Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,492 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import pandas as pd
|
| 3 |
+
import numpy as np
|
| 4 |
+
import json
|
| 5 |
+
import re
|
| 6 |
+
from datetime import datetime
|
| 7 |
+
from typing import List, Dict, Tuple
|
| 8 |
+
from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassification
|
| 9 |
+
import plotly.graph_objects as go
|
| 10 |
+
from plotly.subplots import make_subplots
|
| 11 |
+
import sqlite3
|
| 12 |
+
import hashlib
|
| 13 |
+
import time
|
| 14 |
+
|
| 15 |
+
# Initialize models
|
| 16 |
+
sentiment_analyzer = pipeline("sentiment-analysis", model="cardiffnlp/twitter-roberta-base-sentiment-latest")
|
| 17 |
+
tokenizer = AutoTokenizer.from_pretrained("cardiffnlp/twitter-roberta-base-sentiment-latest")
|
| 18 |
+
|
| 19 |
+
class ReviewAnalyzer:
|
| 20 |
+
def __init__(self):
|
| 21 |
+
self.db_path = "reviews.db"
|
| 22 |
+
self._init_db()
|
| 23 |
+
|
| 24 |
+
def _init_db(self):
|
| 25 |
+
conn = sqlite3.connect(self.db_path)
|
| 26 |
+
conn.execute('''
|
| 27 |
+
CREATE TABLE IF NOT EXISTS usage_log (
|
| 28 |
+
id INTEGER PRIMARY KEY,
|
| 29 |
+
user_id TEXT,
|
| 30 |
+
timestamp DATETIME,
|
| 31 |
+
analysis_type TEXT,
|
| 32 |
+
items_count INTEGER
|
| 33 |
+
)
|
| 34 |
+
''')
|
| 35 |
+
conn.close()
|
| 36 |
+
|
| 37 |
+
def preprocess_text(self, text: str) -> str:
|
| 38 |
+
"""Clean and preprocess review text"""
|
| 39 |
+
text = re.sub(r'http\S+', '', text) # Remove URLs
|
| 40 |
+
text = re.sub(r'[^\w\s]', '', text) # Remove special chars
|
| 41 |
+
text = text.strip().lower()
|
| 42 |
+
return text
|
| 43 |
+
|
| 44 |
+
def analyze_sentiment(self, reviews: List[str]) -> Dict:
|
| 45 |
+
"""Analyze sentiment of reviews"""
|
| 46 |
+
results = []
|
| 47 |
+
sentiments = {'positive': 0, 'negative': 0, 'neutral': 0}
|
| 48 |
+
|
| 49 |
+
for review in reviews:
|
| 50 |
+
if not review.strip():
|
| 51 |
+
continue
|
| 52 |
+
|
| 53 |
+
clean_review = self.preprocess_text(review)
|
| 54 |
+
result = sentiment_analyzer(clean_review)[0]
|
| 55 |
+
|
| 56 |
+
label = result['label'].lower()
|
| 57 |
+
score = result['score']
|
| 58 |
+
|
| 59 |
+
# Map labels to standard format
|
| 60 |
+
if 'pos' in label:
|
| 61 |
+
sentiment = 'positive'
|
| 62 |
+
elif 'neg' in label:
|
| 63 |
+
sentiment = 'negative'
|
| 64 |
+
else:
|
| 65 |
+
sentiment = 'neutral'
|
| 66 |
+
|
| 67 |
+
sentiments[sentiment] += 1
|
| 68 |
+
results.append({
|
| 69 |
+
'text': review[:100] + '...' if len(review) > 100 else review,
|
| 70 |
+
'sentiment': sentiment,
|
| 71 |
+
'confidence': round(score, 3)
|
| 72 |
+
})
|
| 73 |
+
|
| 74 |
+
total = len(results)
|
| 75 |
+
sentiment_percentages = {k: round(v/total*100, 1) for k, v in sentiments.items()}
|
| 76 |
+
|
| 77 |
+
return {
|
| 78 |
+
'summary': sentiment_percentages,
|
| 79 |
+
'details': results,
