Added new features to the project
This commit is contained in:
ajaysi
344
lib/content_performance_predictor/README.md
Normal file
344
lib/content_performance_predictor/README.md
Normal file
@@ -0,0 +1,344 @@
|
||||
# 🎯 AI Content Performance Predictor
|
||||
|
||||
**LLM-Powered Content Success Prediction for Solo Developers**
|
||||
|
||||
The AI Content Performance Predictor is an intelligent feature that leverages Large Language Models (LLMs) to analyze your content and predict its potential success before you publish. Perfect for solo developers and entrepreneurs who need smart content insights without complex ML infrastructure.
|
||||
|
||||
## 📋 Table of Contents
|
||||
|
||||
- [Overview](#overview)
|
||||
- [Features](#features)
|
||||
- [Installation](#installation)
|
||||
- [Usage](#usage)
|
||||
- [Architecture](#architecture)
|
||||
- [AI Analysis Engine](#ai-analysis-engine)
|
||||
- [API Reference](#api-reference)
|
||||
- [File Structure](#file-structure)
|
||||
- [Configuration](#configuration)
|
||||
- [Development](#development)
|
||||
- [Performance Metrics](#performance-metrics)
|
||||
- [Troubleshooting](#troubleshooting)
|
||||
- [Contributing](#contributing)
|
||||
|
||||
## 🔍 Overview
|
||||
|
||||
The AI Content Performance Predictor uses advanced LLM capabilities to provide intelligent content analysis and predictions:
|
||||
|
||||
- **LLM-Powered Analysis**: Uses your existing `llm_text_gen` integration for smart predictions
|
||||
- **Platform-Specific Insights**: Tailored analysis for Twitter, LinkedIn, Facebook, Instagram, and more
|
||||
- **Zero Training Required**: No ML model training needed - works immediately
|
||||
- **Solo Developer Friendly**: Designed for resource-constrained environments
|
||||
- **Real-time Predictions**: Instant analysis and recommendations
|
||||
|
||||
### Key Benefits
|
||||
|
||||
- **🧠 AI-Powered Intelligence**: Leverages LLM understanding for content analysis
|
||||
- **⚡ Instant Predictions**: No waiting for model training or data collection
|
||||
- **📊 Smart Insights**: Platform-specific recommendations and optimization tips
|
||||
- **🎯 Success Scoring**: Comprehensive performance scoring system
|
||||
- **🔄 Adaptive Learning**: Improves recommendations based on platform best practices
|
||||
- **🎨 Multi-platform**: Optimized for different social media platforms
|
||||
|
||||
## ✨ Features
|
||||
|
||||
### Core Features
|
||||
|
||||
#### 1. **AI Prediction Engine**
|
||||
- Overall performance score (0-100)
|
||||
- Success probability percentage
|
||||
- Platform-specific optimization
|
||||
- Content quality assessment
|
||||
|
||||
#### 2. **LLM Integration**
|
||||
- Uses existing Alwrity LLM infrastructure
|
||||
- No additional API costs or setup
|
||||
- Intelligent content understanding
|
||||
- Context-aware analysis
|
||||
|
||||
#### 3. **Platform Optimization**
|
||||
- Twitter: Character limits, hashtag optimization, engagement factors
|
||||
- LinkedIn: Professional tone, optimal length, business focus
|
||||
- Facebook: Community engagement, storytelling elements
|
||||
- Instagram: Visual content readiness, hashtag strategy
|
||||
|
||||
#### 4. **Smart Recommendations**
|
||||
- Content improvement suggestions
|
||||
- Optimal posting strategies
|
||||
- Engagement enhancement tips
|
||||
- SEO optimization advice
|
||||
|
||||
#### 5. **Interactive UI**
|
||||
- Clean Streamlit interface
|
||||
- Real-time analysis
|
||||
- Visual performance indicators
|
||||
- Actionable insights display
|
||||
|
||||
### Analysis Categories
|
||||
|
||||
1. **📈 Engagement Potential**: Predicted likes, comments, shares
|
||||
2. **🎯 Content Quality**: Overall content effectiveness score
|
||||
3. **⏰ Timing Insights**: Optimal posting time recommendations
|
||||
4. **🔍 SEO Score**: Search engine optimization assessment
|
||||
5. **🏷️ Hashtag Strategy**: Hashtag effectiveness analysis
|
||||
6. **👥 Audience Alignment**: Content-audience fit assessment
|
||||
|
||||
## 🚀 Installation
|
||||
|
||||
### Prerequisites
|
||||
|
||||
```bash
|
||||
# Already included in Alwrity - no additional installation required!
|
||||
# Uses existing dependencies: streamlit, llm_text_gen
|
||||
```
|
||||
|
||||
### Setup
|
||||
|
||||
1. **Auto-Integration** (already included):
|
||||
```python
|
||||
# Available in AI Writer Dashboard
|
||||
# Access via: "AI Content Performance Predictor"
|
||||
```
|
||||
|
||||
2. **Direct Usage**:
|
||||
```python
|
||||
from lib.content_performance_predictor.ai_performance_predictor import AIContentPerformancePredictor
|
||||
```
|
||||
|
||||
3. **UI Component**:
|
||||
```python
|
||||
from lib.content_performance_predictor.ai_performance_predictor import render_ai_predictor_ui
