File size: 19,963 Bytes
0a69927 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 |
import time
import re
import os
from typing import List, Dict, Tuple, Union
from colorama import Fore, Style
import logging
import sys
from io import StringIO
from web_scraper import get_web_content, can_fetch
from llm_config import get_llm_config
from llm_response_parser import UltimateLLMResponseParser
from llm_wrapper import LLMWrapper
from urllib.parse import urlparse
# Set up logging
log_directory = 'logs'
if not os.path.exists(log_directory):
os.makedirs(log_directory)
# Configure logger
logger = logging.getLogger(__name__)
logger.setLevel(logging.INFO)
log_file = os.path.join(log_directory, 'llama_output.log')
file_handler = logging.FileHandler(log_file)
formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')
file_handler.setFormatter(formatter)
logger.handlers = []
logger.addHandler(file_handler)
logger.propagate = False
# Suppress other loggers
for name in ['root', 'duckduckgo_search', 'requests', 'urllib3']:
logging.getLogger(name).setLevel(logging.WARNING)
logging.getLogger(name).handlers = []
logging.getLogger(name).propagate = False
class OutputRedirector:
def __init__(self, stream=None):
self.stream = stream or StringIO()
self.original_stdout = sys.stdout
self.original_stderr = sys.stderr
def __enter__(self):
sys.stdout = self.stream
sys.stderr = self.stream
return self.stream
def __exit__(self, exc_type, exc_val, exc_tb):
sys.stdout = self.original_stdout
sys.stderr = self.original_stderr
class EnhancedSelfImprovingSearch:
def __init__(self, llm: LLMWrapper, parser: UltimateLLMResponseParser, max_attempts: int = 5):
self.llm = llm
self.parser = parser
self.max_attempts = max_attempts
self.llm_config = get_llm_config()
@staticmethod
def initialize_llm():
llm_wrapper = LLMWrapper()
return llm_wrapper
def print_thinking(self):
print(Fore.MAGENTA + "🧠 Thinking..." + Style.RESET_ALL)
def print_searching(self):
print(Fore.MAGENTA + "📝 Searching..." + Style.RESET_ALL)
def search_and_improve(self, user_query: str) -> str:
attempt = 0
while attempt < self.max_attempts:
print(f"\n{Fore.CYAN}Search attempt {attempt + 1}:{Style.RESET_ALL}")
self.print_searching()
try:
formulated_query, time_range = self.formulate_query(user_query, attempt)
print(f"{Fore.YELLOW}Original query: {user_query}{Style.RESET_ALL}")
print(f"{Fore.YELLOW}Formulated query: {formulated_query}{Style.RESET_ALL}")
print(f"{Fore.YELLOW}Time range: {time_range}{Style.RESET_ALL}")
if not formulated_query:
print(f"{Fore.RED}Error: Empty search query. Retrying...{Style.RESET_ALL}")
attempt += 1
continue
search_results = self.perform_search(formulated_query, time_range)
if not search_results:
print(f"{Fore.RED}No results found. Retrying with a different query...{Style.RESET_ALL}")
attempt += 1
continue
self.display_search_results(search_results)
selected_urls = self.select_relevant_pages(search_results, user_query)
if not selected_urls:
print(f"{Fore.RED}No relevant URLs found. Retrying...{Style.RESET_ALL}")
attempt += 1
continue
print(Fore.MAGENTA + "⚙️ Scraping selected pages..." + Style.RESET_ALL)
# Scraping is done without OutputRedirector to ensure messages are visible
scraped_content = self.scrape_content(selected_urls)
if not scraped_content:
print(f"{Fore.RED}Failed to scrape content. Retrying...{Style.RESET_ALL}")
attempt += 1
continue
self.display_scraped_content(scraped_content)
self.print_thinking()
with OutputRedirector() as output:
evaluation, decision = self.evaluate_scraped_content(user_query, scraped_content)
llm_output = output.getvalue()
logger.info(f"LLM Output in evaluate_scraped_content:\n{llm_output}")
print(f"{Fore.MAGENTA}Evaluation: {evaluation}{Style.RESET_ALL}")
print(f"{Fore.MAGENTA}Decision: {decision}{Style.RESET_ALL}")
if decision == "answer":
return self.generate_final_answer(user_query, scraped_content)
elif decision == "refine":
print(f"{Fore.YELLOW}Refining search...{Style.RESET_ALL}")
attempt += 1
else:
print(f"{Fore.RED}Unexpected decision. Proceeding to answer.{Style.RESET_ALL}")
return self.generate_final_answer(user_query, scraped_content)
except Exception as e:
print(f"{Fore.RED}An error occurred during search attempt. Check the log file for details.{Style.RESET_ALL}")
logger.error(f"An error occurred during search: {str(e)}", exc_info=True)
attempt += 1
return self.synthesize_final_answer(user_query)
def evaluate_scraped_content(self, user_query: str, scraped_content: Dict[str, str]) -> Tuple[str, str]:
user_query_short = user_query[:200]
prompt = f"""
Evaluate if the following scraped content contains sufficient information to answer the user's question comprehensively:
User's question: "{user_query_short}"
Scraped Content:
{self.format_scraped_content(scraped_content)}
Your task:
1. Determine if the scraped content provides enough relevant and detailed information to answer the user's question thoroughly.
