YixuanWang
commited on
Update app.py
Browse files
app.py
CHANGED
@@ -8,6 +8,8 @@ from typing import List, Dict, Tuple
|
|
8 |
from dataclasses import dataclass
|
9 |
from pathlib import Path
|
10 |
import logging
|
|
|
|
|
11 |
|
12 |
# Configure logging
|
13 |
logging.basicConfig(level=logging.INFO)
|
@@ -29,7 +31,7 @@ class TweetPreprocessor:
|
|
29 |
"""Load and validate the dataset."""
|
30 |
try:
|
31 |
data = pd.read_csv(data_path)
|
32 |
-
required_columns = {'Text', 'Retweets', 'Likes'}
|
33 |
if not required_columns.issubset(data.columns):
|
34 |
raise ValueError(f"Missing required columns: {required_columns - set(data.columns)}")
|
35 |
return data
|
@@ -37,12 +39,40 @@ class TweetPreprocessor:
|
|
37 |
logger.error(f"Error loading data: {e}")
|
38 |
raise
|
39 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
def calculate_metrics(self) -> pd.DataFrame:
|
41 |
"""Calculate sentiment and popularity metrics."""
|
42 |
-
|
|
|
|
|
|
|
|
|
|
|
43 |
self.data['Popularity'] = self._normalize_popularity()
|
|
|
|
|
|
|
44 |
return self.data
|
45 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
46 |
@staticmethod
|
47 |
def _get_sentiment(text: str) -> float:
|
48 |
"""Calculate sentiment polarity for a text."""
|
@@ -55,45 +85,7 @@ class TweetPreprocessor:
|
|
55 |
def _normalize_popularity(self) -> pd.Series:
|
56 |
"""Normalize popularity scores using min-max scaling."""
|
57 |
popularity = self.data['Retweets'] + self.data['Likes']
|
58 |
-
return (popularity - popularity.
|
59 |
-
|
60 |
-
class FakeNewsClassifier:
|
61 |
-
def __init__(self, model_name: str):
|
62 |
-
"""Initialize the fake news classifier."""
|
63 |
-
self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
64 |
-
self.model_name = model_name
|
65 |
-
self.model, self.tokenizer = self._load_model()
|
66 |
-
|
67 |
-
def _load_model(self) -> Tuple[AutoModelForSequenceClassification, AutoTokenizer]:
|
68 |
-
"""Load the model and tokenizer."""
|
69 |
-
try:
|
70 |
-
tokenizer = AutoTokenizer.from_pretrained(self.model_name)
|
71 |
-
model = AutoModelForSequenceClassification.from_pretrained(self.model_name).to(self.device)
|
72 |
-
return model, tokenizer
|
73 |
-
except Exception as e:
|
74 |
-
logger.error(f"Error loading model: {e}")
|
75 |
-
raise
|
76 |
-
|
77 |
-
@torch.no_grad()
|
78 |
-
def predict_batch(self, texts: List[str], batch_size: int = 100) -> np.ndarray:
|
79 |
-
"""Predict fake news probability for a batch of texts."""
|
80 |
-
predictions = []
|
81 |
-
|
82 |
-
for i in range(0, len(texts), batch_size):
|
83 |
-
batch_texts = texts[i:i + batch_size]
|
84 |
-
inputs = self.tokenizer(
|
85 |
-
batch_texts,
|
86 |
-
return_tensors="pt",
|
87 |
-
padding=True,
|
88 |
-
truncation=True,
|
89 |
-
max_length=128
|
90 |
-
).to(self.device)
|
91 |
-
|
92 |
-
outputs = self.model(**inputs)
|
93 |
-
batch_predictions = outputs.logits.argmax(dim=1).cpu().numpy()
|
94 |
-
predictions.extend(batch_predictions)
|
95 |
-
|
96 |
-
return np.array(predictions)
|
97 |
|
98 |
class RecommendationSystem:
|
99 |
def __init__(self, data_path: Path, model_name: str):
|
@@ -106,13 +98,13 @@ class RecommendationSystem:
|
|
106 |
def setup_system(self):
|
107 |
"""Set up the recommendation system."""
