Spaces:
Paused
Paused
File size: 44,661 Bytes
0f5b528 219c095 0f5b528 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 | import os
import json
import time
import feedparser
import requests
from bs4 import BeautifulSoup
import re
from datetime import datetime, timedelta
import gtts
from googletrans import Translator
import urllib.parse
from deep_translator import GoogleTranslator
from dotenv import load_dotenv
# LangChain imports
from langchain_google_genai import GoogleGenerativeAI
from langchain_chroma import Chroma
from langchain_huggingface import HuggingFaceEmbeddings
from langchain.schema import Document
from langchain_core.prompts import ChatPromptTemplate
from langchain.memory import ConversationSummaryBufferMemory
from langchain.tools import Tool
from langchain.agents import AgentExecutor, create_react_agent
from langchain_community.utilities import DuckDuckGoSearchAPIWrapper
# Load environment variables
load_dotenv()
class NewsAgent:
def __init__(self):
print("π Initializing News Agent...")
self.setup_llm()
self.setup_embeddings()
self.setup_vector_store()
self.test_vector_db() # Test the vector DB
self.delete_old_news() # Delete old news on startup
self.setup_memory()
self.setup_search_tools()
self.setup_tools()
self.setup_agent()
self.locations = set() # Track locations we've already fetched
print("β
News Agent initialized and ready!")
def setup_llm(self):
"""Initialize the Gemini model."""
try:
api_key = os.getenv("GOOGLE_API_KEY")
if not api_key:
raise ValueError("GOOGLE_API_KEY environment variable not set")
self.llm = GoogleGenerativeAI(
model="gemini-1.5-flash",
google_api_key=api_key,
temperature=0.2,
top_p=0.8,
max_output_tokens=2048
)
print("β
Gemini 1.5 Flash model initialized")
except Exception as e:
print(f"β Error initializing Gemini model: {e}")
raise
def setup_embeddings(self):
"""Initialize the embedding model."""
try:
self.embedding_model = HuggingFaceEmbeddings(
model_name="sentence-transformers/all-MiniLM-L6-v2",
cache_folder="/app/cache"
)
print("β
Embedding model initialized")
except Exception as e:
print(f"β Error initializing embedding model: {e}")
raise
def setup_vector_store(self):
"""Initialize ChromaDB vector store."""
try:
self.vector_store = Chroma(
persist_directory="./chroma_db",
embedding_function=self.embedding_model
)
print("β
Vector store initialized")
except Exception as e:
print(f"β Error initializing vector store: {e}")
raise
def test_vector_db(self):
"""Test if the vector database is working properly."""
try:
# Check if DB is empty
db_info = self.vector_store.get()
print(f"Vector DB contains {len(db_info['ids'])} documents")
if len(db_info['ids']) > 0:
# Try a simple search
results = self.vector_store.similarity_search("test", k=1)
print(f"Test search returned {len(results)} results")
if results:
print(f"Sample document: {results[0].metadata['title']}")
return True
else:
print("Vector DB is empty")
return False
except Exception as e:
print(f"β Error testing vector DB: {e}")
return False
def is_recent_news_available(self, location, max_age_minutes=180):
"""Check if recent news for a location is available in the database."""
try:
now = datetime.now()
# Search for news related to the location
results = self.vector_store.similarity_search(location, k=10)
# Filter results to those within max_age_minutes
recent_news = []
for doc in results:
metadata = doc.metadata
if metadata.get('location', '').lower() == location.lower():
timestamp_str = metadata.get('timestamp')
if timestamp_str:
try:
timestamp = datetime.fromisoformat(timestamp_str)
if now - timestamp <= timedelta(minutes=max_age_minutes):
recent_news.append(doc)
except Exception:
# Ignore parsing errors
continue
print(f"Found {len(recent_news)} recent news items for {location} in database")
return recent_news
except Exception as e:
print(f"β Error checking recent news: {e}")
return []
def delete_old_news(self, max_age_minutes=60):
"""Delete news older than the specified age from the database."""
