CodeAnalysis MCP Server

#!/usr/bin/env python3 """Feedback Sentiment Analyzer This script analyzes the sentiment of feedback text, categorizing feedback as positive, negative, or neutral, and identifying key themes. Maturity: beta Why: - Manual sentiment analysis is subjective and time-consuming - Automated analysis provides consistent categorization - Identifying themes helps prioritize feedback - Sentiment trends over time reveal the impact of changes """ import argparse import json import os import re from datetime import datetime from pathlib import Path import matplotlib.pyplot as plt import numpy as np from collections import Counter, defaultdict # Note: This script uses simple rule-based sentiment analysis # For production use, consider using a more sophisticated NLP library # such as NLTK, spaCy, or a cloud-based sentiment analysis API class SentimentAnalyzer: """Analyzes sentiment and themes in feedback text.""" def __init__(self): # Simple sentiment lexicons self.positive_words = set([ 'good', 'great', 'excellent', 'amazing', 'awesome', 'fantastic', 'wonderful', 'love', 'like', 'helpful', 'easy', 'useful', 'best', 'perfect', 'happy', 'satisfied', 'impressive', 'intuitive', 'fast', 'efficient', 'reliable', 'responsive', 'beautiful', 'clean', 'simple' ]) self.negative_words = set([ 'bad', 'poor', 'terrible', 'awful', 'horrible', 'worst', 'difficult', 'hard', 'confusing', 'confused', 'slow', 'buggy', 'broken', 'error', 'issue', 'problem', 'fail', 'crash', 'hate', 'dislike', 'disappointed', 'frustrating', 'annoying', 'useless', 'complicated', 'inconsistent', 'unreliable', 'ugly', 'expensive' ]) # Theme keywords self.themes = { 'performance': ['slow', 'fast', 'speed', 'performance', 'lag', 'responsive', 'loading'], 'usability': ['easy', 'intuitive', 'confusing', 'difficult', 'simple', 'complicated', 'user-friendly'], 'reliability': ['crash', 'bug', 'error', 'reliable', 'stable', 'broken', 'issue', 'problem'], 'design': ['design', 'layout', 'look', 'ui', 'interface', 'beautiful', 'ugly', 'clean'], 'features': ['feature', 'functionality', 'capability', 'option', 'setting', 'missing'], 'pricing': ['price', 'cost', 'expensive', 'cheap', 'affordable', 'worth', 'value'], 'support': ['support', 'help', 'documentation', 'tutorial', 'guide', 'assistance'] } def analyze_feedback(self, feedback_items): """Analyze sentiment and themes in a list of feedback items.""" results = [] for item in feedback_items: content = item.get('content', '') if not content: continue # Analyze sentiment sentiment_score, sentiment = self._analyze_sentiment(content) # Identify themes themes = self._identify_themes(content) # Extract key phrases key_phrases = self._extract_key_phrases(content) # Add analysis to results results.append({ 'id': item.get('id', ''), 'date': item.get('date', ''), 'source': item.get('source', ''), 'user': item.get('user', ''), 'category': item.get('category', ''), 'content': content, 'rating': item.get('rating', None), 'sentiment_score': sentiment_score, 'sentiment': sentiment, 'themes': themes, 'key_phrases': key_phrases }) return results def _analyze_sentiment(self, text): """Analyze the sentiment of a text.""" # Normalize text text = text.lower() words = re.findall(r'\b\w+\b', text) # Count positive and negative words positive_count = sum(1 for word in words if word in self.positive_words) negative_count = sum(1 for word in words if word in self.negative_words) # Calculate sentiment score (-1 to 1) total_sentiment_words = positive_count + negative_count if total_sentiment_words == 0: sentiment_score = 0 else: sentiment_score = (positive_count - negative_count) / total_sentiment_words # Determine sentiment category if sentiment_score > 0.2: sentiment = 'positive' elif sentiment_score < -0.2: sentiment = 'negative' else: sentiment = 'neutral' return sentiment_score, sentiment def _identify_themes(self, text): """Identify themes in a text.""" text = text.lower() themes = [] for theme, keywords in self.themes.items(): for keyword in keywords: if re.search(r'\b' + re.escape(keyword) + r'\b', text): themes.append(theme) break return themes def _extract_key_phrases(self, text): """Extract key phrases from a text.""" # This is a simplified implementation # For production use, consider using a more sophisticated approach # Split into sentences sentences = re.split(r'[.!?]', text) key_phrases = [] for sentence in sentences: sentence = sentence.strip() if not sentence: continue # Check if sentence contains sentiment words words = re.findall(r'\b\w+\b', sentence.lower()) has_sentiment = any(word in self.positive_words or word in self.negative_words for word in words) if has_sentiment and 3 <= len(words) <= 15: key_phrases.append(sentence) # Limit to top 3 key phrases return key_phrases[:3] def generate_summary(self, analysis_results): """Generate a summary of the sentiment analysis.""" total_items = len(analysis_results) if total_items == 0: return { 'total_items': 0, 'sentiment_distribution': {}, 'theme_distribution': {}, 'average_sentiment_score': 0 } # Count sentiments sentiment_counts = Counter(item['sentiment'] for item in analysis_results) # Count themes