|
|
|
|
|
|
|
|
import os |
|
|
import json |
|
|
import logging |
|
|
from datetime import datetime, timedelta |
|
|
import tiktoken |
|
|
import threading |
|
|
import time |
|
|
|
|
|
|
|
|
DATA_LOCK = threading.Lock() |
|
|
|
|
|
|
|
|
BASE_DIR = os.path.dirname(os.path.abspath(__file__)) |
|
|
DATA_FILE = os.path.join(BASE_DIR, "bot_data.json") |
|
|
LOG_FILE = os.path.join(BASE_DIR, "bot.log") |
|
|
|
|
|
|
|
|
DATA = { |
|
|
"users": {}, |
|
|
"groups": {}, |
|
|
"banned_users": set(), |
|
|
"stats": { |
|
|
"total_messages": 0, |
|
|
"total_users": 0, |
|
|
"total_groups": 0, |
|
|
"avg_response_time": 0.0, |
|
|
"max_response_time": 0.0, |
|
|
"min_response_time": float('inf'), |
|
|
"total_responses": 0 |
|
|
}, |
|
|
"welcome_message": "سلام {user_mention}! 🤖\n\nمن یک ربات هوشمند هستم. هر سوالی دارید بپرسید.", |
|
|
"group_welcome_message": "سلام به همه! 🤖\n\nمن یک ربات هوشمند هستم. برای استفاده از من در گروه:\n1. مستقیم با من چت کنید\n2. یا با منشن کردن من سوال بپرسید", |
|
|
"goodbye_message": "کاربر {user_mention} گروه را ترک کرد. خداحافظ!", |
|
|
"maintenance_mode": False, |
|
|
"blocked_words": [], |
|
|
"scheduled_broadcasts": [], |
|
|
"bot_start_time": datetime.now().strftime('%Y-%m-%d %H:%M:%S'), |
|
|
"context_mode": "hybrid", |
|
|
"system_instructions": { |
|
|
"global": "شما یک دستیار هوشمند فارسی هستید. پاسخهای شما باید مفید، دقیق و دوستانه باشد. از کلمات فارسی صحیح استفاده کنید و در صورت نیاز توضیحات کامل ارائه دهید.", |
|
|
"groups": {}, |
|
|
"users": {} |
|
|
}, |
|
|
"hybrid_settings": { |
|
|
"mode": "advanced", |
|
|
"enable_user_tracking": True, |
|
|
"max_other_users": 5, |
|
|
"personal_context_weight": 1.0, |
|
|
"group_context_weight": 0.8, |
|
|
"other_users_weight": 0.6, |
|
|
"max_combined_tokens": 30000, |
|
|
"enable_cross_user_references": True |
|
|
} |
|
|
} |
|
|
|
|
|
logger = logging.getLogger(__name__) |
|
|
|
|
|
|
|
|
def count_tokens(text: str) -> int: |
|
|
"""شمارش تعداد توکنهای متن با استفاده از tokenizer""" |
|
|
try: |
|
|
encoding = tiktoken.get_encoding("cl100k_base") |
|
|
return len(encoding.encode(text)) |
|
|
except Exception as e: |
|
|
logger.warning(f"Error counting tokens: {e}, using fallback") |
|
|
return max(1, len(text) // 4) |
|
|
|
|
|
def load_data(): |
|
|
"""دادهها را از فایل JSON بارگذاری کرده و در کش گلوبال ذخیره میکند.""" |
|
|
global DATA |
|
|
with DATA_LOCK: |
|
|
try: |
|
|
if not os.path.exists(DATA_FILE): |
|
|
logger.info(f"فایل داده در {DATA_FILE} یافت نشد. یک فایل جدید ایجاد میشود.") |
|
|
save_data() |
|
|
return |
|
|
|
|
|
with open(DATA_FILE, 'r', encoding='utf-8') as f: |
|
|
loaded_data = json.load(f) |
|
|
loaded_data['banned_users'] = set(loaded_data.get('banned_users', [])) |
|
|
|
|
|
|
|
|
if 'groups' not in loaded_data: loaded_data['groups'] = {} |
|
|
if 'context_mode' not in loaded_data: loaded_data['context_mode'] = 'hybrid' |
|
|
if 'group_welcome_message' not in loaded_data: |
|
|
loaded_data['group_welcome_message'] = "سلام به همه! 🤖\n\nمن یک ربات هوشمند هستم..." |
|
|
if 'total_groups' not in loaded_data.get('stats', {}): |
|
|
if 'stats' not in loaded_data: loaded_data['stats'] = {} |
|
|
loaded_data['stats']['total_groups'] = len(loaded_data.get('groups', {})) |
|
|
|
|
|
|
|
|
if 'hybrid_settings' not in loaded_data: |
|
|
loaded_data['hybrid_settings'] = DATA['hybrid_settings'] |
|
|
|
|
|
|
|
|
for group_id in loaded_data.get('groups', {}): |
|
|
if 'members' in loaded_data['groups'][group_id]: |
|
|
if isinstance(loaded_data['groups'][group_id]['members'], list): |
|
|
loaded_data['groups'][group_id]['members'] = set(loaded_data['groups'][group_id]['members']) |
|
|
elif not isinstance(loaded_data['groups'][group_id]['members'], set): |
|
|
loaded_data['groups'][group_id]['members'] = set() |
|
|
else: |
|
|
loaded_data['groups'][group_id]['members'] = set() |
|
|
|
|
|
|
|
|
for group_id in loaded_data.get('groups', {}): |
|
|
if 'context' not in loaded_data['groups'][group_id]: |
|
|
loaded_data['groups'][group_id]['context'] = [] |
|
|
|
|
|
|
|
|
for user_id in loaded_data.get('users', {}): |
|
|
if 'context' not in loaded_data['users'][user_id]: |
|
|
loaded_data['users'][user_id]['context'] = [] |
|
|
|
|
|
|
|
|
if 'stats' in loaded_data: |
|
|
stats_keys = ['avg_response_time', 'max_response_time', 'min_response_time', 'total_responses'] |
|
|
for key in stats_keys: |
|
|
if key not in loaded_data['stats']: |
|
|
if key == 'min_response_time': |
|
|
loaded_data['stats'][key] = float('inf') |
|
|
else: |
|
|
loaded_data['stats'][key] = 0.0 |