try:
now = datetime.now()
# Get all documents
all_docs = self.vector_store.get()
all_ids = all_docs['ids']
all_metadatas = all_docs['metadatas']
# Identify documents older than max_age_minutes
ids_to_delete = []
for doc_id, metadata in zip(all_ids, all_metadatas):
timestamp_str = metadata.get('timestamp') if metadata else None
if timestamp_str:
try:
timestamp = datetime.fromisoformat(timestamp_str)
if now - timestamp > timedelta(minutes=max_age_minutes):
ids_to_delete.append(doc_id)
except Exception:
# Ignore parsing errors
continue
# Delete old documents
if ids_to_delete:
self.vector_store.delete(ids=ids_to_delete)
print(f"β
Deleted {len(ids_to_delete)} old news items from database")
return len(ids_to_delete)
except Exception as e:
print(f"β Error deleting old news: {e}")
return 0
def determine_news_count(self, user_request):
"""Determine how many news articles to fetch based on user request."""
# Check if user is asking for more news
more_patterns = ["more news", "additional news", "more articles", "show more", "get more"]
if any(pattern in user_request.lower() for pattern in more_patterns):
# Check if user specified a number
number_match = re.search(r'(\d+)\s+(more|additional|extra)', user_request.lower())
if number_match:
try:
count = int(number_match.group(1))
# Cap at a reasonable maximum
return min(count, 20)
except ValueError:
pass
return 15 # Return more news if requested without specific number
else:
return 5 # Default number of news
def setup_memory(self):
"""Initialize conversation memory."""
try:
self.memory = ConversationSummaryBufferMemory(
llm=self.llm,
max_token_limit=4000, # Increased token limit for better context retention
return_messages=True,
memory_key="chat_history",
input_key="input", # Explicitly define input key
output_key="output" # Explicitly define output key
)
print("β
Conversation memory initialized")
except Exception as e:
print(f"β Error initializing memory: {e}")
raise
def setup_search_tools(self):
"""Set up search tools."""
try:
# Setup DuckDuckGo search
self.ddg_wrapper = DuckDuckGoSearchAPIWrapper(
time="d", # Search for content from the past day
max_results=5
)
# Setup DuckDuckGo news search
self.ddg_news_wrapper = DuckDuckGoSearchAPIWrapper(
time="d", # Search for content from the past day
max_results=5
)
print("β
Search tools initialized")
except Exception as e:
print(f"β Error initializing search tools: {e}")
raise
def setup_tools(self):
"""Set up tools for the agent."""
self.tools = [
Tool(
name="FetchNews",
func=self.fetch_city_news,
description="Fetches the latest news for a specific city or location. Input should be the name of the city or 'city, number' to specify how many articles to fetch."
),
Tool(
name="SearchNewsArticle",
func=self.search_news_article,
description="Searches for news articles on a specific topic or title and returns summaries. Input should be the topic or title to search for."
),
Tool(
name="GetMoreInfoOnNews",
func=self.get_more_info_on_news,
description="Gets more detailed information about a specific news story. Input should be the news title or topic you want more information about."
),
Tool(
name="GetArticleContent",
func=self.get_article_content,
description="Gets the content of a news article from a URL. Input should be the URL of the article."
),
Tool(
name="SummarizeText",
func=self.summarize_text,
description="Summarizes a text. Input should be the text to summarize."
),
Tool(
name="TextToSpeech",
func=self.text_to_speech,
description="Converts text to speech in a specified language. Input should be a JSON string with 'text' and 'lang' keys."
),
Tool(
name="TranslateText",
func=self.translate_text,
description="Translates text to a specified language. Input should be a JSON string with 'text' and 'lang' keys."
),
Tool(
name="SearchNewsInDB",
func=self.search_news_in_db,
description="Searches for news in the database. Input should be the search query."
),
Tool(
name="GetRecentNewsFromDB",
func=self.get_recent_news_from_db,
description="Gets recent news for a location from the database. Input should be the location name."
)
]
print("β
Agent tools initialized")
def setup_agent(self):
"""Set up the LangChain agent."""
prompt = ChatPromptTemplate.from_messages([
("system", """You are a helpful AI assistant that specializes in providing location-specific news Developed by GFG-KIIT AI/ML Team.
You can fetch news, search for articles, get more information on specific news stories, summarize text, translate content, and convert text to speech.
Always try to understand what location the user is asking about and provide relevant news.
If you're not sure about a location, ask for clarification.
