quantalogic 0.35.0__py3-none-any.whl → 0.50.0__py3-none-any.whl

quantalogic/tools/google_packages/google_news_tool.py

@@ -0,0 +1,480 @@
+"""Advanced tool for retrieving and analyzing news articles from Google News.
+
+This tool provides a sophisticated interface to fetch, analyze, and format news articles
+from Google News using multiple sources and advanced filtering capabilities.
+"""
+
+import asyncio
+from typing import Any, Dict, List
+
+import aiohttp
+import html2text
+from bs4 import BeautifulSoup
+from gnews import GNews
+from loguru import logger
+
+from quantalogic.event_emitter import EventEmitter
+from quantalogic.tools.llm_tool import LLMTool
+from quantalogic.tools.tool import Tool, ToolArgument
+
+
+class NewsArticle:
+    """Represents a news article with enhanced metadata and analysis."""
+    
+    def __init__(self, title: str, url: str, source: str, date: str):
+        self.title = title
+        self.url = url
+        self.source = source
+        self.date = date
+        self.description = ""
+        self.full_text = ""
+        self.keywords = []
+        self.sentiment = {}
+        self.summary = ""
+        
+    async def enrich(self, session: aiohttp.ClientSession):
+        """Enrich article with additional data and analysis."""
+        try:
+            # Fetch article content
+            async with session.get(self.url) as response:
+                if response.status == 200:
+                    html_content = await response.text()
+                    
+                    # Convert HTML to text
+                    h = html2text.HTML2Text()
+                    h.ignore_links = True
+                    h.ignore_images = True
+                    self.full_text = h.handle(html_content)
+                    
+                    # Extract main content using BeautifulSoup
+                    soup = BeautifulSoup(html_content, 'html.parser')
+                    
+                    # Remove unwanted elements
+                    for tag in soup(['script', 'style', 'nav', 'header', 'footer', 'aside']):
+                        tag.decompose()
+                    
+                    # Get main content
+                    main_content = soup.find('main') or soup.find('article') or soup.find('body')
+                    if main_content:
+                        self.full_text = main_content.get_text(strip=True)
+                    
+                    # Basic keyword extraction from title and content
+                    words = set(word.lower() for word in self.title.split() + self.full_text.split()
+                              if len(word) > 3)
+                    self.keywords = list(words)[:10]  # Top 10 keywords
+                    
+                    # Create a basic summary (first 2-3 sentences)
+                    sentences = [s.strip() for s in self.full_text.split('.') if s.strip()]
+                    self.summary = '. '.join(sentences[:3]) + '.'
+                    
+                    # Perform sentiment analysis
+                    # sia = SentimentIntensityAnalyzer()
+                    # self.sentiment = sia.polarity_scores(self.title + " " + self.summary)
+                    self.sentiment = {'pos': 0.0, 'neg': 0.0, 'neu': 1.0, 'compound': 0.0}
+            
+        except Exception as e:
+            logger.warning(f"Error enriching article {self.url}: {str(e)}")
+            # Set basic info even if enrichment fails
+            self.summary = self.title
+            self.keywords = [word.lower() for word in self.title.split() if len(word) > 3]
+            self.sentiment = {'pos': 0.0, 'neg': 0.0, 'neu': 1.0, 'compound': 0.0}
+
+
+class GoogleNewsTool(Tool):
+    """Advanced tool for retrieving and analyzing news articles from Google News.
+
+    Features:
+    - Multi-source news aggregation
+    - Sentiment analysis
+    - Keyword extraction
+    - Article summarization
+    - Advanced filtering
+    - Async processing for better performance
+    """
+
+    name: str = "google_news_tool"
+    description: str = (
+        "Advanced news retrieval and analysis tool with support for sentiment analysis, "
+        "keyword extraction, and article summarization."
+    )
+    arguments: List[ToolArgument] = [
+        ToolArgument(
+            name="query",
+            arg_type="string",
+            description="The news search query",
+            required=True,
+            example="artificial intelligence developments",
+        ),
+        ToolArgument(
+            name="language",
+            arg_type="string",
+            description="Language code (e.g., 'en' for English)",
+            required=False,
+            default="en",
+            example="en",
+        ),
+        ToolArgument(
+            name="period",
+            arg_type="string",
+            description="Time period (1h, 1d, 7d, 1m)",
+            required=False,
+            default="1d",
+            example="7d",
+        ),
+        ToolArgument(
+            name="max_results",
+            arg_type="int",
+            description="Maximum number of results (1-100)",
+            required=False,
+            default="25",
+            example="20",
+        ),
+        ToolArgument(
+            name="country",
+            arg_type="string",
+            description="Country code for news sources",
+            required=False,
+            default="US",
+            example="GB",
+        ),
+        ToolArgument(
+            name="sort_by",
+            arg_type="string",
+            description="Sort by relevance or date",
+            required=False,
+            default="relevance",
+            example="date",
+        ),
+        ToolArgument(
+            name="analyze",
+            arg_type="boolean",
+            description="Perform detailed analysis of articles",
+            required=False,
+            default="True",
+            example="True",
+        ),
+    ]
+
+    def __init__(
+        self,
+        model_name: str | None = None,
+        on_token: Any | None = None,
+        event_emitter: EventEmitter | None = None,
+    ):
+        """Initialize the GoogleNewsTool.