|
| 80 |
+
'total_reviews': total
|
| 81 |
+
}
|
| 82 |
+
|
| 83 |
+
def detect_fake_reviews(self, reviews: List[str]) -> Dict:
|
| 84 |
+
"""Detect potentially fake reviews"""
|
| 85 |
+
fake_scores = []
|
| 86 |
+
|
| 87 |
+
for review in reviews:
|
| 88 |
+
if not review.strip():
|
| 89 |
+
continue
|
| 90 |
+
|
| 91 |
+
# Simple fake detection heuristics
|
| 92 |
+
score = 0
|
| 93 |
+
|
| 94 |
+
# Length check
|
| 95 |
+
if len(review) < 20:
|
| 96 |
+
score += 0.3
|
| 97 |
+
|
| 98 |
+
# Repetitive words
|
| 99 |
+
words = review.lower().split()
|
| 100 |
+
unique_ratio = len(set(words)) / len(words) if words else 0
|
| 101 |
+
if unique_ratio < 0.5:
|
| 102 |
+
score += 0.4
|
| 103 |
+
|
| 104 |
+
# Excessive punctuation
|
| 105 |
+
punct_ratio = len(re.findall(r'[!?.]', review)) / len(review) if review else 0
|
| 106 |
+
if punct_ratio > 0.1:
|
| 107 |
+
score += 0.2
|
| 108 |
+
|
| 109 |
+
# Generic phrases
|
| 110 |
+
generic_phrases = ['amazing', 'perfect', 'best ever', 'highly recommend']
|
| 111 |
+
if any(phrase in review.lower() for phrase in generic_phrases):
|
| 112 |
+
score += 0.1
|
| 113 |
+
|
| 114 |
+
fake_scores.append({
|
| 115 |
+
'text': review[:100] + '...' if len(review) > 100 else review,
|
| 116 |
+
'fake_probability': min(round(score, 3), 1.0),
|
| 117 |
+
'status': 'suspicious' if score > 0.5 else 'authentic'
|
| 118 |
+
})
|
| 119 |
+
|
| 120 |
+
suspicious_count = sum(1 for item in fake_scores if item['fake_probability'] > 0.5)
|
| 121 |
+
|
| 122 |
+
return {
|
| 123 |
+
'summary': {
|
| 124 |
+
'total_reviews': len(fake_scores),
|
| 125 |
+
'suspicious_reviews': suspicious_count,
|
| 126 |
+
'authenticity_rate': round((len(fake_scores) - suspicious_count) / len(fake_scores) * 100, 1) if fake_scores else 0
|
| 127 |
+
},
|
| 128 |
+
'details': fake_scores
|
| 129 |
+
}
|
| 130 |
+
|
| 131 |
+
def assess_quality(self, reviews: List[str]) -> Dict:
|
| 132 |
+
"""Assess review quality"""
|
| 133 |
+
quality_scores = []
|
| 134 |
+
|
| 135 |
+
for review in reviews:
|
| 136 |
+
if not review.strip():
|
| 137 |
+
continue
|
| 138 |
+
|
| 139 |
+
score = 0
|
| 140 |
+
factors = {}
|
| 141 |
+
|
| 142 |
+
# Length factor
|
| 143 |
+
length_score = min(len(review) / 200, 1.0)
|
| 144 |
+
factors['length'] = round(length_score, 2)
|
| 145 |
+
score += length_score * 0.3
|
| 146 |
+
|
| 147 |
+
# Detail factor (specific words)
|
| 148 |
+
detail_words = ['because', 'however', 'although', 'specifically', 'particularly']
|
| 149 |
+
detail_score = min(sum(1 for word in detail_words if word in review.lower()) / 3, 1.0)
|
| 150 |
+
factors['detail'] = round(detail_score, 2)
|
| 151 |
+
score += detail_score * 0.3
|
| 152 |
+
|
| 153 |
+
# Structure factor
|
| 154 |
+
sentences = len(re.split(r'[.!?]', review))
|
| 155 |
+
structure_score = min(sentences / 5, 1.0)
|
| 156 |
+
factors['structure'] = round(structure_score, 2)