|
||||
```
|
||||
|
||||
## 📖 Usage
|
||||
|
||||
### Through AI Writer Dashboard
|
||||
|
||||
1. Open Alwrity
|
||||
2. Navigate to "AI Writer Dashboard"
|
||||
3. Select "🎯 AI Content Performance Predictor"
|
||||
4. Enter your content and select platform
|
||||
5. Get instant AI-powered predictions!
|
||||
|
||||
### Direct API Usage
|
||||
|
||||
```python
|
||||
from lib.content_performance_predictor.ai_performance_predictor import AIContentPerformancePredictor
|
||||
|
||||
# Initialize predictor
|
||||
predictor = AIContentPerformancePredictor()
|
||||
|
||||
# Analyze content
|
||||
result = await predictor.predict_performance(
|
||||
content="Your amazing content here!",
|
||||
platform="twitter",
|
||||
target_audience="tech entrepreneurs"
|
||||
)
|
||||
|
||||
print(f"Overall Score: {result['overall_score']}")
|
||||
print(f"Recommendations: {result['recommendations']}")
|
||||
```
|
||||
|
||||
### Programmatic Usage
|
||||
|
||||
```python
|
||||
import streamlit as st
|
||||
from lib.content_performance_predictor.ai_performance_predictor import render_ai_predictor_ui
|
||||
|
||||
# Add to your Streamlit app
|
||||
st.title("Content Analysis")
|
||||
render_ai_predictor_ui()
|
||||
```
|
||||
|
||||
### Batch Content Analysis
|
||||
|
||||
```python
|
||||
# Analyze multiple pieces of content
|
||||
contents = [
|
||||
{"content": "Post 1", "platform": "twitter"},
|
||||
{"content": "Post 2", "platform": "linkedin"},
|
||||
{"content": "Post 3", "platform": "facebook"}
|
||||
]
|
||||
|
||||
for content_data in contents:
|
||||
result = await predictor.predict_performance(**content_data)
|
||||
print(f"Content: {content_data['content'][:50]}...")
|
||||
print(f"Score: {result['overall_score']}")
|
||||
print("---")
|
||||
```
|
||||
|
||||
## 🏗️ Architecture
|
||||
|
||||
### System Architecture
|
||||
|
||||
```
|
||||
┌─────────────────────────────────────────────────────────────┐
|
||||
│ STREAMLIT UI │
|
||||
│ (render_ai_predictor_ui) │
|
||||
└─────────────────┬───────────────────────────────────────────┘
|
||||
│
|
||||
┌─────────────────┴───────────────────────────────────────────┐
|
||||
│ AI PREDICTION ENGINE │
|
||||
│ (AIContentPerformancePredictor) │
|
||||
│ ┌─────────────────┐ ┌─────────────────────────────────┐ │
|
||||
│ │ AI Analysis │ │ Platform Configs │ │
|
||||
│ │ (LLM-powered) │ │ (Twitter, LinkedIn, etc.) │ │
|
||||
│ └─────────────────┘ └─────────────────────────────────┘ │
|
||||
└─────────────────┬───────────────────────────────────────────┘
|
||||
│
|
||||
┌─────────────────┴───────────────────────────────────────────┐
|
||||
│ ALWRITY LLM ENGINE │
|
||||
│ (llm_text_gen) │
|
||||
└─────────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
### Component Details
|
||||
|
||||
1. **AIContentPerformancePredictor**: Main prediction class
|
||||
2. **Platform Configurations**: Optimized settings for each platform
|
||||
3. **LLM Integration**: Seamless integration with existing AI infrastructure
|
||||
4. **UI Components**: Interactive Streamlit interface
|
||||
|
||||
## 🧠 AI Analysis Engine
|
||||
|
||||
### LLM-Powered Predictions
|
||||
|
||||
The predictor uses sophisticated prompts to analyze:
|
||||
|
||||
- **Content Quality**: Grammar, readability, engagement potential
|
||||
- **Platform Fit**: Alignment with platform best practices
|
||||
- **Audience Appeal**: Target audience relevance
|
||||
- **Optimization Opportunities**: Specific improvement suggestions
|
||||
|
||||
### Platform-Specific Analysis
|
||||
|
||||
#### Twitter Configuration
|
||||
- Optimal Length: 100-280 characters
|
||||
- Hashtags: 1-3 relevant hashtags
|
||||
- Engagement Factors: Questions, calls-to-action, trending topics
|
||||
|
||||
#### LinkedIn Configuration
|
||||
- Optimal Length: 150-300 words
|
||||
- Professional Tone: Business-focused language
|
||||
- Engagement: Industry insights, professional experiences
|
||||
|
||||
#### Facebook Configuration
|
||||
- Optimal Length: 40-80 characters for high engagement
|
||||
- Community Focus: Shareable, relatable content
|
||||
- Visual Ready: Content that complements images/videos
|
||||
|
||||
#### Instagram Configuration
|
||||
- Visual Emphasis: Content supporting visual storytelling
|
||||
- Hashtags: 5-10 strategic hashtags
|
||||
- Story Potential: Content suitable for Instagram Stories
|
||||
|
||||
## 📊 Performance Metrics
|
||||
|
||||
### Success Indicators
|
||||
|
||||
- **Overall Score**: 0-100 performance prediction