theme_counts = Counter() for item in analysis_results: for theme in item['themes']: theme_counts[theme] += 1 # Calculate average sentiment score avg_sentiment_score = sum(item['sentiment_score'] for item in analysis_results) / total_items # Group by date sentiment_by_date = defaultdict(list) for item in analysis_results: date = item.get('date', '') if date: try: # Extract just the date part if it's a datetime if 'T' in date: date = date.split('T')[0] sentiment_by_date[date].append(item['sentiment_score']) except Exception: pass # Calculate average sentiment by date sentiment_trend = { date: sum(scores) / len(scores) for date, scores in sentiment_by_date.items() } return { 'total_items': total_items, 'sentiment_distribution': { 'positive': sentiment_counts.get('positive', 0), 'neutral': sentiment_counts.get('neutral', 0), 'negative': sentiment_counts.get('negative', 0) }, 'theme_distribution': dict(theme_counts), 'average_sentiment_score': avg_sentiment_score, 'sentiment_trend': dict(sorted(sentiment_trend.items())) } def generate_visualizations(self, summary, output_dir): """Generate visualizations of the sentiment analysis.""" output_dir = Path(output_dir) output_dir.mkdir(parents=True, exist_ok=True) # Sentiment distribution pie chart plt.figure(figsize=(10, 6)) labels = ['Positive', 'Neutral', 'Negative'] sizes = [ summary['sentiment_distribution'].get('positive', 0), summary['sentiment_distribution'].get('neutral', 0), summary['sentiment_distribution'].get('negative', 0) ] colors = ['#4CAF50', '#FFC107', '#F44336'] plt.pie(sizes, labels=labels, colors=colors, autopct='%1.1f%%', startangle=90) plt.axis('equal') plt.title('Feedback Sentiment Distribution') plt.savefig(output_dir / 'sentiment_distribution.png') plt.close() # Theme distribution bar chart if summary['theme_distribution']: plt.figure(figsize=(12, 6)) themes = list(summary['theme_distribution'].keys()) counts = list(summary['theme_distribution'].values()) # Sort by count sorted_indices = np.argsort(counts)[::-1] themes = [themes[i] for i in sorted_indices] counts = [counts[i] for i in sorted_indices] plt.bar(themes, counts, color='#2196F3') plt.xlabel('Theme') plt.ylabel('Count') plt.title('Feedback Themes') plt.xticks(rotation=45, ha='right') plt.tight_layout() plt.savefig(output_dir / 'theme_distribution.png') plt.close() # Sentiment trend line chart if summary['sentiment_trend']: plt.figure(figsize=(12, 6)) dates = list(summary['sentiment_trend'].keys()) scores = list(summary['sentiment_trend'].values()) plt.plot(dates, scores, marker='o', linestyle='-', color='#2196F3') plt.axhline(y=0, color='#757575', linestyle='--', alpha=0.7) plt.xlabel('Date') plt.ylabel('Average Sentiment Score') plt.title('Sentiment Trend Over Time') plt.grid(True, alpha=0.3) plt.xticks(rotation=45, ha='right') plt.tight_layout() plt.savefig(output_dir / 'sentiment_trend.png') plt.close() def main(): parser = argparse.ArgumentParser(description="Analyze sentiment in feedback") parser.add_argument("input_file", help="Input feedback file (JSON)") parser.add_argument("--output-dir", default="sentiment_analysis", help="Output directory") args = parser.parse_args() try: # Load feedback data with open(args.input_file, 'r', encoding='utf-8') as f: data = json.load(f) # Extract feedback items if isinstance(data, list): feedback_items = data elif isinstance(data, dict) and 'items' in data: feedback_items = data['items'] else: print("Error: Invalid feedback data format") return # Create analyzer analyzer = SentimentAnalyzer() # Analyze feedback print(f"Analyzing {len(feedback_items)} feedback items...") analysis_results = analyzer.analyze_feedback(feedback_items) # Generate summary summary = analyzer.generate_summary(analysis_results) # Save analysis results output_dir = Path(args.output_dir) output_dir.mkdir(parents=True, exist_ok=True) with open(output_dir / 'sentiment_analysis.json', 'w', encoding='utf-8') as f: json.dump({ 'metadata': { 'timestamp': datetime.now().isoformat(), 'source_file': args.input_file, 'item_count': len(analysis_results) }, 'summary': summary, 'items': analysis_results }, f, indent=2) # Generate visualizations analyzer.generate_visualizations(summary, output_dir) # Print summary print("\nAnalysis Summary:") print(f"Total items: {summary['total_items']}") print("Sentiment distribution:") print(f" Positive: {summary['sentiment_distribution'].get('positive', 0)} ({summary['sentiment_distribution'].get('positive', 0)/summary['total_items']*100:.1f}%)") print(f" Neutral: {summary['sentiment_distribution'].get('neutral', 0)} ({summary['sentiment_distribution'].get('neutral', 0)/summary['total_items']*100:.1f}%)") print(f" Negative: {summary['sentiment_distribution'].get('negative', 0)} ({summary['sentiment_distribution'].get('negative', 0)/summary['total_items']*100:.1f}%)") print(f"Average sentiment score: {summary['average_sentiment_score']:.2f}") print(f"Top themes: {', '.join(k for k, v in sorted(summary['theme_distribution'].items(), key=lambda x: x[1], reverse=True)[:3])}") print(f"\nResults saved to {output_dir}") except Exception as e: print(f"Error: {e}") if __name__ == "__main__": main()