|
| 157 |
+
score += structure_score * 0.2
|
| 158 |
+
|
| 159 |
+
# Helpfulness factor
|
| 160 |
+
helpful_words = ['pros', 'cons', 'recommend', 'suggest', 'tip', 'advice']
|
| 161 |
+
helpful_score = min(sum(1 for word in helpful_words if word in review.lower()) / 2, 1.0)
|
| 162 |
+
factors['helpfulness'] = round(helpful_score, 2)
|
| 163 |
+
score += helpful_score * 0.2
|
| 164 |
+
|
| 165 |
+
quality_scores.append({
|
| 166 |
+
'text': review[:100] + '...' if len(review) > 100 else review,
|
| 167 |
+
'quality_score': round(score, 3),
|
| 168 |
+
'factors': factors,
|
| 169 |
+
'grade': 'A' if score > 0.8 else 'B' if score > 0.6 else 'C' if score > 0.4 else 'D'
|
| 170 |
+
})
|
| 171 |
+
|
| 172 |
+
avg_quality = sum(item['quality_score'] for item in quality_scores) / len(quality_scores) if quality_scores else 0
|
| 173 |
+
|
| 174 |
+
return {
|
| 175 |
+
'summary': {
|
| 176 |
+
'average_quality': round(avg_quality, 3),
|
| 177 |
+
'total_reviews': len(quality_scores),
|
| 178 |
+
'high_quality_count': sum(1 for item in quality_scores if item['quality_score'] > 0.7)
|
| 179 |
+
},
|
| 180 |
+
'details': quality_scores
|
| 181 |
+
}
|
| 182 |
+
|
| 183 |
+
def compare_competitors(self, product_a_reviews: List[str], product_b_reviews: List[str]) -> Tuple[Dict, go.Figure]:
|
| 184 |
+
"""Compare sentiment between two products"""
|
| 185 |
+
analysis_a = self.analyze_sentiment(product_a_reviews)
|
| 186 |
+
analysis_b = self.analyze_sentiment(product_b_reviews)
|
| 187 |
+
|
| 188 |
+
# Create comparison chart
|
| 189 |
+
fig = make_subplots(
|
| 190 |
+
rows=1, cols=2,
|
| 191 |
+
specs=[[{'type': 'pie'}, {'type': 'pie'}]],
|
| 192 |
+
subplot_titles=['Product A', 'Product B']
|
| 193 |
+
)
|
| 194 |
+
|
| 195 |
+
# Product A pie chart
|
| 196 |
+
fig.add_trace(go.Pie(
|
| 197 |
+
labels=list(analysis_a['summary'].keys()),
|
| 198 |
+
values=list(analysis_a['summary'].values()),
|
| 199 |
+
name="Product A"
|
| 200 |
+
), row=1, col=1)
|
| 201 |
+
|
| 202 |
+
# Product B pie chart
|
| 203 |
+
fig.add_trace(go.Pie(
|
| 204 |
+
labels=list(analysis_b['summary'].keys()),
|
| 205 |
+
values=list(analysis_b['summary'].values()),
|
| 206 |
+
name="Product B"
|
| 207 |
+
), row=1, col=2)
|
| 208 |
+
|
| 209 |
+
fig.update_layout(title_text="Sentiment Comparison")
|
| 210 |
+
|
| 211 |
+
comparison = {
|
| 212 |
+
'product_a': analysis_a,
|
| 213 |
+
'product_b': analysis_b,
|
| 214 |
+
'winner': 'Product A' if analysis_a['summary']['positive'] > analysis_b['summary']['positive'] else 'Product B'
|
| 215 |
+
}
|
| 216 |
+
|
| 217 |
+
return comparison, fig
|
| 218 |
+
|
| 219 |
+
def generate_report(self, analysis_data: Dict, report_type: str = "basic") -> str:
|
| 220 |
+
"""Generate analysis report"""
|
| 221 |
+
timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
|
| 222 |
+
|
| 223 |
+
if report_type == "sentiment":