IMPORTANT: Maintain conversation context. When the user asks follow-up questions about previously mentioned news articles,
use your memory of the conversation to understand which article they're referring to. If they ask for more details about a
news story you've mentioned, use the GetMoreInfoOnNews tool with the appropriate title.
When providing news:
1. Always ensure you're providing the most recent news (from today if possible)
2. First check if recent news is available in the database before fetching from the web
3. If a user asks for more information about a specific news story, use the GetMoreInfoOnNews tool
4. Always include relevant links when providing detailed information about news
5. Summarize news articles in a concise and informative way
6. If a user asks for more news, provide additional articles (up to 15)
7. Remember which news articles you've already mentioned in the conversation
You have access to the following tools:
{tools}
Use the following format:
Question: the input question you must answer
Thought: you should always think about what to do
Action: the action to take, should be one of [{tool_names}]
Action Input: the input to the action
Observation: the result of the action
... (this Thought/Action/Action Input/Observation can repeat N times)
Thought: I now know the final answer
Final Answer: the final answer to the original input question
Chat History: {chat_history}
"""),
("human", "{input}"),
("ai", "{agent_scratchpad}")
])
self.agent = create_react_agent(
llm=self.llm,
tools=self.tools,
prompt=prompt
)
self.agent_executor = AgentExecutor(
agent=self.agent,
tools=self.tools,
memory=self.memory,
verbose=True,
handle_parsing_errors=True,
return_intermediate_steps=True # Return intermediate steps for better debugging
)
print("β
Agent executor initialized")
def get_recent_news_from_db(self, location):
"""Gets recent news for a location from the database."""
try:
recent_news = self.is_recent_news_available(location)
if not recent_news:
return f"No recent news found in database for {location}. Try fetching fresh news."
response = f"π° Recent News from {location} (from database):\n\n"
for i, doc in enumerate(recent_news, 1):
metadata = doc.metadata
response += f"{i}. {metadata.get('title', 'Unknown Title')}\n"
response += f" Source: {metadata.get('source', 'Unknown Source')}\n"
response += f" Published: {metadata.get('date', 'Unknown Date')}\n"
response += f" Link: {metadata.get('link', 'No Link Available')}\n"
# Extract summary from content
content = doc.page_content
summary_match = re.search(r"SUMMARY: (.*?)(?:CONTENT:|$)", content, re.DOTALL)
if summary_match:
summary = summary_match.group(1).strip()
response += f" Summary: {summary}\n"
response += "\n"
return response
except Exception as e:
print(f"β Error getting recent news from DB: {e}")
return f"Error retrieving recent news for {location} from database."
def search_news_article(self, query):
"""Search for news articles on a specific topic using DuckDuckGo News."""
try:
print(f"π Searching for news articles on: {query}")
# Parse input for number of results if provided
parts = query.split(',')
search_query = parts[0].strip()
max_results = 5
if len(parts) > 1:
try:
max_results = int(parts[1].strip())
max_results = min(max_results, 20) # Cap at 20 results
except ValueError:
pass
# Use DuckDuckGo search with news-specific query
search_results = self.ddg_news_wrapper.results(f"{search_query} news", max_results=max_results)
if not search_results:
return f"No news articles found for: {search_query}"
# Process search results
articles = []
for i, result in enumerate(search_results[:max_results]):
title = result.get("title", "No title")
link = result.get("link", "No link")
snippet = result.get("snippet", "No snippet")
published_date = result.get("published", datetime.now().strftime("%a, %d %b %Y %H:%M:%S"))
source = result.get("source", "Unknown source")
# Create article object
article = {
"title": title,
"source": source,
"link": link,
"published": published_date,
"snippet": snippet,
"query": search_query
}
articles.append(article)
# Store in vector database for RAG
self.store_article_in_db(article)
# Format response
response = f"π° Latest News Articles on '{search_query}':\n\n"
for i, article in enumerate(articles, 1):
response += f"{i}. {article['title']}\n"
response += f" Source: {article['source']}\n"
response += f" Published: {article['published']}\n"
response += f" Link: {article['link']}\n"
response += f" Summary: {article['snippet']}\n\n"
return response
except Exception as e:
print(f"β Error searching for news articles: {e}")
return f"Error searching for news articles on '{query}': {str(e)}"
def get_more_info_on_news(self, news_title):
"""Gets more detailed information about a specific news story."""