|
|
|
|
|
|
|
|
if 'system_instructions' not in loaded_data: |
|
|
loaded_data['system_instructions'] = DATA['system_instructions'] |
|
|
|
|
|
if 'added_admins' not in loaded_data: |
|
|
loaded_data['added_admins'] = [] |
|
|
|
|
|
DATA.update(loaded_data) |
|
|
logger.info(f"دادهها با موفقیت از {DATA_FILE} بارگذاری شدند.") |
|
|
|
|
|
except json.JSONDecodeError as e: |
|
|
logger.error(f"خطا در خواندن JSON از {DATA_FILE}: {e}. ربات با دادههای اولیه شروع به کار میکند.") |
|
|
except Exception as e: |
|
|
logger.error(f"خطای غیرمنتظره هنگام بارگذاری دادهها: {e}. ربات با دادههای اولیه شروع به کار میکند.") |
|
|
|
|
|
def save_data(): |
|
|
"""کش گلوبال دادهها را در فایل JSON ذخیره میکند.""" |
|
|
global DATA |
|
|
with DATA_LOCK: |
|
|
try: |
|
|
data_to_save = DATA.copy() |
|
|
data_to_save['banned_users'] = list(DATA['banned_users']) |
|
|
|
|
|
|
|
|
for group_id in data_to_save.get('groups', {}): |
|
|
if 'members' in data_to_save['groups'][group_id]: |
|
|
if isinstance(data_to_save['groups'][group_id]['members'], set): |
|
|
data_to_save['groups'][group_id]['members'] = list(data_to_save['groups'][group_id]['members']) |
|
|
elif not isinstance(data_to_save['groups'][group_id]['members'], list): |
|
|
data_to_save['groups'][group_id]['members'] = [] |
|
|
|
|
|
with open(DATA_FILE, 'w', encoding='utf-8') as f: |
|
|
json.dump(data_to_save, f, indent=4, ensure_ascii=False) |
|
|
logger.debug(f"دادهها با موفقیت در {DATA_FILE} ذخیره شدند.") |
|
|
except Exception as e: |
|
|
logger.error(f"خطای مهلک: امکان ذخیره دادهها در {DATA_FILE} وجود ندارد. خطا: {e}") |
|
|
|
|
|
|
|
|
def update_user_stats(user_id: int, user): |
|
|
"""آمار کاربر را پس از هر پیام بهروز کرده و دادهها را ذخیره میکند.""" |
|
|
global DATA |
|
|
with DATA_LOCK: |
|
|
now_str = datetime.now().strftime('%Y-%m-%d %H:%M:%S') |
|
|
user_id_str = str(user_id) |
|
|
|
|
|
if user_id_str not in DATA['users']: |
|
|
DATA['users'][user_id_str] = { |
|
|
'first_name': user.first_name, |
|
|
'username': user.username, |
|
|
'first_seen': now_str, |
|
|
'message_count': 0, |
|
|
'context': [], |
|
|
'last_group_activity': {}, |
|
|
'user_profile': { |
|
|
'language_preference': 'fa', |
|
|
'interaction_style': 'normal' |
|
|
} |
|
|
} |
|
|
DATA['stats']['total_users'] += 1 |
|
|
logger.info(f"کاربر جدید ثبت شد: {user_id} ({user.first_name})") |
|
|
|
|
|
DATA['users'][user_id_str]['last_seen'] = now_str |
|
|
DATA['users'][user_id_str]['message_count'] += 1 |
|
|
DATA['stats']['total_messages'] += 1 |
|
|
|
|
|
save_data() |
|
|
|
|
|
def update_response_stats(response_time: float): |
|
|
"""آمار زمان پاسخگویی را بهروز میکند.""" |
|
|
global DATA |
|
|
with DATA_LOCK: |
|
|
if DATA['stats']['total_responses'] == 0: |
|
|
DATA['stats']['min_response_time'] = response_time |
|
|
|
|
|
DATA['stats']['total_responses'] += 1 |
|
|
|
|
|
|
|
|
current_avg = DATA['stats']['avg_response_time'] |
|
|
total_responses = DATA['stats']['total_responses'] |
|
|
new_avg = ((current_avg * (total_responses - 1)) + response_time) / total_responses |
|
|
DATA['stats']['avg_response_time'] = new_avg |
|
|
|
|
|
|
|
|
if response_time > DATA['stats']['max_response_time']: |
|
|
DATA['stats']['max_response_time'] = response_time |
|
|
|
|
|
if response_time < DATA['stats']['min_response_time']: |
|
|
DATA['stats']['min_response_time'] = response_time |
|
|
|
|
|
save_data() |
|
|
|
|
|
|
|
|
def is_user_banned(user_id: int) -> bool: |
|
|
"""بررسی میکند آیا کاربر مسدود شده است یا خیر.""" |
|
|
return user_id in DATA['banned_users'] |
|
|
|
|
|
def ban_user(user_id: int): |
|
|
"""کاربر را مسدود کرده و ذخیره میکند.""" |
|
|
with DATA_LOCK: |
|
|
DATA['banned_users'].add(user_id) |
|
|
save_data() |
|
|
|
|
|
def unban_user(user_id: int): |
|
|
"""مسدودیت کاربر را برداشته و ذخیره میکند.""" |
|
|
with DATA_LOCK: |
|
|
DATA['banned_users'].discard(user_id) |
|
|
save_data() |
|
|
|
|
|
def contains_blocked_words(text: str) -> bool: |
|
|
"""بررسی میکند آیا متن حاوی کلمات مسدود شده است یا خیر.""" |
|
|
if not DATA['blocked_words']: |
|
|
return False |
|
|
|
|
|
text_lower = text.lower() |
|
|
for word in DATA['blocked_words']: |
|
|
if word in text_lower: |
|
|
return True |
|
|
|
|
|
return False |
|
|
|
|
|
|
|
|
def get_active_users(days: int) -> list: |
|
|
"""لیست کاربران فعال در بازه زمانی مشخص را برمیگرداند.""" |
|
|
now = datetime.now() |
|
|
cutoff_date = now - timedelta(days=days) |
|
|
|
|
|
active_users = [] |
|
|
for user_id, user_info in DATA['users'].items(): |
|
|
if 'last_seen' in user_info: |