|
108 |
self.data = self.preprocessor.calculate_metrics()
|
109 |
-
predictions = self.classifier.predict_batch(self.data['
|
110 |
self.data['Credibility'] = [1 if pred == 1 else -1 for pred in predictions]
|
111 |
|
112 |
-
def get_recommendations(self, weights: RecommendationWeights, num_recommendations: int = 10) ->
|
113 |
"""Get tweet recommendations based on weights."""
|
114 |
if not self._validate_weights(weights):
|
115 |
-
return "
|
116 |
|
117 |
normalized_weights = self._normalize_weights(weights)
|
118 |
|
@@ -120,7 +112,7 @@ class RecommendationSystem:
|
|
120 |
self.data['Credibility'] * normalized_weights.visibility +
|
121 |
self.data['Sentiment'] * normalized_weights.sentiment +
|
122 |
self.data['Popularity'] * normalized_weights.popularity
|
123 |
-
)
|
124 |
|
125 |
top_recommendations = (
|
126 |
self.data.nlargest(100, 'Final_Score')
|
@@ -129,49 +121,92 @@ class RecommendationSystem:
|
|
129 |
|
130 |
return self._format_recommendations(top_recommendations)
|
131 |
|
132 |
-
|
133 |
-
|
134 |
-
|
135 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
136 |
|
137 |
@staticmethod
|
138 |
-
def
|
139 |
-
"""
|
140 |
-
|
141 |
-
|
142 |
-
|
143 |
-
|
144 |
-
|
145 |
-
sentiment=weights.sentiment / total,
|
146 |
-
popularity=weights.popularity / total
|
147 |
-
)
|
148 |
|
149 |
@staticmethod
|
150 |
-
def
|
151 |
-
"""
|
152 |
-
return
|
153 |
-
|
154 |
-
|
155 |
-
|
|
|
|
|
156 |
|
157 |
def create_gradio_interface(recommendation_system: RecommendationSystem) -> gr.Interface:
|
158 |
"""Create and configure the Gradio interface."""
|
159 |
-
|
160 |
-
|
161 |
-
|
162 |
-
|
163 |
-
|
164 |
-
|
165 |
-
|
166 |
-
|
167 |
-
|
168 |
-
|
169 |
-
|
170 |
-
|
171 |
-
|
172 |
-
|
173 |
-
|
174 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
175 |
|
176 |
def main():
|
177 |
"""Main function to run the application."""
|
@@ -180,8 +215,8 @@ def main():
|
|
180 |
data_path=Path('twitter_dataset.csv'),
|
181 |
model_name="hamzab/roberta-fake-news-classification"
|
182 |
)
|
183 |
-
|
184 |
-
|
185 |
except Exception as e:
|
186 |
logger.error(f"Application failed to start: {e}")
|
187 |
raise
|
|
|
8 |
from dataclasses import dataclass
|
9 |
from pathlib import Path
|
10 |
import logging
|
11 |
+
import re
|
12 |
+
from datetime import datetime
|
13 |
|
14 |
# Configure logging
|
15 |
logging.basicConfig(level=logging.INFO)
|
|
|
31 |
"""Load and validate the dataset."""
|
32 |
try:
|
33 |
data = pd.read_csv(data_path)
|
34 |
+
required_columns = {'Text', 'Retweets', 'Likes', 'Timestamp'} # 添加时间戳列
|
35 |
if not required_columns.issubset(data.columns):
|
36 |
raise ValueError(f"Missing required columns: {required_columns - set(data.columns)}")
|
37 |
return data
|
|
|
39 |
logger.error(f"Error loading data: {e}")
|
40 |
raise
|
41 |
|
42 |
+
def _clean_text(self, text: str) -> str:
|
43 |
+
"""清理文本内容,移除无意义的内容"""
|
44 |
+
if pd.isna(text) or len(str(text).strip()) < 10: # 排除过短或空的文本
|
45 |
+
return ""
|
46 |
+
|
47 |
+
# 移除URL
|
48 |
+
text = re.sub(r'http\S+|www.\S+', '', str(text))
|
49 |
+
# 移除特殊字符
|
50 |
+
text = re.sub(r'[^\w\s]', '', text)
|
51 |
+
# 移除多余空格
|
52 |
+
text = ' '.join(text.split())
|
53 |
+
return text
|
54 |
+
|
55 |
def calculate_metrics(self) -> pd.DataFrame:
|
56 |
"""Calculate sentiment and popularity metrics."""