|
| 224 |
+
return f"""
|
| 225 |
+
# Sentiment Analysis Report
|
| 226 |
+
Generated: {timestamp}
|
| 227 |
+
|
| 228 |
+
## Summary
|
| 229 |
+
- Total Reviews: {analysis_data.get('total_reviews', 0)}
|
| 230 |
+
- Positive: {analysis_data.get('summary', {}).get('positive', 0)}%
|
| 231 |
+
- Negative: {analysis_data.get('summary', {}).get('negative', 0)}%
|
| 232 |
+
- Neutral: {analysis_data.get('summary', {}).get('neutral', 0)}%
|
| 233 |
+
|
| 234 |
+
## Key Insights
|
| 235 |
+
- Overall sentiment trend: {'Positive' if analysis_data.get('summary', {}).get('positive', 0) > 50 else 'Mixed'}
|
| 236 |
+
- Customer satisfaction level: {'High' if analysis_data.get('summary', {}).get('positive', 0) > 70 else 'Moderate' if analysis_data.get('summary', {}).get('positive', 0) > 40 else 'Low'}
|
| 237 |
+
|
| 238 |
+
## Recommendations
|
| 239 |
+
- Focus on addressing negative feedback themes
|
| 240 |
+
- Leverage positive reviews for marketing
|
| 241 |
+
- Monitor sentiment trends over time
|
| 242 |
+
"""
|
| 243 |
+
|
| 244 |
+
elif report_type == "fake":
|
| 245 |
+
return f"""
|
| 246 |
+
# Fake Review Detection Report
|
| 247 |
+
Generated: {timestamp}
|
| 248 |
+
|
| 249 |
+
## Summary
|
| 250 |
+
- Total Reviews Analyzed: {analysis_data.get('summary', {}).get('total_reviews', 0)}
|
| 251 |
+
- Suspicious Reviews: {analysis_data.get('summary', {}).get('suspicious_reviews', 0)}
|
| 252 |
+
- Authenticity Rate: {analysis_data.get('summary', {}).get('authenticity_rate', 0)}%
|
| 253 |
+
|
| 254 |
+
## Risk Assessment
|
| 255 |
+
- Review Quality: {'High Risk' if analysis_data.get('summary', {}).get('authenticity_rate', 0) < 70 else 'Low Risk'}
|
| 256 |
+
- Recommendation: {'Investigate suspicious reviews' if analysis_data.get('summary', {}).get('suspicious_reviews', 0) > 0 else 'Reviews appear authentic'}
|
| 257 |
+
"""
|
| 258 |
+
|
| 259 |
+
return "Report generated successfully"
|
| 260 |
+
|
| 261 |
+
# Global analyzer instance
|
| 262 |
+
analyzer = ReviewAnalyzer()
|
| 263 |
+
|
| 264 |
+
def process_reviews_input(text: str) -> List[str]:
|
| 265 |
+
"""Process review input text into list"""
|
| 266 |
+
if not text.strip():
|
| 267 |
+
return []
|
| 268 |
+
|
| 269 |
+
# Split by lines or by common separators
|
| 270 |
+
reviews = []
|
| 271 |
+
for line in text.split('\n'):
|
| 272 |
+
line = line.strip()
|
| 273 |
+
if line and len(line) > 10: # Minimum length check
|
| 274 |
+
reviews.append(line)
|
| 275 |
+
|
| 276 |
+
return reviews
|
| 277 |
+
|
| 278 |
+
def sentiment_analysis_interface(reviews_text: str):
|
| 279 |
+
"""Interface for sentiment analysis"""
|
| 280 |
+
if not reviews_text.strip():
|
| 281 |
+
return "Please enter some reviews to analyze.", None
|
| 282 |
+
|
| 283 |
+
reviews = process_reviews_input(reviews_text)
|
| 284 |
+
if not reviews:
|
| 285 |
+