2. If the information is sufficient, decide to 'answer'. If more information or clarification is needed, decide to 'refine' the search.
Respond using EXACTLY this format:
Evaluation: [Your evaluation of the scraped content]
Decision: [ONLY 'answer' if content is sufficient, or 'refine' if more information is needed]
"""
max_retries = 3
for attempt in range(max_retries):
try:
response_text = self.llm.generate(prompt, max_tokens=200, stop=None)
evaluation, decision = self.parse_evaluation_response(response_text)
if decision in ['answer', 'refine']:
return evaluation, decision
except Exception as e:
logger.warning(f"Error in evaluate_scraped_content (attempt {attempt + 1}): {str(e)}")
logger.warning("Failed to get a valid decision in evaluate_scraped_content. Defaulting to 'refine'.")
return "Failed to evaluate content.", "refine"
def parse_evaluation_response(self, response: str) -> Tuple[str, str]:
evaluation = ""
decision = ""
for line in response.strip().split('\n'):
if line.startswith('Evaluation:'):
evaluation = line.split(':', 1)[1].strip()
elif line.startswith('Decision:'):
decision = line.split(':', 1)[1].strip().lower()
return evaluation, decision
def formulate_query(self, user_query: str, attempt: int) -> Tuple[str, str]:
user_query_short = user_query[:200]
prompt = f"""
Based on the following user question, formulate a concise and effective search query:
"{user_query_short}"
Your task:
1. Create a search query of 2-5 words that will yield relevant results.
2. Determine if a specific time range is needed for the search.
Time range options:
- 'd': Limit results to the past day. Use for very recent events or rapidly changing information.
- 'w': Limit results to the past week. Use for recent events or topics with frequent updates.
- 'm': Limit results to the past month. Use for relatively recent information or ongoing events.
- 'y': Limit results to the past year. Use for annual events or information that changes yearly.
- 'none': No time limit. Use for historical information or topics not tied to a specific time frame.
Respond in the following format:
Search query: [Your 2-5 word query]
Time range: [d/w/m/y/none]
Do not provide any additional information or explanation.
