Spaces:
Sleeping
Sleeping
Initial commit with full functionality
Browse files- app.py +61 -0
- requirements.txt +6 -0
- smart_web_analyzer.py +54 -0
- space.yml +5 -0
app.py
ADDED
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# app.py
|
2 |
+
import gradio as gr
|
3 |
+
from smart_web_analyzer import WebAnalyzer
|
4 |
+
|
5 |
+
analyzer = WebAnalyzer()
|
6 |
+
|
7 |
+
def format_results(results: dict) -> dict:
|
8 |
+
"""Format analysis results for Gradio tabs"""
|
9 |
+
outputs = {}
|
10 |
+
if 'error' in results:
|
11 |
+
return {"π Error": f"β {results['error']}"}
|
12 |
+
|
13 |
+
outputs["π Clean Text"] = results.get('clean_text', 'No text extracted')
|
14 |
+
|
15 |
+
if 'summary' in results:
|
16 |
+
outputs["π Summary"] = f"**AI Summary:**\n{results['summary']}"
|
17 |
+
|
18 |
+
if 'sentiment' in results:
|
19 |
+
outputs["π Sentiment"] = f"**Sentiment Score:**\n{results['sentiment']}"
|
20 |
+
|
21 |
+
if 'topics' in results:
|
22 |
+
topics = "\n".join([f"- **{k}**: {v:.2f}" for k,v in results['topics'].items()])
|
23 |
+
outputs["π Topics"] = f"**Detected Topics:**\n{topics}"
|
24 |
+
|
25 |
+
return outputs
|
26 |
+
|
27 |
+
with gr.Blocks(title="Smart Web Analyzer Plus") as demo:
|
28 |
+
gr.Markdown("# π Smart Web Analyzer Plus")
|
29 |
+
|
30 |
+
with gr.Row():
|
31 |
+
url_input = gr.Textbox(label="Enter URL", placeholder="https://example.com")
|
32 |
+
modes = gr.CheckboxGroup(["summarize", "sentiment", "topics"],
|
33 |
+
label="Analysis Types")
|
34 |
+
submit_btn = gr.Button("Analyze", variant="primary")
|
35 |
+
|
36 |
+
with gr.Tabs():
|
37 |
+
with gr.Tab("π Clean Text"):
|
38 |
+
clean_text = gr.Markdown()
|
39 |
+
with gr.Tab("π Summary"):
|
40 |
+
summary = gr.Markdown()
|
41 |
+
with gr.Tab("π Sentiment"):
|
42 |
+
sentiment = gr.Markdown()
|
43 |
+
with gr.Tab("π Topics"):
|
44 |
+
topics = gr.Markdown()
|
45 |
+
|
46 |
+
examples = gr.Examples(
|
47 |
+
examples=[
|
48 |
+
["https://www.bbc.com/news/technology-67881954", ["summarize", "sentiment"]],
|
49 |
+
["https://arxiv.org/html/2312.17296v1", ["topics", "summarize"]]
|
50 |
+
],
|
51 |
+
inputs=[url_input, modes]
|
52 |
+
)
|
53 |
+
|
54 |
+
submit_btn.click(
|
55 |
+
fn=lambda url, m: format_results(analyzer.analyze(url, m)),
|
56 |
+
inputs=[url_input, modes],
|
57 |
+
outputs=[clean_text, summary, sentiment, topics]
|
58 |
+
)
|
59 |
+
|
60 |
+
if __name__ == "__main__":
|
61 |
+
demo.launch()
|
requirements.txt
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# requirements.txt
|
2 |
+
gradio>=4.0.0
|
3 |
+
beautifulsoup4>=4.12.0
|
4 |
+
requests>=2.31.0
|
5 |
+
transformers>=4.40.0
|
6 |
+
torch>=2.2.0
|
smart_web_analyzer.py
ADDED
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# smart_web_analyzer.py
|
2 |
+
import requests
|
3 |
+
from bs4 import BeautifulSoup
|
4 |
+
from transformers import pipeline
|
5 |
+
import torch
|
6 |
+
|
7 |
+
class WebAnalyzer:
|
8 |
+
def __init__(self):
|
9 |
+
self.device = 0 if torch.cuda.is_available() else -1
|
10 |
+
self.models = {
|
11 |
+
'summarize': pipeline("summarization", model="facebook/bart-large-cnn"),
|
12 |
+
'sentiment': pipeline("text-classification",
|
13 |
+
model="nlptown/bert-base-multilingual-uncased-sentiment"),
|
14 |
+
'topics': pipeline("zero-shot-classification",
|
15 |
+
model="facebook/bart-large-mnli")
|
16 |
+
}
|
17 |
+
|
18 |
+
def fetch_content(self, url: str) -> str:
|
19 |
+
"""Fetch webpage content with custom headers"""
|
20 |
+
headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36'}
|
21 |
+
response = requests.get(url, headers=headers, timeout=15)
|
22 |
+
response.raise_for_status()
|
23 |
+
return response.text
|
24 |
+
|
25 |
+
def clean_html(self, html: str) -> str:
|
26 |
+
"""Basic HTML cleaning preserving all tags"""
|
27 |
+
soup = BeautifulSoup(html, 'html.parser')
|
28 |
+
return soup.prettify()
|
29 |
+
|
30 |
+
def analyze(self, url: str, modes: list) -> dict:
|
31 |
+
"""Core analysis pipeline"""
|
32 |
+
results = {}
|
33 |
+
try:
|
34 |
+
html = self.fetch_content(url)
|
35 |
+
results['clean_text'] = self.clean_html(html)
|
36 |
+
|
37 |
+
if 'summarize' in modes:
|
38 |
+
results['summary'] = self.models['summarize'](html, max_length=150)[0]['summary_text']
|
39 |
+
|
40 |
+
if 'sentiment' in modes:
|
41 |
+
sentiment = self.models['sentiment'](html[:512])[0]
|
42 |
+
results['sentiment'] = f"{sentiment['label']} ({sentiment['score']:.2f})"
|
43 |
+
|
44 |
+
if 'topics' in modes:
|
45 |
+
topics = self.models['topics'](html[:512],
|
46 |
+
candidate_labels=["Technology", "AI", "Business",
|
47 |
+
"Science", "Politics"])
|
48 |
+
results['topics'] = {topic: score for topic, score
|
49 |
+
in zip(topics['labels'], topics['scores'])}
|
50 |
+
|
51 |
+
except Exception as e:
|
52 |
+
results['error'] = str(e)
|
53 |
+
|
54 |
+
return results
|
space.yml
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# space.yml
|
2 |
+
title: Content Web Analyzer Plus
|
3 |
+
sdk: gradio
|
4 |
+
python:
|
5 |
+
version: "3.10"
|