try:
print(f"π Getting more information on: {news_title}")
# First, search for the news in our database
db_results = self.search_news_in_db(news_title, k=1)
# If we found something relevant in the database
if "No relevant news found" not in db_results:
# Extract the URL from the database results
url_match = re.search(r"Link: (https?://[^\s]+)", db_results)
if url_match:
article_url = url_match.group(1)
# Get the full content of the article
content = self.get_article_content(article_url)
# Summarize the content
summary = self.summarize_text(content)
return f"π° More Information on '{news_title}':\n\n{summary}\n\nSource: {article_url}"
# If we didn't find anything in the database or couldn't extract the URL,
# search for the news using DuckDuckGo
search_results = self.ddg_wrapper.results(f"{news_title} latest news", max_results=5)
if not search_results:
return f"Could not find more information on: {news_title}"
# Get the first result
result = search_results[0]
article_url = result.get("link")
if not article_url:
return f"Could not find a relevant article for: {news_title}"
# Get the content of the article
content = self.get_article_content(article_url)
# Summarize the content
summary = self.summarize_text(content)
# Store in vector database for future reference
self.store_article_in_db({
"title": news_title,
"link": article_url,
"content": content,
"summary": summary,
"source": result.get("source", "Unknown source"),
"published": datetime.now().strftime("%a, %d %b %Y")
})
return f"π° More Information on '{news_title}':\n\n{summary}\n\nSource: {article_url}"
except Exception as e:
print(f"β Error getting more information: {e}")
return f"Error getting more information on '{news_title}': {str(e)}"
def get_article_content(self, url):
"""Extract content from a news article URL."""
try:
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36'
}
# Check if URL is valid
if not url.startswith('http'):
return "Invalid URL. Please provide a URL starting with http:// or https://"
# Send request
response = requests.get(url, headers=headers, timeout=10)
response.raise_for_status() # Raise exception for 4XX/5XX status codes
# Parse HTML
soup = BeautifulSoup(response.text, 'html.parser')
# Remove script, style, and nav elements
for element in soup(['script', 'style', 'nav', 'header', 'footer', 'aside']):
element.decompose()
# Try to find the main content
main_content = None
# Look for article tag
article = soup.find('article')
if article:
main_content = article
# Look for main tag if article not found
if not main_content:
main_tag = soup.find('main')
if main_tag:
main_content = main_tag
# Look for div with content-related class names
if not main_content:
content_div = soup.find('div', class_=lambda c: c and any(x in c.lower() for x in ['content', 'article', 'story', 'entry', 'post']))
if content_div:
main_content = content_div
# Extract text from main content or fallback to body
if main_content:
paragraphs = main_content.find_all('p')
else:
paragraphs = soup.find_all('p')
# Join paragraphs
content = '\n\n'.join([p.get_text().strip() for p in paragraphs if len(p.get_text().strip()) > 40])
# If content is too short, try a different approach
if len(content) < 200:
# Get all text from body
body = soup.find('body')
if body:
content = body.get_text(separator='\n')
# Clean up content
lines = [line.strip() for line in content.split('\n') if line.strip()]
content = '\n'.join(lines)
# If still no content, return error
if not content or len(content) < 100:
return "Could not extract meaningful content from the article."
# Truncate if too long
if len(content) > 8000:
content = content[:8000] + "...[content truncated]"
return content
except requests.exceptions.RequestException as e:
return f"Error fetching article: {str(e)}"
except Exception as e:
return f"Error extracting content: {str(e)}"
def summarize_text(self, text):
"""Summarize text using the LLM."""
try:
if not text or len(text) < 100:
return "Text is too short to summarize."
# Truncate text if it's too long
if len(text) > 10000:
text = text[:10000] + "...[content truncated]"
prompt = f"""
Summarize the following news article in a concise way (3-5 sentences), highlighting the key points:
{text}
Summary:
"""
response = self.llm.invoke(prompt)
return response
except Exception as e:
print(f"β Error summarizing text: {e}")
return "Could not generate summary due to an error."
def fetch_city_news(self, city_input, max_articles=5):
"""Fetch news for a specific city using Google News RSS first, then enhance with search."""