|
|
try: |
|
|
last_seen = datetime.strptime(user_info['last_seen'], '%Y-%m-%d %H:%M:%S') |
|
|
if last_seen >= cutoff_date: |
|
|
active_users.append(int(user_id)) |
|
|
except ValueError: |
|
|
continue |
|
|
|
|
|
return active_users |
|
|
|
|
|
def get_users_by_message_count(min_count: int) -> list: |
|
|
"""لیست کاربران با تعداد پیام بیشتر یا مساوی مقدار مشخص را برمیگرداند.""" |
|
|
users = [] |
|
|
for user_id, user_info in DATA['users'].items(): |
|
|
if user_info.get('message_count', 0) >= min_count: |
|
|
users.append(int(user_id)) |
|
|
|
|
|
return users |
|
|
|
|
|
|
|
|
def add_to_user_context(user_id: int, role: str, content: str): |
|
|
"""اضافه کردن پیام به context کاربر و محدود کردن به ۱۰۲۴ توکن""" |
|
|
add_to_user_context_with_reply(user_id, role, content, None) |
|
|
|
|
|
def add_to_user_context_with_reply(user_id: int, role: str, content: str, reply_info: dict = None): |
|
|
"""اضافه کردن پیام به context کاربر با اطلاعات ریپلای""" |
|
|
with DATA_LOCK: |
|
|
user_id_str = str(user_id) |
|
|
|
|
|
if user_id_str not in DATA['users']: |
|
|
return |
|
|
|
|
|
if 'context' not in DATA['users'][user_id_str]: |
|
|
DATA['users'][user_id_str]['context'] = [] |
|
|
|
|
|
|
|
|
message = { |
|
|
'role': role, |
|
|
'content': content, |
|
|
'timestamp': datetime.now().strftime('%Y-%m-%d %H:%M:%S'), |
|
|
'has_reply': False |
|
|
} |
|
|
|
|
|
if reply_info and 'replied_text' in reply_info and reply_info['replied_text']: |
|
|
message['has_reply'] = True |
|
|
message['reply_to_user_id'] = reply_info.get('replied_user_id') |
|
|
message['reply_to_user_name'] = reply_info.get('replied_user_name', 'Unknown') |
|
|
message['reply_to_content'] = reply_info['replied_text'] |
|
|
message['is_reply_to_bot'] = reply_info.get('is_reply_to_bot', False) |
|
|
message['reply_timestamp'] = reply_info.get('timestamp', time.time()) |
|
|
|
|
|
DATA['users'][user_id_str]['context'].append(message) |
|
|
|
|
|
|
|
|
trim_user_context(user_id, max_tokens=16384) |
|
|
|
|
|
save_data() |
|
|
|
|
|
def trim_user_context(user_id: int, max_tokens: int = 16384): |
|
|
"""کوتاه کردن context کاربر به تعداد توکن مشخص""" |
|
|
user_id_str = str(user_id) |
|
|
|
|
|
if user_id_str not in DATA['users'] or 'context' not in DATA['users'][user_id_str]: |
|
|
return |
|
|
|
|
|
context = DATA['users'][user_id_str]['context'] |
|
|
|
|
|
if not context: |
|
|
return |
|
|
|
|
|
|
|
|
total_tokens = sum(count_tokens(msg['content']) for msg in context) |
|
|
|
|
|
|
|
|
while total_tokens > max_tokens and len(context) > 1: |
|
|
removed_message = context.pop(0) |
|
|
total_tokens -= count_tokens(removed_message['content']) |
|
|
|
|
|
|
|
|
if total_tokens > max_tokens and context: |
|
|
last_message = context[-1] |
|
|
content = last_message['content'] |
|
|
|
|
|
tokens = count_tokens(content) |
|
|
if tokens > max_tokens: |
|
|
context.pop() |
|
|
else: |
|
|
while tokens > (max_tokens - (total_tokens - tokens)) and content: |
|
|
words = content.split() |
|
|
if len(words) > 1: |
|
|
content = ' '.join(words[:-1]) |
|
|
else: |
|
|
content = content[:-1] if len(content) > 1 else '' |
|
|
tokens = count_tokens(content) |
|
|
|
|
|
if content: |
|
|
context[-1]['content'] = content |
|
|
else: |
|
|
context.pop() |
|
|
|
|
|
save_data() |
|
|
|
|
|
def get_user_context(user_id: int) -> list: |
|
|
"""دریافت context کاربر""" |
|
|
user_id_str = str(user_id) |
|
|
|
|
|
if user_id_str not in DATA['users']: |
|
|
return [] |
|
|
|
|
|
return DATA['users'][user_id_str].get('context', []).copy() |
|
|
|
|
|
def clear_user_context(user_id: int): |
|
|
"""پاک کردن context کاربر""" |
|
|
with DATA_LOCK: |
|
|
user_id_str = str(user_id) |
|
|
|
|
|
if user_id_str in DATA['users']: |
|
|
if 'context' in DATA['users'][user_id_str]: |
|
|
DATA['users'][user_id_str]['context'] = [] |
|
|
save_data() |
|
|
logger.info(f"Context cleared for user {user_id}") |
|
|
|
|
|
def get_context_summary(user_id: int) -> str: |
|
|
"""خلاصهای از context کاربر""" |
|
|
context = get_user_context(user_id) |
|
|
if not context: |
|
|
return "بدون تاریخچه" |
|
|
|
|
|
total_messages = len(context) |
|
|
total_tokens = sum(count_tokens(msg['content']) for msg in context) |
|
|
|
|
|
|
|
|
reply_count = sum(1 for msg in context if msg.get('has_reply', False)) |
|
|
|
|
|
|
|
|
last_message = context[-1] if context else {} |
|
|
last_content = last_message.get('content', '')[:50] |
|
|
if len(last_message.get('content', '')) > 50: |
|
|
last_content += "..." |
|
|
|
|
|