|
||||
- **Platform Alignment**: How well content fits the platform
|
||||
- **Engagement Prediction**: Expected interaction levels
|
||||
- **Optimization Score**: Room for improvement rating
|
||||
|
||||
### Recommendation Categories
|
||||
|
||||
1. **Content Improvements**: Direct text enhancements
|
||||
2. **Platform Optimization**: Platform-specific adjustments
|
||||
3. **Timing Suggestions**: Optimal posting strategies
|
||||
4. **Engagement Boosters**: Tactics to increase interaction
|
||||
|
||||
## 🔧 Configuration
|
||||
|
||||
### Platform Settings
|
||||
|
||||
Located in `ai_performance_predictor.py`:
|
||||
|
||||
```python
|
||||
PLATFORM_CONFIGS = {
|
||||
"twitter": {
|
||||
"optimal_length": {"min": 100, "max": 280},
|
||||
"hashtag_range": {"min": 1, "max": 3},
|
||||
"engagement_factors": ["questions", "cta", "trending"]
|
||||
},
|
||||
# ... other platforms
|
||||
}
|
||||
```
|
||||
|
||||
### Customization
|
||||
|
||||
You can modify:
|
||||
- Platform-specific parameters
|
||||
- Analysis prompts
|
||||
- Scoring algorithms
|
||||
- UI components
|
||||
|
||||
## 🚀 Development
|
||||
|
||||
### Adding New Platforms
|
||||
|
||||
1. Add platform config to `PLATFORM_CONFIGS`
|
||||
2. Update analysis prompts
|
||||
3. Test with platform-specific content
|
||||
|
||||
### Enhancing AI Analysis
|
||||
|
||||
1. Modify prompts in `_create_analysis_prompt()`
|
||||
2. Add new scoring criteria
|
||||
3. Implement additional recommendation types
|
||||
|
||||
## 🔍 Troubleshooting
|
||||
|
||||
### Common Issues
|
||||
|
||||
**No Predictions Generated**:
|
||||
- Check LLM service availability
|
||||
- Verify content input format
|
||||
- Ensure platform is supported
|
||||
|
||||
**Low Accuracy Scores**:
|
||||
- Content may be too short/long for platform
|
||||
- Platform mismatch with content style
|
||||
- Generic content without specific appeal
|
||||
|
||||
**UI Not Loading**:
|
||||
- Check Streamlit dependencies
|
||||
- Verify import paths
|
||||
- Ensure LLM service is configured
|
||||
|
||||
### Debug Mode
|
||||
|
||||
Enable detailed logging:
|
||||
```python
|
||||
import logging
|
||||
logging.basicConfig(level=logging.DEBUG)
|
||||
```
|
||||
|
||||
## 📈 Performance Tips
|
||||
|
||||
1. **Content Length**: Follow platform-specific optimal lengths
|
||||
2. **Platform Selection**: Choose the right platform for your content type
|
||||
3. **Target Audience**: Specify your audience for better predictions
|
||||
4. **Iterate**: Use recommendations to improve content before posting
|
||||
|
||||
## 🤝 Contributing
|
||||
|
||||
1. Fork the repository
|
||||
2. Create a feature branch
|
||||
3. Make your changes
|
||||
4. Test with different content types
|
||||
5. Submit a pull request
|
||||
|
||||
### Development Setup
|
||||
|
||||
```bash
|
||||
# No additional setup required!
|
||||
# Uses existing Alwrity infrastructure
|
||||
```
|
||||
|
||||
## 📝 License
|
||||
|
||||
Part of the Alwrity AI Content Creation Suite.
|
||||
|
||||
---
|
||||
|
||||
**Ready to predict your content's success? Access the AI Content Performance Predictor through the AI Writer Dashboard now!**
|
||||
662
lib/content_performance_predictor/ai_performance_predictor.py
Normal file
662
lib/content_performance_predictor/ai_performance_predictor.py
Normal file
@@ -0,0 +1,662 @@
|
||||
"""
|
||||
AI-Powered Content Performance Predictor
|
||||
|
||||
This module uses AI (LLM) to predict content performance instead of traditional ML models.
|
||||
Perfect for solo developers who want competitive intelligence without expensive ML infrastructure.
|
||||
"""
|
||||
|
||||
import asyncio
|
||||
import json
|
||||
from datetime import datetime, timedelta
|
||||
from typing import Dict, Any, List, Optional
|
||||
from loguru import logger
|
||||
import streamlit as st
|
||||
|
||||
# Import existing Alwrity modules
|
||||
from lib.database.twitter_service import TwitterDatabaseService
|
||||
from lib.ai_web_researcher.google_trends_researcher import do_google_trends_analysis
|
||||
from lib.gpt_providers.text_generation.main_text_generation import llm_text_gen
|
||||
|
||||
|
||||
class AIContentPerformancePredictor:
|
||||
"""
|
||||
AI-powered content performance predictor using LLM intelligence.
|
||||
No ML training required - uses AI's existing knowledge of content patterns.
|
||||
"""
|
||||
|
||||
def __init__(self):
|
||||
"""Initialize the AI predictor."""