return "No valid reviews found. Please check your input.", None
|
| 286 |
+
|
| 287 |
+
try:
|
| 288 |
+
result = analyzer.analyze_sentiment(reviews)
|
| 289 |
+
|
| 290 |
+
# Create visualization
|
| 291 |
+
fig = go.Figure(data=[
|
| 292 |
+
go.Bar(x=list(result['summary'].keys()),
|
| 293 |
+
y=list(result['summary'].values()),
|
| 294 |
+
marker_color=['green', 'red', 'gray'])
|
| 295 |
+
])
|
| 296 |
+
fig.update_layout(title="Sentiment Distribution", yaxis_title="Percentage")
|
| 297 |
+
|
| 298 |
+
return json.dumps(result, indent=2), fig
|
| 299 |
+
except Exception as e:
|
| 300 |
+
return f"Error: {str(e)}", None
|
| 301 |
+
|
| 302 |
+
def fake_detection_interface(reviews_text: str):
|
| 303 |
+
"""Interface for fake review detection"""
|
| 304 |
+
if not reviews_text.strip():
|
| 305 |
+
return "Please enter some reviews to analyze."
|
| 306 |
+
|
| 307 |
+
reviews = process_reviews_input(reviews_text)
|
| 308 |
+
if not reviews:
|
| 309 |
+
return "No valid reviews found. Please check your input."
|
| 310 |
+
|
| 311 |
+
try:
|
| 312 |
+
result = analyzer.detect_fake_reviews(reviews)
|
| 313 |
+
return json.dumps(result, indent=2)
|
| 314 |
+
except Exception as e:
|
| 315 |
+
return f"Error: {str(e)}"
|
| 316 |
+
|
| 317 |
+
def quality_assessment_interface(reviews_text: str):
|
| 318 |
+
"""Interface for quality assessment"""
|
| 319 |
+
if not reviews_text.strip():
|
| 320 |
+
return "Please enter some reviews to analyze."
|
| 321 |
+
|
| 322 |
+
reviews = process_reviews_input(reviews_text)
|
| 323 |
+
if not reviews:
|
| 324 |
+
return "No valid reviews found. Please check your input."
|
| 325 |
+
|
| 326 |
+
try:
|
| 327 |
+
result = analyzer.assess_quality(reviews)
|
| 328 |
+
return json.dumps(result, indent=2)
|
| 329 |
+
except Exception as e:
|
| 330 |
+
return f"Error: {str(e)}"
|
| 331 |
+
|
| 332 |
+
def competitor_comparison_interface(product_a_text: str, product_b_text: str):
|
| 333 |
+
"""Interface for competitor comparison"""
|
| 334 |
+
if not product_a_text.strip() or not product_b_text.strip():
|
| 335 |
+
return "Please enter reviews for both products.", None
|
| 336 |
+
|
| 337 |
+
reviews_a = process_reviews_input(product_a_text)
|
| 338 |
+
reviews_b = process_reviews_input(product_b_text)
|
| 339 |
+
|
| 340 |
+
if not reviews_a or not reviews_b:
|
| 341 |
+
return "Please provide valid reviews for both products.", None
|
| 342 |
+
|
| 343 |
+
try:
|
| 344 |
+
result, fig = analyzer.compare_competitors(reviews_a, reviews_b)
|
| 345 |
+
return json.dumps(result, indent=2), fig
|
| 346 |
+
except Exception as e:
|
| 347 |
+
return f"Error: {str(e)}", None
|
| 348 |
+
|
| 349 |
+
def generate_report_interface(analysis_result: str, report_type: str):
|
| 350 |
+
"""Interface for report generation"""
|
| 351 |
+
if not analysis_result.strip():
|
| 352 |
+
return "No analysis data available. Please run an analysis first."
