info.name}"
session_id = get_chat_session_id(file_path)
if session_id:
session_ids.add(session_id)
except Exception as e:
log.error(f"Error scanning sandbox folder {sandbox_folder}: {e}")
return session_ids
def check_vote_conditions(file_path):
"""Check if the last line of the file has type:vote and feedback dict with 6 keys"""
try:
with smbclient.open_file(file_path, mode='r') as f:
# Read all lines and get the last non-empty line
lines = [line.strip() for line in f if line.strip()]
if not lines:
return False
last_line = lines[-1]
try:
data = json.loads(last_line)
feedback = data.get('feedback')
return (data.get('type') == 'vote' and
isinstance(feedback, dict) and
len(feedback) == 6)
except json.JSONDecodeError:
return False
except Exception as e:
log.error(f"Error checking vote conditions in file {file_path}: {e}")
return False
def get_file_data(file_path):
"""Read file and return IP and vote condition status"""
try:
with smbclient.open_file(file_path, mode='r') as f:
lines = [line.strip() for line in f if line.strip()]
if not lines:
return None, False
# Get IP from first line
try:
first_line_data = json.loads(lines[0])
ip = first_line_data.get('ip')
# Early return if IP is not in whitelist
if ip not in WHITELIST_IPS:
return None, False
except json.JSONDecodeError:
ip = None
# Check vote conditions from last line
try:
last_line_data = json.loads(lines[-1])
feedback = last_line_data.get('feedback')
vote_conditions_met = (last_line_data.get('type') == 'vote' and
isinstance(feedback, dict) and
len(feedback) == 6)
except json.JSONDecodeError:
vote_conditions_met = False
return ip, vote_conditions_met
except Exception as e:
log.error(f"Error reading file {file_path}: {e}")
return None, False
def count_files_per_ip(smb_url, start_date_str="2025_02_18"):
"""Count files per IP address from the given start date"""
# Remove 'smb://' prefix and parse URL components
url = smb_url[6:]
creds_server, share = url.split('/', 1)
creds, server = creds_server.rsplit('@', 1)
username, password = creds.split(':', 1)
password = unquote(password)
# Register the SMB session
smbclient.register_session(server, username=username, password=password, port=8080)
# Convert start date string to datetime
start_date = datetime.strptime(start_date_str, "%Y_%m_%d")
ip_counts = defaultdict(int)
try:
# Get current date for iteration
current_date = start_date
today = datetime.now()
while current_date <= today:
date_str = current_date.strftime("%Y_%m_%d")
folder_path = f"\\\\{server}\\{share}\\{date_str}\\conv_logs\\battle_anony"
try:
# List all JSON files in the battle_anony folder
if smbclient.path.exists(folder_path):
for file_info in smbclient.scandir(folder_path, search_pattern="conv-log-*.json"):
file_path = f"{folder_path}\\{file_info.name}"
ip, vote_conditions_met = get_file_data(file_path)
if vote_conditions_met and ip:
ip_counts[ip] += 1
except Exception as e:
log.error(f"Error processing folder {date_str}: {e}")
# Move to next day
current_date += timedelta(days=1)
except Exception as e:
log.error(f"Error accessing SMB share: {e}")
return dict(ip_counts)
def download_files_by_ip(smb_url, start_date_str="2025_02_18", check_sandbox=True):
"""Download files and organize them by IP address
Args:
smb_url (str): The SMB URL to connect to
start_date_str (str): The start date in YYYY_MM_DD format
check_sandbox (bool): Whether to check for matching sandbox logs
"""
# Remove 'smb://' prefix and parse URL components
url = smb_url[6:]
creds_server, share = url.split('/', 1)
creds, server = creds_server.rsplit('@', 1)
username, password = creds.split(':', 1)
password = unquote(password)
# Register the SMB session
smbclient.register_session(server, username=username, password=password)
# Create base data directory
data_dir = os.path.join(os.getcwd(), "data")
os.makedirs(data_dir, exist_ok=True)
# Convert start date string to datetime
start_date = datetime.strptime(start_date_str, "%Y_%m_%d")
try:
# Get current date for iteration
current_date = start_date
today = datetime.now()
while current_date <= today:
date_str = current_date.strftime("%Y_%m_%d")
folder_path = f"\\\\{server}\\{share}\\{date_str}\\conv_logs\\battle_anony"
# Get all sandbox session IDs for this date
sandbox_session_ids = get_sandbox_session_ids(server, share, date_str) if check_sandbox else set()
try:
# List all JSON files in the battle_anony folder
if smbclient.path.exists(folder_path):
for file_info in smbclient.scandir(folder_path):
# Skip macOS metadata files
if file_info.name.startswith('._'):
continue
if file_info.name.endswith('.json'):
file_path = f"{folder_path}\\{file_info.name}"
ip = get_ip_from_jsonl(file_path)
if ip:
# Create directory structure for this IP
ip_dir = os.path.join(data_dir, ip)
valid_dir = os.path.join(ip_dir, "valid")
invalid_dir = os.path.join(ip_dir, "invalid")
os.makedirs(valid_dir, exist_ok=True)
os.makedirs(invalid_dir, exist_ok=True)
# Check if chat_session_id exists in sandbox logs
if check_sandbox:
chat_session_id = get_chat_session_id(file_path)
has_sandbox = chat_session_id in sandbox_session_ids if chat_session_id else False
target_dir = valid_dir if has_sandbox else invalid_dir
else:
# When sandbox checking is disabled, put everything in valid
target_dir = valid_dir
# Download the file
local_file_path = os.path.join(target_dir, file_info.name)
try:
with smbclient.open_file(file_path, mode='rb') as remote_file:
with open(local_file_path, 'wb') as local_file:
shutil.copyfileobj(remote_file, local_file)
log.info(f"Downloaded {file_info.name} to {target_dir}")
except Exception as e:
log.error(f"Error downloading file {file_info.name}: {e}")
except Exception as e:
log.error(f"Error processing folder {date_str}: {e}")
# Move to next day
current_date += timedelta(days=1)
except Exception as e:
log.error(f"Error accessing SMB share: {e}")
def main():
smb_url = os.getenv("SMB_URL")
# Add argument parser for optional parameters
parser = argparse.ArgumentParser(description='Download and organize conversation files by IP')
parser.add_argument('--sandbox-check', action='store_true', help='Check for matching sandbox logs')
parser.add_argument('--download', action='store_true', help='Enable file download')
args = parser.parse_args()
# Download files if enabled
if args.download:
print("\nDownloading files and organizing by IP address...")
download_files_by_ip(smb_url, check_sandbox=args.sandbox_check)
# Count and display statistics
ip_counts = count_files_per_ip(smb_url)
print("\nFile counts per IP address:")
for ip, count in sorted(ip_counts.items(), key=lambda x: x[1], reverse=True):
print(f"IP: {ip:<15} Count: {count}")
if __name__ == "__main__":
main()