|
||||
self.twitter_service = TwitterDatabaseService()
|
||||
self.platform_configs = {
|
||||
'twitter': {
|
||||
'optimal_length': 120,
|
||||
'hashtag_range': (1, 3),
|
||||
'best_times': [9, 12, 15, 18, 21],
|
||||
'engagement_factors': ['questions', 'hashtags', 'mentions', 'visuals']
|
||||
},
|
||||
'linkedin': {
|
||||
'optimal_length': 1500,
|
||||
'hashtag_range': (3, 7),
|
||||
'best_times': [8, 12, 17],
|
||||
'engagement_factors': ['professional_insights', 'industry_expertise', 'networking']
|
||||
},
|
||||
'facebook': {
|
||||
'optimal_length': 200,
|
||||
'hashtag_range': (1, 5),
|
||||
'best_times': [12, 15, 18],
|
||||
'engagement_factors': ['visual_content', 'community_building', 'emotional_connection']
|
||||
},
|
||||
'instagram': {
|
||||
'optimal_length': 150,
|
||||
'hashtag_range': (5, 15),
|
||||
'best_times': [11, 13, 17, 19],
|
||||
'engagement_factors': ['visual_appeal', 'storytelling', 'trending_hashtags']
|
||||
}
|
||||
}
|
||||
|
||||
logger.info("AI Content Performance Predictor initialized")
|
||||
|
||||
async def predict_content_performance(self, content_data: Dict[str, Any]) -> Dict[str, Any]:
|
||||
"""
|
||||
Predict content performance using AI analysis.
|
||||
|
||||
Args:
|
||||
content_data: Dictionary containing content and metadata
|
||||
|
||||
Returns:
|
||||
AI-powered performance prediction with insights
|
||||
"""
|
||||
try:
|
||||
st.info("🧠 AI is analyzing your content...")
|
||||
|
||||
# Extract content details
|
||||
content = content_data.get('content', '')
|
||||
platform = content_data.get('platform', 'twitter')
|
||||
hashtags = content_data.get('hashtags', [])
|
||||
posting_time = content_data.get('posting_time', datetime.now())
|
||||
|
||||
# Get current trends for context
|
||||
trending_context = await self._get_trending_context(platform)
|
||||
|
||||
# Create comprehensive AI prompt for prediction
|
||||
prediction_prompt = self._create_prediction_prompt(
|
||||
content, platform, hashtags, posting_time, trending_context
|
||||
)
|
||||
|
||||
# Get AI prediction
|
||||
ai_response = llm_text_gen(
|
||||
prediction_prompt,
|
||||
system_prompt="You are an expert social media analyst with deep knowledge of content performance patterns across all platforms. Provide specific, actionable predictions."
|
||||
)
|
||||
|
||||
# Parse AI response into structured prediction
|
||||
structured_prediction = self._parse_ai_prediction(ai_response, content_data)
|
||||
|
||||
# Add platform-specific insights
|
||||
platform_insights = self._get_platform_insights(content_data, platform)
|
||||
|
||||
# Generate actionable recommendations
|
||||
recommendations = await self._generate_ai_recommendations(content_data, structured_prediction)
|
||||
|
||||
return {
|
||||
'success': True,
|
||||
'content_analyzed': content[:100] + "..." if len(content) > 100 else content,
|
||||
'platform': platform,
|
||||
'ai_prediction': structured_prediction,
|
||||
'platform_insights': platform_insights,
|
||||
'recommendations': recommendations,
|
||||
'trending_context': trending_context,
|
||||
'analysis_timestamp': datetime.now().isoformat(),
|
||||
'confidence_level': self._calculate_confidence_level(content_data)
|
||||
}
|
||||
|
||||
except Exception as e:
|
||||
error_msg = f"Error in AI prediction: {str(e)}"
|
||||
logger.error(error_msg, exc_info=True)
|
||||
return {'error': error_msg}
|
||||
|
||||
def _create_prediction_prompt(
|
||||
self,
|
||||
content: str,
|
||||
platform: str,
|
||||
hashtags: List[str],
|
||||
posting_time: datetime,
|
||||
trending_context: Dict[str, Any]
|
||||
) -> str:
|
||||
"""Create a comprehensive prompt for AI prediction."""
|
||||
|
||||
config = self.platform_configs.get(platform, {})
|
||||
|
||||
prompt = f"""
|
||||
Analyze this {platform} content and predict its performance:
|
||||
|
||||
CONTENT TO ANALYZE:
|
||||
"{content}"
|
||||
|
||||
METADATA:
|
||||
- Platform: {platform}
|
||||
- Hashtags: {hashtags}
|
||||
- Posting Time: {posting_time.strftime('%A %I:%M %p')}
|
||||
- Content Length: {len(content)} characters
|
||||
- Word Count: {len(content.split())} words
|
||||
|
||||
PLATFORM CONTEXT:
|
||||
- Optimal Length: {config.get('optimal_length', 'N/A')} characters
|
||||
- Recommended Hashtags: {config.get('hashtag_range', 'N/A')}
|
||||
- Best Posting Times: {config.get('best_times', 'N/A')}
|
||||
|
||||
CURRENT TRENDS:
|
||||
{json.dumps(trending_context, indent=2)}
|
||||
|
||||
PREDICTION REQUIREMENTS:
|
||||
Please provide a detailed analysis with these specific predictions:
|
||||
|
||||
1. ENGAGEMENT PREDICTION:
|
||||
- Estimated engagement rate (0-10%)
|
||||
- Estimated likes (number)
|
||||
- Estimated shares/retweets (number)
|
||||
- Estimated comments (number)
|
||||
|
||||
2. PERFORMANCE ANALYSIS:
|
||||
- Strengths of this content
|
||||
- Weaknesses to address
|
||||
- Viral potential (Low/Medium/High)
|
||||
- Audience appeal rating (1-10)
|
||||
|
||||
3. OPTIMIZATION OPPORTUNITIES:
|
||||
- How to improve engagement potential
|
||||
- Better hashtag suggestions
|
||||
- Content format improvements
|
||||
- Timing optimization
|
||||
|
||||
4. COMPETITIVE ASSESSMENT:
|
||||
- How this compares to typical content in this niche
|
||||
- Unique elements that stand out
|
||||
- Missing elements competitors usually include
|
||||
|
||||
Format your response as a detailed analysis with specific numbers and actionable insights.