|
57 |
+
# 清理文本
|
58 |
+
self.data['Clean_Text'] = self.data['Text'].apply(self._clean_text)
|
59 |
+
# 过滤掉无效的文本
|
60 |
+
self.data = self.data[self.data['Clean_Text'].str.len() > 0]
|
61 |
+
|
62 |
+
self.data['Sentiment'] = self.data['Clean_Text'].apply(self._get_sentiment)
|
63 |
self.data['Popularity'] = self._normalize_popularity()
|
64 |
+
|
65 |
+
# 添加时间衰减因子
|
66 |
+
self.data['Time_Weight'] = self._calculate_time_weight()
|
67 |
return self.data
|
68 |
|
69 |
+
def _calculate_time_weight(self) -> pd.Series:
|
70 |
+
"""计算时间权重,越新的内容权重越高"""
|
71 |
+
current_time = datetime.now()
|
72 |
+
self.data['Timestamp'] = pd.to_datetime(self.data['Timestamp'])
|
73 |
+
time_diff = (current_time - self.data['Timestamp']).dt.total_seconds()
|
74 |
+
return np.exp(-time_diff / (7 * 24 * 3600)) # 7天的衰减周期
|
75 |
+
|
76 |
@staticmethod
|
77 |
def _get_sentiment(text: str) -> float:
|
78 |
"""Calculate sentiment polarity for a text."""
|
|
|
85 |
def _normalize_popularity(self) -> pd.Series:
|
86 |
"""Normalize popularity scores using min-max scaling."""
|
87 |
popularity = self.data['Retweets'] + self.data['Likes']
|
88 |
+
return (popularity - popularity.min()) / (popularity.max() - popularity.min() + 1e-6)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
89 |
|
90 |
class RecommendationSystem:
|
91 |
def __init__(self, data_path: Path, model_name: str):
|
|
|
98 |
def setup_system(self):
|
99 |
"""Set up the recommendation system."""
|
100 |
self.data = self.preprocessor.calculate_metrics()
|
101 |
+
predictions = self.classifier.predict_batch(self.data['Clean_Text'].tolist())
|
102 |
self.data['Credibility'] = [1 if pred == 1 else -1 for pred in predictions]
|
103 |
|
104 |
+
def get_recommendations(self, weights: RecommendationWeights, num_recommendations: int = 10) -> Dict:
|
105 |
"""Get tweet recommendations based on weights."""
|
106 |
if not self._validate_weights(weights):
|
107 |
+
return {"error": "Invalid weights provided"}
|
108 |
|
109 |
normalized_weights = self._normalize_weights(weights)
|
110 |
|
|
|
112 |
self.data['Credibility'] * normalized_weights.visibility +
|
113 |
self.data['Sentiment'] * normalized_weights.sentiment +
|
114 |
self.data['Popularity'] * normalized_weights.popularity
|
115 |
+
) * self.data['Time_Weight'] # 考虑时间因素
|
116 |
|
117 |
top_recommendations = (
|
118 |
self.data.nlargest(100, 'Final_Score')
|
|
|
121 |
|
122 |
return self._format_recommendations(top_recommendations)
|
123 |
|
124 |
+
def _format_recommendations(self, recommendations: pd.DataFrame) -> Dict:
|
125 |
+
"""Format recommendations for display."""
|
126 |
+
formatted_results = []
|
127 |
+
for _, row in recommendations.iterrows():
|
128 |
+
score_details = {
|
129 |
+
"总分": f"{row['Final_Score']:.2f}",
|
130 |
+
"可信度": "可信" if row['Credibility'] > 0 else "存疑",
|
131 |
+
"情感倾向": self._get_sentiment_label(row['Sentiment']),
|
132 |
+
"热度": f"{row['Popularity']:.2f}",
|
133 |
+
"互动数": f"点赞 {row['Likes']} · 转发 {row['Retweets']}"
|
134 |
+
}
|
135 |
+
|
136 |
+
formatted_results.append({
|
137 |
+
"text": row['Clean_Text'],
|
138 |
+
"scores": score_details,
|
139 |
+
"timestamp": row['Timestamp'].strftime("%Y-%m-%d %H:%M")
|
140 |
+
})
|
141 |
+
|
142 |
+
return {
|
143 |
+
"recommendations": formatted_results,
|
144 |
+
"score_explanation": self._get_score_explanation()
|
145 |
+
}
|
146 |
|
147 |
@staticmethod
|
148 |
+
def _get_sentiment_label(sentiment_score: float) -> str:
|
149 |
+
"""Convert sentiment score to human-readable label."""