return f"{total_messages} پیام ({total_tokens} توکن) - {reply_count} ریپلای - آخرین: {last_message.get('role', '')}: {last_content}" |
|
|
|
|
|
def get_context_for_api(user_id: int) -> list: |
|
|
"""فرمت context برای ارسال به API""" |
|
|
context = get_user_context(user_id) |
|
|
|
|
|
api_context = [] |
|
|
for msg in context: |
|
|
api_context.append({ |
|
|
'role': msg['role'], |
|
|
'content': msg['content'] |
|
|
}) |
|
|
|
|
|
return api_context |
|
|
|
|
|
def get_context_token_count(user_id: int) -> int: |
|
|
"""تعداد کل توکنهای context کاربر""" |
|
|
context = get_user_context(user_id) |
|
|
return sum(count_tokens(msg['content']) for msg in context) |
|
|
|
|
|
|
|
|
def update_group_stats(chat_id: int, chat, user_id: int = None): |
|
|
"""آمار گروه را بهروز کرده و دادهها را ذخیره میکند.""" |
|
|
global DATA |
|
|
with DATA_LOCK: |
|
|
now_str = datetime.now().strftime('%Y-%m-%d %H:%M:%S') |
|
|
chat_id_str = str(chat_id) |
|
|
|
|
|
if chat_id_str not in DATA['groups']: |
|
|
DATA['groups'][chat_id_str] = { |
|
|
'title': chat.title if hasattr(chat, 'title') else 'Private Group', |
|
|
'type': chat.type, |
|
|
'first_seen': now_str, |
|
|
'message_count': 0, |
|
|
'members': set(), |
|
|
'context': [], |
|
|
'recent_users': [], |
|
|
'last_activity': now_str |
|
|
} |
|
|
DATA['stats']['total_groups'] += 1 |
|
|
logger.info(f"گروه جدید ثبت شد: {chat_id} ({chat.title if hasattr(chat, 'title') else 'Private'})") |
|
|
|
|
|
DATA['groups'][chat_id_str]['last_seen'] = now_str |
|
|
DATA['groups'][chat_id_str]['last_activity'] = now_str |
|
|
DATA['groups'][chat_id_str]['message_count'] += 1 |
|
|
|
|
|
|
|
|
if user_id is not None: |
|
|
if 'members' not in DATA['groups'][chat_id_str]: |
|
|
DATA['groups'][chat_id_str]['members'] = set() |
|
|
elif isinstance(DATA['groups'][chat_id_str]['members'], list): |
|
|
DATA['groups'][chat_id_str]['members'] = set(DATA['groups'][chat_id_str]['members']) |
|
|
|
|
|
DATA['groups'][chat_id_str]['members'].add(user_id) |
|
|
|
|
|
|
|
|
if 'recent_users' not in DATA['groups'][chat_id_str]: |
|
|
DATA['groups'][chat_id_str]['recent_users'] = [] |
|
|
|
|
|
|
|
|
recent_users = DATA['groups'][chat_id_str]['recent_users'] |
|
|
recent_users = [uid for uid in recent_users if uid != user_id] |
|
|
|
|
|
|
|
|
recent_users.insert(0, user_id) |
|
|
|
|
|
|
|
|
if len(recent_users) > 20: |
|
|
recent_users = recent_users[:20] |
|
|
|
|
|
DATA['groups'][chat_id_str]['recent_users'] = recent_users |
|
|
|
|
|
save_data() |
|
|
|
|
|
def add_to_group_context(chat_id: int, role: str, content: str, user_name: str = None): |
|
|
"""اضافه کردن پیام به context گروه و محدود کردن به 16384 توکن""" |
|
|
add_to_group_context_with_reply(chat_id, role, content, user_name, None, None) |
|
|
|
|
|
def add_to_group_context_with_reply(chat_id: int, role: str, content: str, user_name: str = None, |
|
|
reply_info: dict = None, user_id: int = None): |
|
|
"""اضافه کردن پیام به context گروه با اطلاعات ریپلای""" |
|
|
with DATA_LOCK: |
|
|
chat_id_str = str(chat_id) |
|
|
|
|
|
if chat_id_str not in DATA['groups']: |
|
|
return |
|
|
|
|
|
if 'context' not in DATA['groups'][chat_id_str]: |
|
|
DATA['groups'][chat_id_str]['context'] = [] |
|
|
|
|
|
|
|
|
if user_name and role == 'user': |
|
|
content_with_name = f"{user_name}: {content}" |
|
|
else: |
|
|
content_with_name = content |
|
|
|
|
|
|
|
|
message = { |
|
|
'role': role, |
|
|
'content': content_with_name if role == 'user' else content, |
|
|
'timestamp': datetime.now().strftime('%Y-%m-%d %H:%M:%S'), |
|
|
'user_name': user_name if role == 'user' else None, |
|
|
'user_id': user_id, |
|
|
'has_reply': False, |
|
|
'message_type': 'group' |
|
|
} |
|
|
|
|
|
if reply_info and 'replied_text' in reply_info and reply_info['replied_text']: |
|
|
message['has_reply'] = True |
|
|
message['reply_to_user_id'] = reply_info.get('replied_user_id') |
|
|
message['reply_to_user_name'] = reply_info.get('replied_user_name', 'Unknown') |
|
|
message['reply_to_content'] = reply_info['replied_text'] |
|
|
message['is_reply_to_bot'] = reply_info.get('is_reply_to_bot', False) |
|
|
message['reply_timestamp'] = reply_info.get('timestamp', time.time()) |
|
|
|
|
|
DATA['groups'][chat_id_str]['context'].append(message) |
|
|
|
|
|
|
|
|
trim_group_context(chat_id, max_tokens=32768) |
|
|
|
|
|
save_data() |
|
|
|
|
|
def optimize_group_context(chat_id: int): |
|
|
"""بهینهسازی context گروه برای حفظ مهمترین پیامها""" |
|
|
chat_id_str = str(chat_id) |
|
|
|
|
|
if chat_id_str not in DATA['groups'] or 'context' not in DATA['groups'][chat_id_str]: |