|
||||
Be realistic but optimistic in your predictions.
|
||||
"""
|
||||
|
||||
return prompt
|
||||
|
||||
async def _get_trending_context(self, platform: str) -> Dict[str, Any]:
|
||||
"""Get current trending context for better predictions."""
|
||||
try:
|
||||
# Use existing Twitter integration if available
|
||||
if platform == 'twitter' and hasattr(self.twitter_service, 'get_trending_topics'):
|
||||
trending_topics = self.twitter_service.get_trending_topics()
|
||||
else:
|
||||
# Fallback to general trends
|
||||
trending_topics = [
|
||||
'AI and technology',
|
||||
'Content creation',
|
||||
'Social media marketing',
|
||||
'Digital transformation',
|
||||
'Remote work'
|
||||
]
|
||||
|
||||
return {
|
||||
'trending_topics': trending_topics[:5],
|
||||
'platform': platform,
|
||||
'analysis_date': datetime.now().isoformat()
|
||||
}
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error getting trending context: {str(e)}")
|
||||
return {
|
||||
'trending_topics': ['General content', 'Engagement tips'],
|
||||
'platform': platform,
|
||||
'analysis_date': datetime.now().isoformat()
|
||||
}
|
||||
|
||||
def _parse_ai_prediction(self, ai_response: str, content_data: Dict[str, Any]) -> Dict[str, Any]:
|
||||
"""Parse AI response into structured prediction data."""
|
||||
try:
|
||||
# Extract numerical predictions using simple parsing
|
||||
# This is a simplified version - in production, you might want more sophisticated parsing
|
||||
|
||||
prediction = {
|
||||
'engagement_rate': self._extract_percentage(ai_response, 'engagement rate'),
|
||||
'estimated_likes': self._extract_number(ai_response, 'likes'),
|
||||
'estimated_shares': self._extract_number(ai_response, ['shares', 'retweets']),
|
||||
'estimated_comments': self._extract_number(ai_response, 'comments'),
|
||||
'viral_potential': self._extract_rating(ai_response, 'viral potential'),
|
||||
'audience_appeal': self._extract_rating(ai_response, 'audience appeal'),
|
||||
'strengths': self._extract_list_items(ai_response, 'strengths'),
|
||||
'weaknesses': self._extract_list_items(ai_response, 'weaknesses'),
|
||||
'full_analysis': ai_response
|
||||
}
|
||||
|
||||
return prediction
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error parsing AI prediction: {str(e)}")
|
||||
return {
|
||||
'engagement_rate': 2.5, # Default reasonable prediction
|
||||
'estimated_likes': 50,
|
||||
'estimated_shares': 10,
|
||||
'estimated_comments': 5,
|
||||
'viral_potential': 'Medium',
|
||||
'audience_appeal': 7,
|
||||
'full_analysis': ai_response
|
||||
}
|
||||
|
||||
def _get_platform_insights(self, content_data: Dict[str, Any], platform: str) -> Dict[str, Any]:
|
||||
"""Get platform-specific insights."""
|
||||
config = self.platform_configs.get(platform, {})
|
||||
content = content_data.get('content', '')
|
||||
hashtags = content_data.get('hashtags', [])
|
||||
|
||||
insights = {
|
||||
'platform_optimization': [],
|
||||
'timing_analysis': {},
|
||||
'format_analysis': {},
|
||||
'hashtag_analysis': {}
|
||||
}
|
||||
|
||||
# Length analysis
|
||||
optimal_length = config.get('optimal_length', 200)
|
||||
current_length = len(content)
|
||||
|
||||
if abs(current_length - optimal_length) > 50:
|
||||
insights['platform_optimization'].append(
|
||||
f"Content length ({current_length}) differs from optimal ({optimal_length}) for {platform}"
|
||||
)
|
||||
else:
|
||||
insights['platform_optimization'].append(
|
||||
f"Content length is well-optimized for {platform}"
|
||||
)
|
||||
|
||||
# Hashtag analysis
|
||||
hashtag_range = config.get('hashtag_range', (1, 5))
|
||||
hashtag_count = len(hashtags)
|
||||
|
||||
if hashtag_count < hashtag_range[0]:
|
||||
insights['hashtag_analysis']['recommendation'] = f"Add more hashtags (optimal: {hashtag_range[0]}-{hashtag_range[1]})"
|
||||
elif hashtag_count > hashtag_range[1]:
|
||||
insights['hashtag_analysis']['recommendation'] = f"Consider reducing hashtags (optimal: {hashtag_range[0]}-{hashtag_range[1]})"
|
||||
else:
|
||||
insights['hashtag_analysis']['recommendation'] = "Hashtag count is optimal"
|
||||
|
||||
# Timing analysis
|
||||
best_times = config.get('best_times', [])
|
||||
current_hour = datetime.now().hour
|
||||
|
||||
insights['timing_analysis'] = {
|
||||
'best_times': best_times,
|
||||
'current_timing': 'Optimal' if current_hour in best_times else 'Suboptimal',
|
||||
'suggestion': f"Consider posting at {best_times} for better engagement" if current_hour not in best_times else "Current timing is optimal"
|
||||
}
|
||||
|
||||
return insights
|
||||
|
||||
async def _generate_ai_recommendations(
|
||||
self,
|
||||
content_data: Dict[str, Any],
|
||||
prediction: Dict[str, Any]
|
||||
) -> List[Dict[str, str]]:
|
||||
"""Generate AI-powered recommendations for improvement."""