|
| 353 |
+
|
| 354 |
+
try:
|
| 355 |
+
data = json.loads(analysis_result)
|
| 356 |
+
report = analyzer.generate_report(data, report_type.lower())
|
| 357 |
+
return report
|
| 358 |
+
except Exception as e:
|
| 359 |
+
return f"Error generating report: {str(e)}"
|
| 360 |
+
|
| 361 |
+
# Create Gradio interface
|
| 362 |
+
with gr.Blocks(title="SmartReview Pro", theme=gr.themes.Soft()) as demo:
|
| 363 |
+
gr.Markdown("# π SmartReview Pro")
|
| 364 |
+
gr.Markdown("Professional review analysis platform for e-commerce businesses")
|
| 365 |
+
|
| 366 |
+
with gr.Tab("π Sentiment Analysis"):
|
| 367 |
+
gr.Markdown("### Analyze customer sentiment from reviews")
|
| 368 |
+
with gr.Row():
|
| 369 |
+
with gr.Column():
|
| 370 |
+
sentiment_input = gr.Textbox(
|
| 371 |
+
lines=10,
|
| 372 |
+
placeholder="Enter reviews (one per line):\nGreat product, love it!\nTerrible quality, waste of money.\nOkay product, nothing special.",
|
| 373 |
+
label="Reviews"
|
| 374 |
+
)
|
| 375 |
+
sentiment_btn = gr.Button("Analyze Sentiment", variant="primary")
|
| 376 |
+
with gr.Column():
|
| 377 |
+
sentiment_output = gr.Textbox(label="Analysis Results", lines=15)
|
| 378 |
+
sentiment_chart = gr.Plot(label="Sentiment Distribution")
|
| 379 |
+
|
| 380 |
+
sentiment_btn.click(
|
| 381 |
+
sentiment_analysis_interface,
|
| 382 |
+
inputs=[sentiment_input],
|
| 383 |
+
outputs=[sentiment_output, sentiment_chart]
|
| 384 |
+
)
|
| 385 |
+
|
| 386 |
+
with gr.Tab("π Fake Review Detection"):
|
| 387 |
+
gr.Markdown("### Detect potentially fake or suspicious reviews")
|
| 388 |
+
with gr.Row():
|
| 389 |
+
with gr.Column():
|
| 390 |
+
fake_input = gr.Textbox(
|
| 391 |
+
lines=10,
|
| 392 |
+
placeholder="Enter reviews to check for authenticity...",
|
| 393 |
+
label="Reviews"
|
| 394 |
+
)
|
| 395 |
+
fake_btn = gr.Button("Detect Fake Reviews", variant="primary")
|
| 396 |
+
with gr.Column():
|
| 397 |
+
fake_output = gr.Textbox(label="Detection Results", lines=15)
|
| 398 |
+
|
| 399 |
+
fake_btn.click(
|
| 400 |
+
fake_detection_interface,
|
| 401 |
+
inputs=[fake_input],
|
| 402 |
+
outputs=[fake_output]
|
| 403 |
+
)
|
| 404 |
+
|
| 405 |
+
with gr.Tab("β Quality Assessment"):
|
| 406 |
+
gr.Markdown("### Assess the quality and helpfulness of reviews")
|
| 407 |
+
with gr.Row():
|
| 408 |
+
with gr.Column():
|
| 409 |
+
quality_input = gr.Textbox(
|
| 410 |
+
lines=10,
|
| 411 |
+
placeholder="Enter reviews to assess quality...",
|
| 412 |
+
label="Reviews"
|
| 413 |
+
)
|
| 414 |
+
quality_btn = gr.Button("Assess Quality", variant="primary")
|
| 415 |
+
with gr.Column():
|
| 416 |
+
quality_output = gr.Textbox(label="Quality Assessment", lines=15)
|
| 417 |
+
|
| 418 |
+
quality_btn.click(
|
| 419 |
+
quality_assessment_interface,
|
| 420 |
+
inputs=[quality_input],
|
| 421 |
+