|
|
return |
|
|
|
|
|
context = DATA['groups'][chat_id_str]['context'] |
|
|
if len(context) <= 10: |
|
|
return |
|
|
|
|
|
|
|
|
scored_messages = [] |
|
|
for i, msg in enumerate(context): |
|
|
score = 0 |
|
|
|
|
|
|
|
|
content_length = len(msg['content']) |
|
|
score += min(content_length / 100, 10) |
|
|
|
|
|
|
|
|
recency = len(context) - i |
|
|
score += min(recency / 10, 20) |
|
|
|
|
|
|
|
|
if msg['role'] == 'assistant': |
|
|
score += 15 |
|
|
|
|
|
|
|
|
if msg.get('has_reply', False): |
|
|
score += 10 |
|
|
|
|
|
|
|
|
if content_length < 20: |
|
|
score -= 5 |
|
|
|
|
|
scored_messages.append((score, msg)) |
|
|
|
|
|
|
|
|
scored_messages.sort(key=lambda x: x[0], reverse=True) |
|
|
|
|
|
|
|
|
keep_count = int(len(scored_messages) * 0.9) |
|
|
kept_messages = [msg for _, msg in scored_messages[:keep_count]] |
|
|
|
|
|
|
|
|
kept_messages.sort(key=lambda x: x.get('timestamp', '')) |
|
|
|
|
|
DATA['groups'][chat_id_str]['context'] = kept_messages |
|
|
save_data() |
|
|
|
|
|
logger.info(f"Optimized group {chat_id} context: {len(context)} -> {len(kept_messages)} messages") |
|
|
|
|
|
def trim_group_context(chat_id: int, max_tokens: int = 32768): |
|
|
"""کوتاه کردن context گروه به تعداد توکن مشخص""" |
|
|
chat_id_str = str(chat_id) |
|
|
|
|
|
if chat_id_str not in DATA['groups'] or 'context' not in DATA['groups'][chat_id_str]: |
|
|
return |
|
|
|
|
|
context = DATA['groups'][chat_id_str]['context'] |
|
|
|
|
|
if not context: |
|
|
return |
|
|
|
|
|
|
|
|
total_tokens = sum(count_tokens(msg['content']) for msg in context) |
|
|
|
|
|
|
|
|
if total_tokens > max_tokens * 1.5: |
|
|
optimize_group_context(chat_id) |
|
|
context = DATA['groups'][chat_id_str]['context'] |
|
|
total_tokens = sum(count_tokens(msg['content']) for msg in context) |
|
|
|
|
|
|
|
|
while total_tokens > max_tokens and len(context) > 1: |
|
|
removed_message = context.pop(0) |
|
|
total_tokens -= count_tokens(removed_message['content']) |
|
|
|
|
|
|
|
|
if total_tokens > max_tokens and context: |
|
|
last_message = context[-1] |
|
|
content = last_message['content'] |
|
|
|
|
|
tokens = count_tokens(content) |
|
|
if tokens > max_tokens: |
|
|
context.pop() |
|
|
else: |
|
|
while tokens > (max_tokens - (total_tokens - tokens)) and content: |
|
|
words = content.split() |
|
|
if len(words) > 1: |
|
|
content = ' '.join(words[:-1]) |
|
|
else: |
|
|
content = content[:-1] if len(content) > 1 else '' |
|
|
tokens = count_tokens(content) |
|
|
|
|
|
if content: |
|
|
context[-1]['content'] = content |
|
|
else: |
|
|
context.pop() |
|
|
|
|
|
save_data() |
|
|
|
|
|
def get_group_context(chat_id: int) -> list: |
|
|
"""دریافت context گروه""" |
|
|
chat_id_str = str(chat_id) |
|
|
|
|
|
if chat_id_str not in DATA['groups']: |
|
|
return [] |
|
|
|
|
|
return DATA['groups'][chat_id_str].get('context', []).copy() |
|
|
|
|
|
def get_context_for_api_group(chat_id: int) -> list: |
|
|
"""فرمت context گروه برای ارسال به API""" |
|
|
context = get_group_context(chat_id) |
|
|
|
|
|
api_context = [] |
|
|
for msg in context: |
|
|
api_context.append({ |
|
|
'role': msg['role'], |
|
|
'content': msg['content'] |
|
|
}) |
|
|
|
|
|
return api_context |
|
|
|
|
|
def clear_group_context(chat_id: int): |
|
|
"""پاک کردن context گروه""" |
|
|
with DATA_LOCK: |
|
|
chat_id_str = str(chat_id) |
|
|
|
|
|
if chat_id_str in DATA['groups']: |
|
|
if 'context' in DATA['groups'][chat_id_str]: |
|
|
DATA['groups'][chat_id_str]['context'] = [] |
|
|
save_data() |
|
|
logger.info(f"Context cleared for group {chat_id}") |
|
|
|
|
|
def get_group_context_summary(chat_id: int) -> str: |
|
|
"""خلاصهای از context گروه""" |
|
|
context = get_group_context(chat_id) |
|
|
if not context: |
|
|
return "بدون تاریخچه" |
|
|
|
|
|
total_messages = len(context) |
|
|
total_tokens = sum(count_tokens(msg['content']) for msg in context) |
|
|
|
|
|
|
|
|
reply_count = sum(1 for msg in context if msg.get('has_reply', False)) |
|
|
|
|
|
|
|
|
unique_users = set() |
|
|
for msg in context: |
|
|
if msg.get('user_id'): |
|
|
unique_users.add(msg['user_id']) |
|
|
|
|
|
last_message = context[-1] if context else {} |
|
|
last_content = last_message.get('content', '')[:50] |
|
|
if len(last_message.get('content', '')) > 50: |
|
|
last_content += "..." |
|
|
|
|
|
return f"{total_messages} پیام ({total_tokens} توکن) - {len(unique_users)} کاربر - {reply_count} ریپلای - آخرین: {last_content}" |
|
|
|
|
|
|
|
|
def add_to_hybrid_context_advanced(user_id: int, chat_id: int, role: str, content: str, |
