tool.py

from smolagents import Tool
from typing import Any, Optional

class SimpleTool(Tool):
    name = "web_analyzer"
    description = "Advanced web content analyzer with AI-powered analysis."
    inputs = {"url":{"type":"string","description":"The webpage URL to analyze."},"mode":{"type":"string","nullable":True,"description":"Analysis mode ('analyze', 'summarize', 'sentiment', 'topics')."}}
    output_type = "string"

    def forward(self, url: str, mode: str = "analyze") -> str:
        """Advanced web content analyzer with AI-powered analysis.

        Args:
            url: The webpage URL to analyze.
            mode: Analysis mode ('analyze', 'summarize', 'sentiment', 'topics'). 

        Returns:
            str: AI-enhanced analysis of web content.
        """
        import requests
        from bs4 import BeautifulSoup
        import re
        from transformers import pipeline

        try:
            # Setup headers
            headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64)'}

            # Fetch content
            response = requests.get(url, headers=headers, timeout=10)
            response.raise_for_status()

            # Parse content
            soup = BeautifulSoup(response.text, 'html.parser')
            for tag in soup(['script', 'style', 'meta']):
                tag.decompose()

            # Get cleaned text
            title = soup.title.string if soup.title else "No title found"
            title = re.sub(r'\s+', ' ', title).strip()
            text_content = re.sub(r'\s+', ' ', soup.get_text()).strip()

            # Initialize ML models based on mode
            if mode == "analyze":
                # Basic analysis with summary
                summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
                classifier = pipeline("text-classification", 
                                    model="nlptown/bert-base-multilingual-uncased-sentiment")

                # Get summary and sentiment
                summary = summarizer(text_content[:1024], max_length=100, min_length=30)[0]['summary_text']
                sentiment = classifier(text_content[:512])[0]
                sent_score = int(sentiment['label'][0])
                sent_text = ["Very Negative", "Negative", "Neutral", "Positive", "Very Positive"][sent_score-1]

                # Format output
                return f"""📊 Content Analysis

    Title: {title}
    Length: {len(text_content)} characters

    📝 AI Summary:
    {summary}

    😊 Overall Sentiment: {sent_text} ({sent_score}/5)"""

            elif mode == "summarize":
                summarizer = pipeline("summarization", model="facebook/bart-large-cnn")

                # Process in chunks
                chunk_size = 1024
                summaries = []

                for i in range(0, min(len(text_content), 3072), chunk_size):
                    chunk = text_content[i:i+chunk_size]
                    if len(chunk) > 100:
                        summary = summarizer(chunk, max_length=100, min_length=30)[0]['summary_text']
                        summaries.append(summary)

                return f"""📝 Multi-Section Summary

    Title: {title}

    {' '.join(summaries)}"""

            elif mode == "sentiment":
                classifier = pipeline("text-classification", 
                                   model="nlptown/bert-base-multilingual-uncased-sentiment")

                # Analyze paragraphs
                paragraphs = soup.find_all('p')
                sentiments = ""
                count = 0

                for p in paragraphs:
                    text = p.text.strip()
                    if len(text) > 50:
                        result = classifier(text[:512])[0]
                        score = int(result['label'][0])
                        mood = ["Very Negative", "Negative", "Neutral", "Positive", "Very Positive"][score-1]
                        sentiments += f"\nSection {count + 1}: {mood} ({score}/5 stars)"
                        count += 1
                        if count >= 5:
                            break

                return f"""😊 Sentiment Analysis

    Title: {title}
    {sentiments}"""

            elif mode == "topics":
                classifier = pipeline("zero-shot-classification",
                                   model="facebook/bart-large-mnli")

                topics = [
                    "Technology", "AI/ML", "Business", "Science",
                    "Innovation", "Research", "Industry News"
                ]

                results = classifier(text_content[:512], topics)

                topic_analysis = "Detected Topics:\n"
                for topic, score in zip(results['labels'], results['scores']):
                    if score > 0.1:
                        topic_analysis += f"- {topic}: {score*100:.1f}% confidence\n"

                return f"""🎯 Topic Classification

    Title: {title}

    {topic_analysis}"""

            else:
                return f"Error: Unknown mode '{mode}'"

        except Exception as e:
            return f"Error processing webpage: {str(e)}"