outputs=[quality_output]
|
| 422 |
+
)
|
| 423 |
+
|
| 424 |
+
with gr.Tab("π Competitor Comparison"):
|
| 425 |
+
gr.Markdown("### Compare sentiment between competing products")
|
| 426 |
+
with gr.Row():
|
| 427 |
+
with gr.Column():
|
| 428 |
+
comp_product_a = gr.Textbox(
|
| 429 |
+
lines=8,
|
| 430 |
+
placeholder="Product A reviews...",
|
| 431 |
+
label="Product A Reviews"
|
| 432 |
+
)
|
| 433 |
+
comp_product_b = gr.Textbox(
|
| 434 |
+
lines=8,
|
| 435 |
+
placeholder="Product B reviews...",
|
| 436 |
+
label="Product B Reviews"
|
| 437 |
+
)
|
| 438 |
+
comp_btn = gr.Button("Compare Products", variant="primary")
|
| 439 |
+
with gr.Column():
|
| 440 |
+
comp_output = gr.Textbox(label="Comparison Results", lines=15)
|
| 441 |
+
comp_chart = gr.Plot(label="Comparison Chart")
|
| 442 |
+
|
| 443 |
+
comp_btn.click(
|
| 444 |
+
competitor_comparison_interface,
|
| 445 |
+
inputs=[comp_product_a, comp_product_b],
|
| 446 |
+
outputs=[comp_output, comp_chart]
|
| 447 |
+
)
|
| 448 |
+
|
| 449 |
+
with gr.Tab("π Report Generation"):
|
| 450 |
+
gr.Markdown("### Generate professional analysis reports")
|
| 451 |
+
with gr.Row():
|
| 452 |
+
with gr.Column():
|
| 453 |
+
report_data = gr.Textbox(
|
| 454 |
+
lines=10,
|
| 455 |
+
placeholder="Paste analysis results here...",
|
| 456 |
+
label="Analysis Data (JSON)"
|
| 457 |
+
)
|
| 458 |
+
report_type = gr.Dropdown(
|
| 459 |
+
choices=["sentiment", "fake", "quality"],
|
| 460 |
+
value="sentiment",
|
| 461 |
+
label="Report Type"
|
| 462 |
+
)
|
| 463 |
+
report_btn = gr.Button("Generate Report", variant="primary")
|
| 464 |
+
with gr.Column():
|
| 465 |
+
report_output = gr.Textbox(label="Generated Report", lines=15)
|
| 466 |
+
|
| 467 |
+
report_btn.click(
|
| 468 |
+
generate_report_interface,
|
| 469 |
+
inputs=[report_data, report_type],
|
| 470 |
+
outputs=[report_output]
|
| 471 |
+
)
|
| 472 |
+
|
| 473 |
+
with gr.Tab("βΉοΈ About"):
|
| 474 |
+
gr.Markdown("""
|
| 475 |
+
## SmartReview Pro Features
|
| 476 |
+
|
| 477 |
+
- **Sentiment Analysis**: Analyze customer emotions and opinions
|
| 478 |
+
- **Fake Review Detection**: Identify suspicious or inauthentic reviews
|
| 479 |
+
- **Quality Assessment**: Evaluate review helpfulness and detail
|
| 480 |
+
- **Competitor Comparison**: Compare sentiment across products
|
| 481 |
+
- **Professional Reports**: Generate detailed analysis reports
|
| 482 |
+
|
| 483 |
+
## Pricing Plans
|
| 484 |
+
- **Free**: 10 analyses per day
|
| 485 |
+
- **Pro ($299/month)**: 1000 analyses per day + advanced features
|
| 486 |
+
- **Enterprise**: Unlimited usage + API access + custom reports
|
| 487 |
+
|
| 488 |
+
Contact us for enterprise solutions and custom integrations.
|
| 489 |
+
""")
|
| 490 |
+
|
| 491 |
+
if __name__ == "__main__":
|
| 492 |
+
demo.launch()
|