|
|
user_name: str = None, reply_info: dict = None): |
|
|
""" |
|
|
ذخیره پیشرفته در context ترکیبی |
|
|
هم در تاریخچه شخصی و هم در تاریخچه گروه ذخیره میشود |
|
|
""" |
|
|
|
|
|
add_to_user_context_with_reply(user_id, role, content, reply_info) |
|
|
|
|
|
|
|
|
add_to_group_context_with_reply(chat_id, role, content, user_name, reply_info, user_id) |
|
|
|
|
|
def get_advanced_hybrid_context_for_api(user_id: int, chat_id: int) -> list: |
|
|
""" |
|
|
دریافت context ترکیبی پیشرفته برای ارسال به API |
|
|
ترکیب هوشمندانه: |
|
|
1. تاریخچه شخصی کاربر |
|
|
2. تاریخچه گروه (بدون تکراریها) |
|
|
3. اطلاعات کاربران دیگر فعال در گروه |
|
|
""" |
|
|
|
|
|
system_instruction = get_system_instruction(user_id=user_id, chat_id=chat_id) |
|
|
|
|
|
|
|
|
user_context = get_context_for_api(user_id) |
|
|
|
|
|
|
|
|
group_context_raw = get_group_context(chat_id) |
|
|
|
|
|
|
|
|
user_messages_in_personal = set() |
|
|
for msg in user_context: |
|
|
if msg['role'] == 'user': |
|
|
|
|
|
content = msg['content'] |
|
|
user_messages_in_personal.add(content[:100]) |
|
|
|
|
|
filtered_group_context = [] |
|
|
for msg in group_context_raw: |
|
|
if msg.get('user_id') == user_id and msg['role'] == 'user': |
|
|
|
|
|
content_snippet = msg['content'][:100] |
|
|
if content_snippet in user_messages_in_personal: |
|
|
continue |
|
|
filtered_group_context.append(msg) |
|
|
|
|
|
|
|
|
other_users_info = get_other_users_summary_in_group(user_id, chat_id) |
|
|
|
|
|
|
|
|
combined_context = [] |
|
|
|
|
|
|
|
|
if system_instruction: |
|
|
combined_context.append({"role": "system", "content": system_instruction}) |
|
|
|
|
|
|
|
|
hybrid_instruction = f""" |
|
|
شما در حال پاسخ به کاربر {user_id} ({get_user_name(user_id)}) در یک گروه هستید. |
|
|
|
|
|
اطلاعات وضعیت: |
|
|
- حالت: Hybrid پیشرفته |
|
|
- کاربر فعلی: {get_user_name(user_id)} (آیدی: {user_id}) |
|
|
- گروه: {get_group_name(chat_id)} |
|
|
- کاربران دیگر فعال در گروه: {len(other_users_info)} کاربر |
|
|
|
|
|
دستورالعملها: |
|
|
1. پاسخهای شما باید مخصوص کاربر فعلی ({get_user_name(user_id)}) باشد. |
|
|
2. میتوانید به کاربران دیگر نیز در پاسخ اشاره کنید اگر مرتبط است. |
|
|
3. از تاریخچه شخصی کاربر و تاریخچه گروه استفاده کنید. |
|
|
4. پاسخهای دقیق و مرتبط بدهید. |
|
|
""" |
|
|
|
|
|
combined_context.append({"role": "system", "content": hybrid_instruction}) |
|
|
|
|
|
|
|
|
if other_users_info and DATA['hybrid_settings']['enable_cross_user_references']: |
|
|
other_users_summary = "📋 **اطلاعات کاربران دیگر در گروه:**\n" |
|
|
for info in other_users_info: |
|
|
other_users_summary += f"- {info['name']}: {info['summary']}\n" |
|
|
|
|
|
combined_context.append({ |
|
|
"role": "system", |
|
|
"content": other_users_summary[:500] + "..." if len(other_users_summary) > 500 else other_users_summary |
|
|
}) |
|
|
|
|
|
|
|
|
combined_context.extend(user_context[-10:]) |
|
|
|
|
|
|
|
|
for msg in filtered_group_context[-20:]: |
|
|
combined_context.append({ |
|
|
'role': msg['role'], |
|
|
'content': msg['content'] |
|
|
}) |
|
|
|
|
|
|
|
|
return limit_context_tokens(combined_context, DATA['hybrid_settings']['max_combined_tokens']) |
|
|
|
|
|
def get_other_users_summary_in_group(current_user_id: int, chat_id: int) -> list: |
|
|
"""دریافت خلاصه اطلاعات کاربران دیگر در گروه""" |
|
|
chat_id_str = str(chat_id) |
|
|
|
|
|
if chat_id_str not in DATA['groups']: |
|
|
return [] |
|
|
|
|
|
group_info = DATA['groups'][chat_id_str] |
|
|
recent_users = group_info.get('recent_users', []) |
|
|
|
|
|
other_users_info = [] |
|
|
|
|
|
for other_user_id in recent_users[:DATA['hybrid_settings']['max_other_users']]: |
|
|
if other_user_id == current_user_id: |
|
|
continue |
|
|
|
|
|
user_info = DATA['users'].get(str(other_user_id), {}) |
|
|
if not user_info: |
|
|
continue |
|
|
|
|
|
user_name = user_info.get('first_name', 'کاربر') |
|
|
message_count = user_info.get('message_count', 0) |
|
|
last_seen = user_info.get('last_seen', 'نامشخص') |
|
|
|
|
|
|
|
|
last_group_message = None |
|
|
for msg in reversed(group_info.get('context', [])): |
|
|
if msg.get('user_id') == other_user_id and msg['role'] == 'user': |
|
|
last_group_message = msg['content'][:100] |
|
|
if len(msg['content']) > 100: |
|
|
last_group_message += "..." |
|
|
break |
|
|
|
|
|
|
|
|
summary = f"{message_count} پیام" |
|
|
if last_group_message: |
|
|
summary += f" - آخرین: {last_group_message}" |
|
|
|
|
|
other_users_info.append({ |
|
|
'user_id': other_user_id, |
|
|
'name': user_name, |
|
|
