Luciferalive commited on
Commit
8b672e9
1 Parent(s): 2538e49

Create app.py

Browse files
Files changed (1) hide show
  1. app.py +68 -0
app.py ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import gradio as gr
2
+ from transformers import pipeline
3
+
4
+ # Load the model
5
+ model_name = "knowledgator/comprehend_it-base"
6
+ classifier = pipeline("zero-shot-classification", model=model_name, device="cpu")
7
+
8
+ # Keywords associated with the "Value" label
9
+ value_keywords = [
10
+ "cheap", "expensive", "worth", "waste", "value for money", "overpriced", "bargain",
11
+ "affordable", "pricey", "costly", "economical", "deal", "rip-off", "budget-friendly",
12
+ "high-priced", "low-priced", "discounted", "premium", "luxurious", "inexpensive",
13
+ "priced right", "steal", "splurge", "bang for the buck", "investment", "saver",
14
+ "money's worth", "exorbitant", "reasonable", "unreasonable", "priced well",
15
+ "cost-effective", "overvalued", "undervalued", "fair price", "high cost", "low cost",
16
+ "good deal", "bad deal", "profitable", "loss", "savings", "spendy", "wallet-friendly",
17
+ "financially smart", "economic", "lavish", "modestly priced", "upscale", "downscale"
18
+ ]
19
+
20
+
21
+ # Function to check for value-related keywords in feedback
22
+ def contains_value_keywords(feedback_text):
23
+ for keyword in value_keywords:
24
+ if keyword in feedback_text.lower():
25
+ return True
26
+ return False
27
+
28
+ # Function to classify feedback
29
+ def classify_feedback(feedback_text):
30
+ # Classify feedback using the loaded model
31
+ labels = ["Value", "Facilities", "Experience", "Functionality", "Quality"]
32
+ result = classifier(feedback_text, labels, multi_label=True)
33
+
34
+ # Check for value-related keywords and adjust scores if necessary
35
+ if contains_value_keywords(feedback_text):
36
+ # Find the index of the "Value" label
37
+ try:
38
+ value_index = result["labels"].index("Value")
39
+ # Promote the score of the "Value" label
40
+ result["scores"][value_index] += 0.2 # Adjust the promotion strength as needed
41
+ # Ensure the score does not exceed 1
42
+ result["scores"][value_index] = min(result["scores"][value_index], 1.0)
43
+ except ValueError:
44
+ pass # "Value" label not in the top results
45
+
46
+ # Get the top two labels associated with the feedback, after possible adjustment
47
+ top_labels_scores = sorted(zip(result["labels"], result["scores"]), key=lambda x: x[1], reverse=True)[:2]
48
+ top_labels, scores = zip(*top_labels_scores)
49
+
50
+ # Generate HTML content for displaying the scores as meters/progress bars
51
+ html_content = ""
52
+ for i in range(len(top_labels)):
53
+ score_percentage = scores[i] * 100 # Convert score to percentage
54
+ html_content += f"<div><b>{top_labels[i]}:</b> {scores[i]:.2f} <div style='background-color: #e0e0e0; border-radius: 10px;'><div style='height: 24px; width: {score_percentage}%; background-color: #76b900; border-radius: 10px;'></div></div></div>"
55
+
56
+ return html_content
57
+
58
+ # Create Gradio interface
59
+ feedback_textbox = gr.Textbox(label="Enter your feedback:")
60
+ feedback_output = gr.HTML(label="Top 2 Labels with Scores:")
61
+
62
+ gr.Interface(
63
+ fn=classify_feedback,
64
+ inputs=feedback_textbox,
65
+ outputs=feedback_output,
66
+ title="Feedback Classifier",
67
+ description="Enter your feedback and get the top 2 associated labels with scores."
68
+ ).launch()