# Parse input for city and optional count
parts = city_input.split(',')
city = parts[0].strip()
if len(parts) > 1:
try:
max_articles = int(parts[1].strip())
max_articles = min(max_articles, 20) # Cap at 20 articles
except ValueError:
pass
print(f"π Fetching {max_articles} news articles for: {city}")
# Check if we have recent news in the database
recent_news = self.is_recent_news_available(city)
if recent_news and len(recent_news) >= max_articles:
print(f"β
Found {len(recent_news)} recent news items in database for {city}")
response = f"π° Latest News from {city} (from database):\n\n"
for i, doc in enumerate(recent_news[:max_articles], 1):
metadata = doc.metadata
response += f"{i}. {metadata.get('title', 'Unknown Title')}\n"
response += f" Source: {metadata.get('source', 'Unknown Source')}\n"
response += f" Published: {metadata.get('date', 'Unknown Date')}\n"
response += f" Link: {metadata.get('link', 'No Link Available')}\n"
# Extract summary from content
content = doc.page_content
summary_match = re.search(r"SUMMARY: (.*?)(?:CONTENT:|$)", content, re.DOTALL)
if summary_match:
summary = summary_match.group(1).strip()
response += f" Summary: {summary}\n"
response += "\n"
return response
# Clean the city name to avoid URL issues
clean_city = city.strip().replace("\n", "").replace("\r", "")
encoded_city = urllib.parse.quote(clean_city)
try:
# First get news from Google News RSS
rss_url = f"https://news.google.com/rss/search?q={encoded_city}+when:1d&hl=en-US&gl=US&ceid=US:en"
feed = feedparser.parse(rss_url)
if not feed.entries:
return f"No news found for {city}"
# Process articles from RSS feed
articles = []
for entry in feed.entries[:max_articles]:
# Extract title and source
title_parts = entry.title.split(" - ")
title = title_parts[0].strip() if len(title_parts) > 1 else entry.title.strip()
source = title_parts[-1].strip() if len(title_parts) > 1 else "Unknown"
# Get the article link
google_news_link = entry.link
# Extract publication date
published_date = entry.get("published", datetime.now().strftime("%a, %d %b %Y"))
print(f"π° Found news: {title}")
print(f"π Searching for more details about: {title}")
# Now search for more details about this specific news
try:
search_results = self.ddg_wrapper.results(f"{title} {city} news", max_results=3)
if search_results:
# Get the first result
result = search_results[0]
article_url = result.get("link")
# Get the content of the article
content = self.get_article_content(article_url)
# Summarize the content
summary = self.summarize_text(content)
else:
article_url = google_news_link
content = ""
summary = "No additional details available."
except Exception as e:
print(f"β Error getting more details: {e}")
article_url = google_news_link
content = ""
summary = "Could not retrieve additional details due to an error."
# Create article object
article = {
"title": title,
"source": source,
"link": article_url,
"published": published_date,
"location": city,
"summary": summary,
"content": content if 'content' in locals() else ""
}
articles.append(article)
# Store in vector database for RAG
self.store_article_in_db(article)
# Add location to tracked locations
self.locations.add(city.lower())
# Format response
response = f"π° Latest News from {city}:\n\n"
for i, article in enumerate(articles, 1):
response += f"{i}. {article['title']}\n"
response += f" Source: {article['source']}\n"
response += f" Published: {article['published']}\n"
response += f" Link: {article['link']}\n"
response += f" Summary: {article['summary']}\n\n"
return response
except Exception as e:
print(f"β Error fetching news: {e}")
return f"Error fetching news for {city}: {str(e)}"
def store_article_in_db(self, article):
"""Store an article in the vector database."""
try:
# Create document text
doc_text = f"""
TITLE: {article.get('title', 'Unknown Title')}
SOURCE: {article.get('source', 'Unknown Source')}
PUBLISHED: {article.get('published', datetime.now().strftime('%a, %d %b %Y'))}
LOCATION: {article.get('location', 'Unknown Location')}
LINK: {article.get('link', 'No Link Available')}
SUMMARY: {article.get('summary', article.get('snippet', 'No Summary Available'))}
CONTENT: {article.get('content', 'No Content Available')}
"""
# Add metadata
metadata = {
"title": article.get('title', 'Unknown Title'),
"source": article.get('source', 'Unknown Source'),
"location": article.get('location', 'Unknown Location'),
"date": article.get('published', datetime.now().strftime('%a, %d %b %Y')),
"link": article.get('link', 'No Link Available'),
"type": "news",
"timestamp": datetime.now().isoformat() # Add timestamp for recency filtering
}
# Create document
document = Document(page_content=doc_text, metadata=metadata)
# Add to vector store - this automatically persists the data
self.vector_store.add_documents([document])
# Verify storage
print(f"β
Stored article in vector database: {article.get('title', 'Unknown Title')}")
try:
db_info = self.vector_store.get()
print(f" Current DB size: {len(db_info['ids'])} documents")
except:
print(" Could not verify DB size")
return True
except Exception as e:
print(f"β Error storing article: {e}")
print(f"Article data: {article}")
return False
def text_to_speech(self, input_json):
"""Convert text to speech in the specified language."""