"""
max_retries = 3
for retry in range(max_retries):
with OutputRedirector() as output:
response_text = self.llm.generate(prompt, max_tokens=50, stop=None)
llm_output = output.getvalue()
logger.info(f"LLM Output in formulate_query:\n{llm_output}")
query, time_range = self.parse_query_response(response_text)
if query and time_range:
return query, time_range
return self.fallback_query(user_query), "none"
def parse_query_response(self, response: str) -> Tuple[str, str]:
query = ""
time_range = "none"
for line in response.strip().split('\n'):
if ":" in line:
key, value = line.split(":", 1)
key = key.strip().lower()
value = value.strip()
if "query" in key:
query = self.clean_query(value)
elif "time" in key or "range" in key:
time_range = self.validate_time_range(value)
return query, time_range
def clean_query(self, query: str) -> str:
query = re.sub(r'["\'\[\]]', '', query)
query = re.sub(r'\s+', ' ', query)
return query.strip()[:100]
def validate_time_range(self, time_range: str) -> str:
valid_ranges = ['d', 'w', 'm', 'y', 'none']
time_range = time_range.lower()
return time_range if time_range in valid_ranges else 'none'
def fallback_query(self, user_query: str) -> str:
words = user_query.split()
return " ".join(words[:5])
def perform_search(self, query: str, time_range: str) -> List[Dict]:
if not query:
return []
from duckduckgo_search import DDGS
with DDGS() as ddgs:
try:
with OutputRedirector() as output:
if time_range and time_range != 'none':
results = list(ddgs.text(query, timelimit=time_range, max_results=10))
else:
results = list(ddgs.text(query, max_results=10))
ddg_output = output.getvalue()
logger.info(f"DDG Output in perform_search:\n{ddg_output}")
print(f"{Fore.GREEN}Search query sent to DuckDuckGo: {query}{Style.RESET_ALL}")
print(f"{Fore.GREEN}Time range sent to DuckDuckGo: {time_range}{Style.RESET_ALL}")
print(f"{Fore.GREEN}Number of results: {len(results)}{Style.RESET_ALL}")
return [{'number': i+1, **result} for i, result in enumerate(results)]
except Exception as e:
print(f"{Fore.RED}Search error: {str(e)}{Style.RESET_ALL}")
return []
def display_search_results(self, results: List[Dict]):
print(f"\n{Fore.CYAN}Search Results:{Style.RESET_ALL}")
for result in results:
print(f"{Fore.GREEN}Result {result['number']}:{Style.RESET_ALL}")
print(f"Title: {result.get('title', 'N/A')}")
print(f"Snippet: {result.get('body', 'N/A')[:200]}...")
print(f"URL: {result.get('href', 'N/A')}\n")
def select_relevant_pages(self, search_results: List[Dict], user_query: str) -> List[str]:
prompt = f"""
Given the following search results for the user's question: "{user_query}"
Select the 2 most relevant results to scrape and analyze. Explain your reasoning for each selection.
Search Results:
{self.format_results(search_results)}
Instructions:
1. You MUST select exactly 2 result numbers from the search results.
2. Choose the results that are most likely to contain comprehensive and relevant information to answer the user's question.
3. Provide a brief reason for each selection.
You MUST respond using EXACTLY this format and nothing else:
Selected Results: [Two numbers corresponding to the selected results]
Reasoning: [Your reasoning for the selections]
"""
max_retries = 3
for retry in range(max_retries):
with OutputRedirector() as output:
response_text = self.llm.generate(prompt, max_tokens=200, stop=None)
llm_output = output.getvalue()
logger.info(f"LLM Output in select_relevant_pages:\n{llm_output}")
parsed_response = self.parse_page_selection_response(response_text)
if parsed_response and self.validate_page_selection_response(parsed_response, len(search_results)):
selected_urls = [result['href'] for result in search_results if result['number'] in parsed_response['selected_results']]
allowed_urls = [url for url in selected_urls if can_fetch(url)]
if allowed_urls:
return allowed_urls
else:
print(f"{Fore.YELLOW}Warning: All selected URLs are disallowed by robots.txt. Retrying selection.{Style.RESET_ALL}")
else:
print(f"{Fore.YELLOW}Warning: Invalid page selection. Retrying.{Style.RESET_ALL}")
print(f"{Fore.YELLOW}Warning: All attempts to select relevant pages failed. Falling back to top allowed results.{Style.RESET_ALL}")
allowed_urls = [result['href'] for result in search_results if can_fetch(result['href'])][:2]
return allowed_urls
def parse_page_selection_response(self, response: str) -> Dict[str, Union[List[int], str]]:
lines = response.strip().split('\n')
parsed = {}
for line in lines:
if line.startswith('Selected Results:'):
parsed['selected_results'] = [int(num.strip()) for num in re.findall(r'\d+', line)]
elif line.startswith('Reasoning:'):
parsed['reasoning'] = line.split(':', 1)[1].strip()
return parsed if 'selected_results' in parsed and 'reasoning' in parsed else None