|
||||
|
||||
recommendations_prompt = f"""
|
||||
Based on this content analysis, provide specific improvement recommendations:
|
||||
|
||||
CONTENT: "{content_data.get('content', '')[:200]}..."
|
||||
PLATFORM: {content_data.get('platform', 'twitter')}
|
||||
PREDICTED ENGAGEMENT: {prediction.get('engagement_rate', 'N/A')}%
|
||||
VIRAL POTENTIAL: {prediction.get('viral_potential', 'N/A')}
|
||||
|
||||
Provide 5-7 specific, actionable recommendations to improve this content's performance:
|
||||
|
||||
1. Content optimization suggestions
|
||||
2. Hashtag improvements
|
||||
3. Timing recommendations
|
||||
4. Format enhancements
|
||||
5. Engagement boosters
|
||||
6. Audience targeting tips
|
||||
7. Platform-specific optimizations
|
||||
|
||||
Format each recommendation as:
|
||||
- Category: [category]
|
||||
- Action: [specific action to take]
|
||||
- Expected Impact: [what improvement to expect]
|
||||
- Priority: [High/Medium/Low]
|
||||
|
||||
Focus on quick wins and high-impact changes.
|
||||
"""
|
||||
|
||||
try:
|
||||
ai_recommendations = llm_text_gen(
|
||||
recommendations_prompt,
|
||||
system_prompt="You are a content optimization expert. Provide specific, actionable recommendations that can be implemented immediately."
|
||||
)
|
||||
|
||||
# Parse recommendations into structured format
|
||||
return self._parse_recommendations(ai_recommendations)
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error generating AI recommendations: {str(e)}")
|
||||
return [
|
||||
{
|
||||
'category': 'Content Enhancement',
|
||||
'action': 'Add more engaging elements like questions or calls-to-action',
|
||||
'expected_impact': 'Increase engagement by 20-30%',
|
||||
'priority': 'High'
|
||||
},
|
||||
{
|
||||
'category': 'Hashtag Optimization',
|
||||
'action': 'Research and add 2-3 trending relevant hashtags',
|
||||
'expected_impact': 'Improve discoverability',
|
||||
'priority': 'Medium'
|
||||
}
|
||||
]
|
||||
|
||||
def _extract_percentage(self, text: str, keyword: str) -> float:
|
||||
"""Extract percentage value from AI response."""
|
||||
import re
|
||||
patterns = [
|
||||
rf'{keyword}.*?(\d+\.?\d*)%',
|
||||
rf'(\d+\.?\d*)%.*?{keyword}',
|
||||
rf'{keyword}.*?(\d+\.?\d*) percent'
|
||||
]
|
||||
|
||||
for pattern in patterns:
|
||||
match = re.search(pattern, text, re.IGNORECASE)
|
||||
if match:
|
||||
return float(match.group(1))
|
||||
|
||||
return 2.5 # Default reasonable engagement rate
|
||||
|
||||
def _extract_number(self, text: str, keywords: List[str]) -> int:
|
||||
"""Extract number from AI response."""
|
||||
import re
|
||||
|
||||
if isinstance(keywords, str):
|
||||
keywords = [keywords]
|
||||
|
||||
for keyword in keywords:
|
||||
patterns = [
|
||||
rf'{keyword}.*?(\d+)',
|
||||
rf'(\d+).*?{keyword}'
|
||||
]
|
||||
|
||||
for pattern in patterns:
|
||||
match = re.search(pattern, text, re.IGNORECASE)
|
||||
if match:
|
||||
return int(match.group(1))
|
||||
|
||||
return 25 # Default reasonable number
|
||||
|
||||
def _extract_rating(self, text: str, keyword: str) -> str:
|
||||
"""Extract rating (High/Medium/Low) from AI response."""
|
||||
import re
|
||||
|
||||
pattern = rf'{keyword}.*?(High|Medium|Low)'
|
||||
match = re.search(pattern, text, re.IGNORECASE)
|
||||
|
||||
if match:
|
||||
return match.group(1).capitalize()
|
||||
|
||||
return 'Medium' # Default
|
||||
|
||||
def _extract_list_items(self, text: str, section: str) -> List[str]:
|
||||
"""Extract list items from a section of AI response."""