'message_count': message_count, |
|
|
'last_seen': last_seen, |
|
|
'last_message': last_group_message, |
|
|
'summary': summary |
|
|
}) |
|
|
|
|
|
return other_users_info |
|
|
|
|
|
def get_recent_users_in_group(chat_id: int, limit: int = 5) -> list: |
|
|
"""دریافت لیست کاربران اخیراً فعال در گروه""" |
|
|
chat_id_str = str(chat_id) |
|
|
|
|
|
if chat_id_str not in DATA['groups']: |
|
|
return [] |
|
|
|
|
|
group_info = DATA['groups'][chat_id_str] |
|
|
recent_users = group_info.get('recent_users', []) |
|
|
|
|
|
return recent_users[:limit] |
|
|
|
|
|
def get_user_name(user_id: int) -> str: |
|
|
"""دریافت نام کاربر""" |
|
|
user_id_str = str(user_id) |
|
|
user_info = DATA['users'].get(user_id_str, {}) |
|
|
return user_info.get('first_name', f"کاربر {user_id}") |
|
|
|
|
|
def get_group_name(chat_id: int) -> str: |
|
|
"""دریافت نام گروه""" |
|
|
chat_id_str = str(chat_id) |
|
|
group_info = DATA['groups'].get(chat_id_str, {}) |
|
|
return group_info.get('title', f"گروه {chat_id}") |
|
|
|
|
|
def limit_context_tokens(context: list, max_tokens: int) -> list: |
|
|
"""محدود کردن context به تعداد توکن مشخص""" |
|
|
if not context: |
|
|
return context |
|
|
|
|
|
total_tokens = sum(count_tokens(msg['content']) for msg in context) |
|
|
|
|
|
if total_tokens <= max_tokens: |
|
|
return context |
|
|
|
|
|
|
|
|
limited_context = context.copy() |
|
|
while total_tokens > max_tokens and len(limited_context) > 1: |
|
|
|
|
|
if limited_context[0]['role'] == 'system': |
|
|
|
|
|
if len(limited_context) > 2: |
|
|
removed = limited_context.pop(1) |
|
|
total_tokens -= count_tokens(removed['content']) |
|
|
else: |
|
|
break |
|
|
else: |
|
|
removed = limited_context.pop(0) |
|
|
total_tokens -= count_tokens(removed['content']) |
|
|
|
|
|
return limited_context |
|
|
|
|
|
def get_hybrid_context_summary(user_id: int, chat_id: int) -> str: |
|
|
"""خلاصهای از context ترکیبی پیشرفته""" |
|
|
personal_summary = get_context_summary(user_id) |
|
|
group_summary = get_group_context_summary(chat_id) |
|
|
|
|
|
|
|
|
other_users = get_recent_users_in_group(chat_id, limit=5) |
|
|
other_users_count = len([uid for uid in other_users if uid != user_id]) |
|
|
|
|
|
hybrid_mode = DATA['hybrid_settings']['mode'] |
|
|
|
|
|
return ( |
|
|
f"📊 **حالت Hybrid {hybrid_mode} فعال**\n\n" |
|
|
f"👤 **شخصی شما:** {personal_summary}\n" |
|
|
f"👥 **گروه:** {group_summary}\n" |
|
|
f"👥 **کاربران فعال دیگر در گروه:** {other_users_count} کاربر\n" |
|
|
f"⚙️ **تنظیمات:**\n" |
|
|
f"• وزن تاریخچه شخصی: {DATA['hybrid_settings']['personal_context_weight']}\n" |
|
|
f"• وزن تاریخچه گروه: {DATA['hybrid_settings']['group_context_weight']}\n" |
|
|
f"• ارجاع به کاربران دیگر: {'فعال ✅' if DATA['hybrid_settings']['enable_cross_user_references'] else 'غیرفعال ❌'}" |
|
|
) |
|
|
|
|
|
|
|
|
def get_context_mode() -> str: |
|
|
"""دریافت حالت فعلی context""" |
|
|
return DATA.get('context_mode', 'hybrid') |
|
|
|
|
|
def set_context_mode(mode: str): |
|
|
"""تنظیم حالت context""" |
|
|
valid_modes = ['separate', 'group_shared', 'hybrid'] |
|
|
if mode in valid_modes: |
|
|
DATA['context_mode'] = mode |
|
|
save_data() |
|
|
logger.info(f"Context mode changed to: {mode}") |
|
|
return True |
|
|
return False |
|
|
|
|
|
def set_hybrid_mode(mode: str): |
|
|
"""تنظیم حالت Hybrid""" |
|
|
if mode in ['basic', 'advanced']: |
|
|
if 'hybrid_settings' not in DATA: |
|
|
DATA['hybrid_settings'] = {} |
|
|
|
|
|
DATA['hybrid_settings']['mode'] = mode |
|
|
|
|
|
if mode == 'advanced': |
|
|
|
|
|
DATA['hybrid_settings']['enable_user_tracking'] = True |
|
|
DATA['hybrid_settings']['max_other_users'] = 5 |
|
|
DATA['hybrid_settings']['personal_context_weight'] = 1.0 |
|
|
DATA['hybrid_settings']['group_context_weight'] = 0.8 |
|
|
DATA['hybrid_settings']['other_users_weight'] = 0.6 |
|
|
DATA['hybrid_settings']['max_combined_tokens'] = 30000 |
|
|
DATA['hybrid_settings']['enable_cross_user_references'] = True |
|
|
else: |
|
|
|
|
|
DATA['hybrid_settings']['enable_user_tracking'] = True |
|
|
DATA['hybrid_settings']['max_other_users'] = 3 |
|
|
DATA['hybrid_settings']['personal_context_weight'] = 1.0 |
|
|
DATA['hybrid_settings']['group_context_weight'] = 0.7 |
|
|
DATA['hybrid_settings']['other_users_weight'] = 0.4 |
|
|
DATA['hybrid_settings']['max_combined_tokens'] = 25000 |
|
|
DATA['hybrid_settings']['enable_cross_user_references'] = False |
|
|
|
|
|
save_data() |
|
|
logger.info(f"Hybrid mode set to: {mode}") |
|
|
return True |
|
|
return False |
|
|
|
|
|
|
|
|