try:
# Parse input JSON
try:
data = json.loads(input_json)
text = data.get("text", "")
lang = data.get("lang", "en")
except json.JSONDecodeError:
# If not valid JSON, assume it's just text
text = input_json
lang = "en"
if not text:
return "No text provided for speech conversion."
# Get supported languages
supported_languages = gtts.lang.tts_langs()
if lang not in supported_languages:
return f"Language '{lang}' is not supported for text-to-speech."
# Generate speech
output_file = f"speech_{int(time.time())}.mp3"
tts = gtts.gTTS(text=text, lang=lang, slow=False)
tts.save(output_file)
return f"Successfully converted text to speech in {supported_languages[lang]}."
except Exception as e:
print(f"β Error in text-to-speech: {e}")
return f"Error in text-to-speech: {str(e)}"
def translate_text(self, input_json):
"""Translate text to the specified language."""
try:
# Parse input JSON
try:
data = json.loads(input_json)
text = data.get("text", "")
lang = data.get("lang", "en")
except json.JSONDecodeError:
# If not valid JSON, assume format is "text|lang"
parts = input_json.split("|")
text = parts[0]
lang = parts[1] if len(parts) > 1 else "en"
if not text:
return "No text provided for translation."
# Translate text using deep-translator
translator = GoogleTranslator(source='auto', target=lang)
translated_text = translator.translate(text)
return f"Translated text: {translated_text}"
except Exception as e:
print(f"β Error in translation: {e}")
return f"Error in translation: {str(e)}"
def search_news_in_db(self, query, k=3):
"""Search for news in the vector database with recency filtering."""
try:
# Get current date
current_date = datetime.now()
# First, perform the similarity search
results = self.vector_store.similarity_search(query, k=k*2) # Get more results than needed for filtering
if not results:
return "No relevant news found in the database."
# Filter for recent news (prioritize news from the last 24 hours)
recent_results = []
older_results = []
for doc in results:
metadata = doc.metadata
timestamp_str = metadata.get("timestamp")
if timestamp_str:
try:
timestamp = datetime.fromisoformat(timestamp_str)
# If news is from the last 24 hours
if current_date - timestamp <= timedelta(days=1):
recent_results.append(doc)
else:
older_results.append(doc)
except (ValueError, TypeError):
older_results.append(doc)
else:
older_results.append(doc)
# Combine recent and older results, prioritizing recent ones
filtered_results = recent_results + older_results
# Limit to the requested number of results
filtered_results = filtered_results[:k]
if not filtered_results:
return "No relevant news found in the database."
response = "π° Related News from Database:\n\n"
for i, doc in enumerate(filtered_results, 1):
metadata = doc.metadata
response += f"{i}. {metadata.get('title', 'Unknown Title')}\n"
response += f" Source: {metadata.get('source', 'Unknown Source')}\n"
response += f" Location: {metadata.get('location', 'Unknown Location')}\n"
response += f" Published: {metadata.get('date', 'Unknown Date')}\n"
response += f" Link: {metadata.get('link', 'No Link Available')}\n\n"
return response
except Exception as e:
print(f"β Error searching news in DB: {e}")
return "Error searching the news database."
def extract_locations(self, query):
"""Extract potential location names from the query."""
try:
prompt = f"""
Extract any city or country names from this text. Return ONLY the names separated by commas, or 'None' if no locations are found:
Text: {query}
"""
response = self.llm.invoke(prompt)
locations = [loc.strip() for loc in response.strip().split(',') if loc.strip().lower() != 'none']
return locations
except Exception:
# Fallback to simple keyword extraction
common_cities = ["new york", "london", "tokyo", "paris", "delhi", "mumbai", "kolkata", "bangalore", "bhubaneswar"]
found = []
for city in common_cities:
if city.lower() in query.lower():
found.append(city)
return found
def process_query(self, query):
"""Process a user query through the agent."""