|
150 |
+
if sentiment_score > 0.3:
|
151 |
+
return "积极"
|
152 |
+
elif sentiment_score < -0.3:
|
153 |
+
return "消极"
|
154 |
+
return "中性"
|
|
|
|
|
|
|
155 |
|
156 |
@staticmethod
|
157 |
+
def _get_score_explanation() -> Dict[str, str]:
|
158 |
+
"""Provide explanation for different score components."""
|
159 |
+
return {
|
160 |
+
"可信度": "基于机器学习模型对内容可信度的评估",
|
161 |
+
"情感倾向": "文本的情感倾向分析结果",
|
162 |
+
"热度": "根据点赞和转发数量计算的归一化热度分数",
|
163 |
+
"时间权重": "考虑内容时效性的权重因子"
|
164 |
+
}
|
165 |
|
166 |
def create_gradio_interface(recommendation_system: RecommendationSystem) -> gr.Interface:
|
167 |
"""Create and configure the Gradio interface."""
|
168 |
+
with gr.Blocks(theme=gr.themes.Soft()) as interface:
|
169 |
+
gr.Markdown("""
|
170 |
+
# 推文推荐系统
|
171 |
+
|
172 |
+
这个系统通过多个维度来为您推荐高质量的推文:
|
173 |
+
- **可信度**: 评估内容的可靠性
|
174 |
+
- **情感倾向**: 分析文本的情感色彩
|
175 |
+
- **热度**: 考虑内容的受欢迎程度
|
176 |
+
- **时效性**: 优先推荐较新的内容
|
177 |
+
""")
|
178 |
+
|
179 |
+
with gr.Row():
|
180 |
+
with gr.Column(scale=1):
|
181 |
+
visibility_weight = gr.Slider(
|
182 |
+
0, 1, 0.5,
|
183 |
+
label="可信度权重",
|
184 |
+
info="调整对内容可信度的重视程度"
|
185 |
+
)
|
186 |
+
sentiment_weight = gr.Slider(
|
187 |
+
0, 1, 0.3,
|
188 |
+
label="情感倾向权重",
|
189 |
+
info="调整对情感倾向的重视程度"
|
190 |
+
)
|
191 |
+
popularity_weight = gr.Slider(
|
192 |
+
0, 1, 0.2,
|
193 |
+
label="热度权重",
|
194 |
+
info="调整对内容热度的重视程度"
|
195 |
+
)
|
196 |
+
submit_btn = gr.Button("获取推荐", variant="primary")
|
197 |
+
|
198 |
+
with gr.Column(scale=2):
|
199 |
+
output = gr.JSON(label="推荐结果")
|
200 |
+
|
201 |
+
submit_btn.click(
|
202 |
+
fn=lambda v, s, p: recommendation_system.get_recommendations(
|
203 |
+
RecommendationWeights(v, s, p)
|
204 |
+
),
|
205 |
+
inputs=[visibility_weight, sentiment_weight, popularity_weight],
|
206 |
+
outputs=output
|
207 |
+
)
|
208 |
+
|
209 |
+
return interface
|
210 |
|
211 |
def main():
|
212 |
"""Main function to run the application."""
|
|
|
215 |
data_path=Path('twitter_dataset.csv'),
|
216 |
model_name="hamzab/roberta-fake-news-classification"
|
217 |
)
|
218 |
+
interface = create_gradio_interface(recommendation_system)
|
219 |
+
interface.launch()
|
220 |
except Exception as e:
|
221 |
logger.error(f"Application failed to start: {e}")
|
222 |
raise
|