+        
+        Args:
+            model_name (str | None): Name of the LLM model to use for summarization
+            on_token (Any | None): Token callback for streaming
+            event_emitter (EventEmitter | None): Event emitter for the tool
+        """
+        super().__init__()
+        self.model_name = model_name
+        self.on_token = on_token
+        self.event_emitter = event_emitter
+        if model_name:
+            self.llm_tool = LLMTool(
+                model_name=model_name,
+                on_token=on_token,
+                event_emitter=event_emitter,
+            )
+
+    def _summarize_article(self, article: Dict[str, Any]) -> str:
+        """Summarize a news article using LLM.
+        
+        Args:
+            article (Dict[str, Any]): Article data including title and description
+            
+        Returns:
+            str: Summarized article content
+        """
+        if not hasattr(self, 'llm_tool'):
+            return article.get('description', '')
+
+        prompt = f"""
+        Summarize this news article concisely and professionally:
+        
+        Title: {article.get('title', '')}
+        Description: {article.get('description', '')}
+        
+        Provide a 2-3 sentence summary that captures the key points.
+        """
+
+        try:
+            summary = self.llm_tool.execute(
+                system_prompt="You are a professional news summarizer. Create clear, accurate, and concise summaries.",
+                prompt=prompt,
+                temperature="0.3"
+            )
+            return summary
+        except Exception as e:
+            logger.error(f"Error summarizing article: {e}")
+            return article.get('description', '')
+
+    def _format_html_output(self, articles: List[Dict[str, Any]], query: str) -> str:
+        """Format articles as HTML with a modern, clean design."""
+        css_styles = """
+            body {
+                font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
+                line-height: 1.6;
+                max-width: 1200px;
+                margin: 0 auto;
+                padding: 20px;
+                background-color: #f5f5f5;
+            }
+            .header {
+                background-color: #2c3e50;
+                color: white;
+                padding: 20px;
+                border-radius: 8px;
+                margin-bottom: 20px;
+            }
+            .article {
+                background-color: white;
+                padding: 20px;
+                margin-bottom: 20px;
+                border-radius: 8px;
+                box-shadow: 0 2px 4px rgba(0,0,0,0.1);
+            }
+            .article h2 {
+                color: #2c3e50;
+                margin-top: 0;
+            }
+            .article-meta {
+                color: #666;
+                font-size: 0.9em;
+                margin-bottom: 10px;
+            }
+            .summary {
+                border-left: 4px solid #2c3e50;
+                padding-left: 15px;
+                margin: 15px 0;
+            }
+            .source-link {
+                display: inline-block;
+                margin-top: 10px;
+                color: #3498db;
+                text-decoration: none;
+            }
+            .source-link:hover {
+                text-decoration: underline;
+            }
+        """
+
+        articles_html = []
+        for article in articles:
+            article_html = f"""
+                <div class="article">
+                    <h2>{article.get('title', 'No Title')}</h2>
+                    <div class="article-meta">
+                        <span>Source: {article.get('source', {}).get('title', 'Unknown')}</span>
+                        <span> • </span>
+                        <span>Published: {article.get('published_date', 'Unknown date')}</span>
+                    </div>
+                    <div class="summary">
+                        {article.get('summary', 'No summary available')}
+                    </div>
+                    <a href="{article.get('link', '#')}" class="source-link" target="_blank">Read full article →</a>
+                </div>
+            """
+            articles_html.append(article_html)
+
+        html_content = f"""
+            <!DOCTYPE html>
+            <html>
+            <head>
+                <meta charset="UTF-8">
+                <meta name="viewport" content="width=device-width, initial-scale=1.0">
+                <style>
+                    {css_styles}
+                </style>
+            </head>
+            <body>
+                <div class="header">
+                    <h1>News Results for: {query}</h1>
+                    <p>Found {len(articles)} articles</p>
+                </div>
+                {''.join(articles_html)}
+            </body>
+            </html>
+        """
+        
+        return html_content.strip()
+
+    def _fetch_article_data(self, articles: List[NewsArticle]) -> List[NewsArticle]:
+        """Fetch detailed data for multiple articles."""
+        try:
+            loop = asyncio.get_event_loop()
+        except RuntimeError:
+            loop = asyncio.new_event_loop()
+            asyncio.set_event_loop(loop)
+
+        async def fetch_all():
+            tasks = []
+            async with aiohttp.ClientSession() as session:
+                for article in articles:
+                    tasks.append(article.enrich(session))
+                await asyncio.gather(*tasks)
+            return articles
+
+        return loop.run_until_complete(fetch_all())
+
+    def _format_results(self, articles: List[NewsArticle], analyze: bool) -> str:
+        """Format news results with optional analysis data."""