# Clean up old news first
self.delete_old_news()
# Get conversation history to provide context
chat_history = self.get_conversation_context()
# Determine how many news to fetch
news_count = self.determine_news_count(query)
# Check if query contains a location
potential_locations = self.extract_locations(query)
# Check if user is asking for more details about a specific news
is_asking_for_details = any(pattern in query.lower() for pattern in
["more details", "tell me more about", "more information on",
"details on", "what about", "tell me about"])
# If asking for details about specific news, try to extract the news title from context
if is_asking_for_details and not any(word in query.lower() for word in ["news", "article"]):
# Try to extract news title from the query or recent conversation
news_title = self.extract_news_title_from_context(query, chat_history)
if news_title:
print(f"π Extracted news title from context: {news_title}")
# Append the extracted title to the query for clarity
query = f"{query} about '{news_title}'"
# For location-based queries
for location in potential_locations:
# Check if we have recent news in the database
recent_news = self.is_recent_news_available(location)
# If user wants more news or we don't have recent news, fetch from web
if not recent_news or "more" in query.lower():
if location.lower() not in [loc.lower() for loc in self.locations]:
print(f"π Detected new location: {location}. Fetching news...")
self.fetch_city_news(f"{location}, {news_count}")
# Process through the agent with enhanced context
try:
chat_history = self.get_conversation_context()
response = self.agent_executor.invoke({
"input": query,
"chat_history": chat_history # This will be included in the system message
})
return response["output"]
except Exception as e:
print(f"β Error processing query: {e}")
return "I'm sorry, I encountered an error while processing your question. Please try again."
def get_conversation_context(self):
"""Get formatted conversation history for context."""
try:
# Get messages from memory
messages = self.memory.chat_memory.messages
if not messages:
return []
return messages
except Exception as e:
print(f"β Error retrieving conversation context: {e}")
return []
def extract_news_title_from_context(self, query, chat_history):
"""Extract relevant news title from conversation context or query."""
try:
# First, check if there are any news titles in the recent AI messages
recent_ai_messages = [msg.content for msg in chat_history[-4:] if hasattr(msg, 'type') and msg.type == 'ai']
# Combine recent AI messages
context_text = " ".join(recent_ai_messages)
# Look for news titles in the format typically used in our responses
title_matches = re.findall(r'\d+\.\s+(.*?)\n', context_text)
if title_matches:
# Use the LLM to determine which title is most relevant to the query
titles_text = "\n".join([f"{i+1}. {title}" for i, title in enumerate(title_matches)])
prompt = f"""
Given the user query and the list of recently mentioned news titles, which title is the user most likely referring to?
Return ONLY the title, or "None" if none seem relevant.
User query: {query}
Recently mentioned titles:
{titles_text}
"""
response = self.llm.invoke(prompt).strip()
if response and response.lower() != "none":
return response
# If we couldn't find a title from context, try to extract it from the query
# This is a fallback for explicit mentions
query_words = query.lower().split()
for i, word in enumerate(query_words):
if word in ["about", "regarding", "concerning", "on"]:
if i+1 < len(query_words):
potential_title = " ".join(query_words[i+1:])
# Remove quotes if present
potential_title = potential_title.strip('"\'')
if len(potential_title) > 3: # Minimum length check
return potential_title
return None
except Exception as e:
print(f"β Error extracting news title from context: {e}")
return None
def main():
print("=" * 50)
print("π Location-Specific News Agent")
print("=" * 50)
print("Initializing system...")
agent = NewsAgent()
print("\nChat with the news agent! Type 'exit' to quit.")
print("Example: 'What's happening in Delhi today?'")
while True:
user_input = input("\nYou: ").strip()
if user_input.lower() in ['exit', 'quit', 'bye']:
print("Thank you for using the news agent. Goodbye!")
break
if not user_input:
continue
response = agent.process_query(user_input)
print(f"\nAI: {response}")
if __name__ == "__main__":
main()
|