def validate_page_selection_response(self, parsed_response: Dict[str, Union[List[int], str]], num_results: int) -> bool:
if len(parsed_response['selected_results']) != 2:
return False
if any(num < 1 or num > num_results for num in parsed_response['selected_results']):
return False
return True
def format_results(self, results: List[Dict]) -> str:
formatted_results = []
for result in results:
formatted_result = f"{result['number']}. Title: {result.get('title', 'N/A')}\n"
formatted_result += f" Snippet: {result.get('body', 'N/A')[:200]}...\n"
formatted_result += f" URL: {result.get('href', 'N/A')}\n"
formatted_results.append(formatted_result)
return "\n".join(formatted_results)
def scrape_content(self, urls: List[str]) -> Dict[str, str]:
scraped_content = {}
blocked_urls = []
for url in urls:
robots_allowed = can_fetch(url)
if robots_allowed:
content = get_web_content([url])
if content:
scraped_content.update(content)
print(Fore.YELLOW + f"Successfully scraped: {url}" + Style.RESET_ALL)
logger.info(f"Successfully scraped: {url}")
else:
print(Fore.RED + f"Robots.txt disallows scraping of {url}" + Style.RESET_ALL)
logger.warning(f"Robots.txt disallows scraping of {url}")
else:
blocked_urls.append(url)
print(Fore.RED + f"Warning: Robots.txt disallows scraping of {url}" + Style.RESET_ALL)
logger.warning(f"Robots.txt disallows scraping of {url}")
print(Fore.CYAN + f"Scraped content received for {len(scraped_content)} URLs" + Style.RESET_ALL)
logger.info(f"Scraped content received for {len(scraped_content)} URLs")
if blocked_urls:
print(Fore.RED + f"Warning: {len(blocked_urls)} URL(s) were not scraped due to robots.txt restrictions." + Style.RESET_ALL)
logger.warning(f"{len(blocked_urls)} URL(s) were not scraped due to robots.txt restrictions: {', '.join(blocked_urls)}")
return scraped_content
def display_scraped_content(self, scraped_content: Dict[str, str]):
print(f"\n{Fore.CYAN}Scraped Content:{Style.RESET_ALL}")
for url, content in scraped_content.items():
print(f"{Fore.GREEN}URL: {url}{Style.RESET_ALL}")
print(f"Content: {content[:4000]}...\n")
def generate_final_answer(self, user_query: str, scraped_content: Dict[str, str]) -> str:
user_query_short = user_query[:200]
prompt = f"""
You are an AI assistant. Provide a comprehensive and detailed answer to the following question using ONLY the information provided in the scraped content. Do not include any references or mention any sources. Answer directly and thoroughly.
Question: "{user_query_short}"
Scraped Content:
{self.format_scraped_content(scraped_content)}
Important Instructions:
1. Do not use phrases like "Based on the absence of selected results" or similar.
2. If the scraped content does not contain enough information to answer the question, say so explicitly and explain what information is missing.
3. Provide as much relevant detail as possible from the scraped content.
Answer:
"""
max_retries = 3
for attempt in range(max_retries):
with OutputRedirector() as output:
response_text = self.llm.generate(prompt, max_tokens=1024, stop=None)
llm_output = output.getvalue()
logger.info(f"LLM Output in generate_final_answer:\n{llm_output}")
if response_text:
logger.info(f"LLM Response:\n{response_text}")
return response_text
error_message = "I apologize, but I couldn't generate a satisfactory answer based on the available information."
logger.warning(f"Failed to generate a response after {max_retries} attempts. Returning error message.")
return error_message
def format_scraped_content(self, scraped_content: Dict[str, str]) -> str:
formatted_content = []
for url, content in scraped_content.items():
content = re.sub(r'\s+', ' ', content)
formatted_content.append(f"Content from {url}:\n{content}\n")
return "\n".join(formatted_content)
def synthesize_final_answer(self, user_query: str) -> str:
prompt = f"""
After multiple search attempts, we couldn't find a fully satisfactory answer to the user's question: "{user_query}"
Please provide the best possible answer you can, acknowledging any limitations or uncertainties.
If appropriate, suggest ways the user might refine their question or where they might find more information.
Respond in a clear, concise, and informative manner.
"""
try:
with OutputRedirector() as output:
response_text = self.llm.generate(prompt, max_tokens=self.llm_config.get('max_tokens', 1024), stop=self.llm_config.get('stop', None))
llm_output = output.getvalue()
logger.info(f"LLM Output in synthesize_final_answer:\n{llm_output}")
if response_text:
return response_text.strip()
except Exception as e:
logger.error(f"Error in synthesize_final_answer: {str(e)}", exc_info=True)
return "I apologize, but after multiple attempts, I wasn't able to find a satisfactory answer to your question. Please try rephrasing your question or breaking it down into smaller, more specific queries."
# End of EnhancedSelfImprovingSearch class
|