+        results = ["=== Advanced Google News Results ===\n"]
+        
+        for i, article in enumerate(articles, 1):
+            results.extend([
+                f"{i}. {article.title}",
+                f"   Source: {article.source} | Date: {article.date}",
+                f"   URL: {article.url}",
+                ""
+            ])
+            
+            if analyze and article.summary:
+                results.extend([
+                    "   Summary:",
+                    f"   {article.summary}",
+                    "",
+                    "   Key Topics:",
+                    f"   {', '.join(article.keywords[:5])}",
+                    "",
+                    "   Sentiment Analysis:",
+                    "   Overall tone: " + self._interpret_sentiment(article.sentiment),
+                    f"   - Positive: {article.sentiment.get('pos', 0)*100:.1f}% ({self._get_sentiment_level(article.sentiment.get('pos', 0))})",
+                    f"   - Negative: {article.sentiment.get('neg', 0)*100:.1f}% ({self._get_sentiment_level(article.sentiment.get('neg', 0))})",
+                    f"   - Neutral:  {article.sentiment.get('neu', 0)*100:.1f}% ({self._get_sentiment_level(article.sentiment.get('neu', 0))})",
+                    ""
+                ])
+            
+            results.append("")
+        
+        return "\n".join(results)
+
+    def _get_sentiment_level(self, score: float) -> str:
+        """Convert sentiment score to descriptive level."""
+        if score >= 0.6:
+            return "Very High"
+        elif score >= 0.4:
+            return "High"
+        elif score >= 0.2:
+            return "Moderate"
+        elif score > 0.1:
+            return "Low"
+        else:
+            return "Very Low"
+
+    def _interpret_sentiment(self, sentiment: Dict[str, float]) -> str:
+        """Interpret the overall sentiment of the text."""
+        compound = sentiment.get('compound', 0)
+        if compound >= 0.5:
+            return "Very Positive"
+        elif compound >= 0.1:
+            return "Slightly Positive"
+        elif compound <= -0.5:
+            return "Very Negative"
+        elif compound <= -0.1:
+            return "Slightly Negative"
+        else:
+            return "Neutral"
+
+    def execute(
+        self,
+        query: str,
+        language: str = "en",
+        period: str = "1m",
+        max_results: int = 30,
+        country: str = "US",
+        sort_by: str = "relevance",
+        analyze: bool = True,
+    ) -> str:
+        """Execute the Google News search with summarization and HTML formatting.
+
+        Args:
+            query (str): Search query
+            language (str, optional): Language code. Defaults to "en".
+            period (str, optional): Time period. Defaults to "1m".
+            max_results (int, optional): Maximum results. Defaults to 30.
+            country (str, optional): Country code. Defaults to "US".
+            sort_by (str, optional): Sort method. Defaults to "relevance".
+            analyze (bool, optional): Whether to analyze results. Defaults to True.
+
+        Returns:
+            str: HTML formatted news results with summaries
+        """
+        try:
+            # Input validation
+            if not query:
+                raise ValueError("Query cannot be empty")
+
+            # Configure GNews
+            google_news = GNews(
+                language=language,
+                country=country,
+                period=period,
+                max_results=max_results,
+            )
+
+            # Fetch news
+            logger.info(f"Fetching news for query: {query}")
+            articles = []
+            try:
+                raw_articles = google_news.get_news(query)
+                for article_data in raw_articles:
+                    articles.append(
+                        NewsArticle(
+                            title=article_data.get("title", ""),
+                            url=article_data.get("url", ""),
+                            source=article_data.get("publisher", {}).get("title", ""),
+                            date=article_data.get("published date", ""),
+                        )
+                    )
+            except Exception as e:
+                logger.error(f"Error fetching articles: {e}")
+                raise RuntimeError(f"Failed to fetch articles: {str(e)}")
+
+            # Enrich articles with additional data if requested
+            if analyze:
+                logger.info("Performing detailed analysis of articles...")
+                articles = self._fetch_article_data(articles)
+            
+            # Sort results if needed
+            if sort_by == "date":
+                articles.sort(key=lambda x: x.date if x.date else "", reverse=True)
+            
+            # Process and summarize each article
+            processed_articles = []
+            for article in articles:
+                article_copy = {
+                    'title': article.title,
+                    'link': article.url,
+                    'source': {'title': article.source},
+                    'published_date': article.date,
+                    'description': article.full_text if hasattr(article, 'full_text') else '',
+                }
+                article_copy['summary'] = self._summarize_article(article_copy)
+                processed_articles.append(article_copy)
+
+            # Format results as HTML
+            html_output = self._format_html_output(processed_articles, query)
+            return html_output
+
+        except Exception as e:
+            logger.error(f"Error in GoogleNewsTool: {e}")
+            raise Exception(f"Failed to fetch news: {str(e)}")
+
+
+if __name__ == "__main__":
+    # Example usage with analysis
+    tool = GoogleNewsTool()
+    try:
+        result = tool.execute(
+            query="XRP crypto coin",
+            language="en",
+            period="7d",
+            max_results=25,
+            analyze=True
+        )
+        print(result)
+    except Exception as e:
+        print(f"Error: {e}")