|
||||
import re
|
||||
|
||||
# Find the section
|
||||
section_pattern = rf'{section}:?\s*(.*?)(?=\n\n|\d\.|[A-Z]+:|$)'
|
||||
match = re.search(section_pattern, text, re.IGNORECASE | re.DOTALL)
|
||||
|
||||
if match:
|
||||
section_text = match.group(1)
|
||||
# Extract bullet points or numbered items
|
||||
items = re.findall(r'[-•]\s*(.+)', section_text)
|
||||
if not items:
|
||||
items = re.findall(r'\d+\.\s*(.+)', section_text)
|
||||
|
||||
return [item.strip() for item in items[:3]] # Return first 3 items
|
||||
|
||||
return []
|
||||
|
||||
def _parse_recommendations(self, ai_recommendations: str) -> List[Dict[str, str]]:
|
||||
"""Parse AI recommendations into structured format."""
|
||||
recommendations = []
|
||||
|
||||
try:
|
||||
# Simple parsing - split by numbers or bullet points
|
||||
import re
|
||||
sections = re.split(r'\d+\.|[-•]', ai_recommendations)
|
||||
|
||||
for section in sections[1:6]: # Take first 5 recommendations
|
||||
if len(section.strip()) > 10: # Only substantial recommendations
|
||||
recommendations.append({
|
||||
'category': 'AI Recommendation',
|
||||
'action': section.strip()[:200], # Limit length
|
||||
'expected_impact': 'Improved engagement',
|
||||
'priority': 'Medium'
|
||||
})
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error parsing recommendations: {str(e)}")
|
||||
|
||||
# Ensure we have at least a few recommendations
|
||||
if len(recommendations) < 3:
|
||||
recommendations.extend([
|
||||
{
|
||||
'category': 'Engagement',
|
||||
'action': 'Add questions to encourage audience interaction',
|
||||
'expected_impact': '20-30% more engagement',
|
||||
'priority': 'High'
|
||||
},
|
||||
{
|
||||
'category': 'Visibility',
|
||||
'action': 'Use trending hashtags relevant to your niche',
|
||||
'expected_impact': 'Better discoverability',
|
||||
'priority': 'Medium'
|
||||
},
|
||||
{
|
||||
'category': 'Timing',
|
||||
'action': 'Post during peak engagement hours for your audience',
|
||||
'expected_impact': '15-25% more reach',
|
||||
'priority': 'Medium'
|
||||
}
|
||||
])
|
||||
|
||||
return recommendations[:7] # Return max 7 recommendations
|
||||
|
||||
def _calculate_confidence_level(self, content_data: Dict[str, Any]) -> str:
|
||||
"""Calculate confidence level of prediction."""
|
||||
confidence_factors = 0
|
||||
|
||||
# More complete data = higher confidence
|
||||
if content_data.get('content'):
|
||||
confidence_factors += 1
|
||||
if content_data.get('hashtags'):
|
||||
confidence_factors += 1
|
||||
if content_data.get('platform'):
|
||||
confidence_factors += 1
|
||||
if content_data.get('posting_time'):
|
||||
confidence_factors += 1
|
||||
|
||||
if confidence_factors >= 4:
|
||||
return 'High'
|
||||
elif confidence_factors >= 2:
|
||||
return 'Medium'
|
||||
else:
|
||||
return 'Low'
|
||||
|
||||
async def analyze_content_batch(self, content_list: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
|
||||
"""Analyze multiple pieces of content."""
|
||||
results = []
|
||||
|
||||
for i, content_data in enumerate(content_list):
|
||||
st.write(f"🔍 Analyzing content {i+1}/{len(content_list)}")
|
||||
result = await self.predict_content_performance(content_data)
|
||||
results.append(result)
|
||||
|
||||
return results
|
||||
|
||||
def get_platform_best_practices(self, platform: str) -> Dict[str, Any]:
|
||||
"""Get best practices for a specific platform."""
|
||||
config = self.platform_configs.get(platform, {})
|
||||
|
||||
return {
|
||||
'platform': platform,
|
||||
'optimal_length': config.get('optimal_length'),
|
||||
'hashtag_range': config.get('hashtag_range'),
|
||||
'best_posting_times': config.get('best_times'),
|
||||
'engagement_factors': config.get('engagement_factors', []),
|
||||
'tips': [
|
||||
f"Keep content around {config.get('optimal_length', 200)} characters",
|
||||
f"Use {config.get('hashtag_range', (1, 5))[0]}-{config.get('hashtag_range', (1, 5))[1]} relevant hashtags",
|
||||
f"Post during peak hours: {config.get('best_times', [])}",
|
||||
"Include engaging elements like questions or calls-to-action",
|
||||
"Use visuals when possible to increase engagement"
|
||||
]
|
||||
}
|
||||
|
||||
|
||||
# Usage example and Streamlit interface
|
||||
def render_ai_predictor_ui():
|
||||
"""Render the AI content performance predictor interface."""
|
||||
st.title("🎯 AI Content Performance Predictor")
|
||||
st.markdown("Get AI-powered predictions for your content performance - no ML training required!")