def get_system_instruction(user_id: int = None, chat_id: int = None) -> str: |
|
|
"""دریافت system instruction مناسب برای کاربر/گروه""" |
|
|
user_id_str = str(user_id) if user_id else None |
|
|
chat_id_str = str(chat_id) if chat_id else None |
|
|
|
|
|
|
|
|
if user_id_str and user_id_str in DATA['system_instructions']['users']: |
|
|
instruction_data = DATA['system_instructions']['users'][user_id_str] |
|
|
if instruction_data.get('enabled', True): |
|
|
return instruction_data.get('instruction', '') |
|
|
|
|
|
|
|
|
if chat_id_str and chat_id_str in DATA['system_instructions']['groups']: |
|
|
instruction_data = DATA['system_instructions']['groups'][chat_id_str] |
|
|
if instruction_data.get('enabled', True): |
|
|
return instruction_data.get('instruction', '') |
|
|
|
|
|
|
|
|
return DATA['system_instructions'].get('global', '') |
|
|
|
|
|
def set_global_system_instruction(instruction: str) -> bool: |
|
|
"""تنظیم system instruction سراسری""" |
|
|
DATA['system_instructions']['global'] = instruction |
|
|
save_data() |
|
|
return True |
|
|
|
|
|
def set_group_system_instruction(chat_id: int, instruction: str, set_by_user_id: int) -> bool: |
|
|
"""تنظیم system instruction برای گروه""" |
|
|
chat_id_str = str(chat_id) |
|
|
DATA['system_instructions']['groups'][chat_id_str] = { |
|
|
'instruction': instruction, |
|
|
'set_by': set_by_user_id, |
|
|
'set_at': datetime.now().strftime('%Y-%m-%d %H:%M:%S'), |
|
|
'enabled': True |
|
|
} |
|
|
save_data() |
|
|
return True |
|
|
|
|
|
def set_user_system_instruction(user_id: int, instruction: str, set_by_user_id: int) -> bool: |
|
|
"""تنظیم system instruction برای کاربر""" |
|
|
user_id_str = str(user_id) |
|
|
DATA['system_instructions']['users'][user_id_str] = { |
|
|
'instruction': instruction, |
|
|
'set_by': set_by_user_id, |
|
|
'set_at': datetime.now().strftime('%Y-%m-%d %H:%M:%S'), |
|
|
'enabled': True |
|
|
} |
|
|
save_data() |
|
|
return True |
|
|
|
|
|
def remove_group_system_instruction(chat_id: int) -> bool: |
|
|
"""حذف system instruction گروه""" |
|
|
chat_id_str = str(chat_id) |
|
|
if chat_id_str in DATA['system_instructions']['groups']: |
|
|
del DATA['system_instructions']['groups'][chat_id_str] |
|
|
save_data() |
|
|
return True |
|
|
return False |
|
|
|
|
|
def remove_user_system_instruction(user_id: int) -> bool: |
|
|
"""حذف system instruction کاربر""" |
|
|
user_id_str = str(user_id) |
|
|
if user_id_str in DATA['system_instructions']['users']: |
|
|
del DATA['system_instructions']['users'][user_id_str] |
|
|
save_data() |
|
|
return True |
|
|
return False |
|
|
|
|
|
def toggle_group_system_instruction(chat_id: int, enabled: bool) -> bool: |
|
|
"""فعال/غیرفعال کردن system instruction گروه""" |
|
|
chat_id_str = str(chat_id) |
|
|
if chat_id_str in DATA['system_instructions']['groups']: |
|
|
DATA['system_instructions']['groups'][chat_id_str]['enabled'] = enabled |
|
|
save_data() |
|
|
return True |
|
|
return False |
|
|
|
|
|
def toggle_user_system_instruction(user_id: int, enabled: bool) -> bool: |
|
|
"""فعال/غیرفعال کردن system instruction کاربر""" |
|
|
user_id_str = str(user_id) |
|
|
if user_id_str in DATA['system_instructions']['users']: |
|
|
DATA['system_instructions']['users'][user_id_str]['enabled'] = enabled |
|
|
save_data() |
|
|
return True |
|
|
return False |
|
|
|
|
|
def get_system_instruction_info(user_id: int = None, chat_id: int = None) -> dict: |
|
|
"""دریافت اطلاعات system instruction""" |
|
|
user_id_str = str(user_id) if user_id else None |
|
|
chat_id_str = str(chat_id) if chat_id else None |
|
|
|
|
|
result = { |
|
|
'has_instruction': False, |
|
|
'type': 'global', |
|
|
'instruction': DATA['system_instructions'].get('global', ''), |
|
|
'details': None |
|
|
} |
|
|
|
|
|
if user_id_str and user_id_str in DATA['system_instructions']['users']: |
|
|
user_data = DATA['system_instructions']['users'][user_id_str] |
|
|
result.update({ |
|
|
'has_instruction': True, |
|
|
'type': 'user', |
|
|
'instruction': user_data.get('instruction', ''), |
|
|
'details': user_data |
|
|
}) |
|
|
elif chat_id_str and chat_id_str in DATA['system_instructions']['groups']: |
|
|
group_data = DATA['system_instructions']['groups'][chat_id_str] |
|
|
result.update({ |
|
|
'has_instruction': True, |
|
|
'type': 'group', |
|
|
'instruction': group_data.get('instruction', ''), |
|
|
'details': group_data |
|
|
}) |
|
|
|
|
|
return result |
|
|
|
|
|
|
|
|
load_data() |
|
|
logger.info(f"Data loaded. Context mode: {get_context_mode()}, Hybrid mode: {DATA['hybrid_settings']['mode']}") |