|
||||
|
||||
# Initialize predictor
|
||||
if 'ai_predictor' not in st.session_state:
|
||||
st.session_state.ai_predictor = AIContentPerformancePredictor()
|
||||
|
||||
predictor = st.session_state.ai_predictor
|
||||
|
||||
# Input section
|
||||
st.header("📝 Content Analysis")
|
||||
|
||||
col1, col2 = st.columns(2)
|
||||
|
||||
with col1:
|
||||
platform = st.selectbox(
|
||||
"Platform",
|
||||
["twitter", "linkedin", "facebook", "instagram"],
|
||||
help="Choose your target platform"
|
||||
)
|
||||
|
||||
posting_time = st.time_input("Posting Time", value=datetime.now().time())
|
||||
|
||||
with col2:
|
||||
hashtags_input = st.text_input(
|
||||
"Hashtags (comma-separated)",
|
||||
value="AI, ContentCreation, Marketing",
|
||||
help="Enter hashtags without # symbol"
|
||||
)
|
||||
|
||||
content = st.text_area(
|
||||
"Content to Analyze",
|
||||
value="Discover how AI is revolutionizing content creation! What's your experience with AI tools? Share your thoughts below! 🚀",
|
||||
height=150,
|
||||
help="Enter the content you want to analyze"
|
||||
)
|
||||
|
||||
# Process hashtags
|
||||
hashtags = [tag.strip() for tag in hashtags_input.split(',') if tag.strip()]
|
||||
|
||||
if st.button("🧠 Analyze Content Performance", type="primary"):
|
||||
if content:
|
||||
# Prepare content data
|
||||
content_data = {
|
||||
'content': content,
|
||||
'platform': platform,
|
||||
'hashtags': hashtags,
|
||||
'posting_time': datetime.combine(datetime.now().date(), posting_time)
|
||||
}
|
||||
|
||||
# Run AI analysis
|
||||
with st.spinner("🤖 AI is analyzing your content..."):
|
||||
results = asyncio.run(predictor.predict_content_performance(content_data))
|
||||
|
||||
if results.get('success'):
|
||||
st.success("✅ Analysis Complete!")
|
||||
|
||||
# Display predictions
|
||||
st.header("📊 AI Performance Prediction")
|
||||
|
||||
prediction = results.get('ai_prediction', {})
|
||||
|
||||
# Key metrics
|
||||
col1, col2, col3, col4 = st.columns(4)
|
||||
|
||||
with col1:
|
||||
st.metric(
|
||||
"Engagement Rate",
|
||||
f"{prediction.get('engagement_rate', 0):.1f}%"
|
||||
)
|
||||
|
||||
with col2:
|
||||
st.metric(
|
||||
"Est. Likes",
|
||||
f"{prediction.get('estimated_likes', 0):,}"
|
||||
)
|
||||
|
||||
with col3:
|
||||
st.metric(
|
||||
"Est. Shares",
|
||||
f"{prediction.get('estimated_shares', 0):,}"
|
||||
)
|
||||
|
||||
with col4:
|
||||
st.metric(
|
||||
"Viral Potential",
|
||||
prediction.get('viral_potential', 'Medium')
|
||||
)
|
||||
|
||||
# Platform insights
|
||||
platform_insights = results.get('platform_insights', {})
|
||||
if platform_insights:
|
||||
st.subheader("🎯 Platform Optimization")
|
||||
|
||||
for insight in platform_insights.get('platform_optimization', []):
|
||||
st.info(f"💡 {insight}")
|
||||
|
||||
# Timing analysis
|
||||
timing = platform_insights.get('timing_analysis', {})
|
||||
if timing:
|
||||
st.write(f"**Timing Analysis:** {timing.get('suggestion', 'N/A')}")
|
||||
|
||||
# Hashtag analysis
|
||||
hashtag_analysis = platform_insights.get('hashtag_analysis', {})
|
||||
if hashtag_analysis:
|
||||
st.write(f"**Hashtag Recommendation:** {hashtag_analysis.get('recommendation', 'N/A')}")
|
||||
|
||||
# AI Recommendations
|
||||
recommendations = results.get('recommendations', [])
|
||||
if recommendations:
|
||||
st.subheader("🚀 AI Recommendations")
|
||||
|
||||
for i, rec in enumerate(recommendations):
|
||||
with st.expander(f"💡 {rec.get('category', 'Recommendation')} - {rec.get('priority', 'Medium')} Priority"):
|
||||
st.write(f"**Action:** {rec.get('action', 'N/A')}")
|
||||
st.write(f"**Expected Impact:** {rec.get('expected_impact', 'N/A')}")
|
||||
|
||||
# Full AI Analysis
|
||||
if prediction.get('full_analysis'):
|
||||
with st.expander("🤖 Complete AI Analysis"):
|
||||
st.write(prediction['full_analysis'])
|
||||
|
||||
else:
|
||||
st.error(f"❌ Analysis failed: {results.get('error')}")
|
||||
else:
|
||||
st.warning("⚠️ Please enter content to analyze")
|
||||
|
||||
# Platform best practices
|
||||
st.sidebar.header("📚 Platform Best Practices")
|
||||
selected_platform = st.sidebar.selectbox("Get tips for:", ["twitter", "linkedin", "facebook", "instagram"])
|
||||
|
||||
best_practices = predictor.get_platform_best_practices(selected_platform)
|
||||
|
||||
st.sidebar.write(f"**{selected_platform.title()} Best Practices:**")
|
||||
for tip in best_practices.get('tips', []):
|
||||
st.sidebar.write(f"• {tip}")
|
||||
|
||||
|
||||
# Main execution
|
||||
if __name__ == "__main__":
|
||||
render_ai_predictor_ui()
|
||||
Reference in New Issue
Block a user