Update app.py

app.py

@@ -1,166 +1,66 @@
+# Load model directly
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, AutoModelForSequenceClassification, TextClassificationPipeline
+import torch
 import gradio as gr
-import pandas as pd
-import numpy as np
-import random
+from openpyxl import load_workbook
+from numpy import mean
 
-def show_stars():
-    """ Function that shows the HTML script for stars -Janika"""
-    stars_html = """
-    <style>
-        /* Title */
-        .title-text {
-        color: white;
-        font-family: sans-serif;
-        font-size: 14px;
-        font-weight: bold;
-        }
+tokenizer = AutoTokenizer.from_pretrained("suriya7/bart-finetuned-text-summarization")
+model = AutoModelForSeq2SeqLM.from_pretrained("suriya7/bart-finetuned-text-summarization")
 
-        /* Empty stars */
-        .star-rating {
-            display: inline-block;
-            font-size: 1.5em;
-            color: lightgray;
-        }
+tokenizer_keywords = AutoTokenizer.from_pretrained("transformer3/H2-keywordextractor")
+model_keywords = AutoModelForSeq2SeqLM.from_pretrained("transformer3/H2-keywordextractor")
 
-        /* Filled stars */
-        .star-rating .star.filled {
-            color: gold;
-        }
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+# Load the fine-tuned model and tokenizer
+new_model = AutoModelForSequenceClassification.from_pretrained('roberta-rating')
+new_tokenizer = AutoTokenizer.from_pretrained('roberta-rating')
 
-        /* Box for the whole star rating component */
-        .rating-box {
-            border: 1px transparent;
-            background-color: #1f2937;
-            padding: 10px;
-            border-radius: 5px;
-            width: 100%;
-            margin-left: 0;
-        }
 
-        /* Box for the title */
-        .title-box {
-            border: 1px transparent;
-            background-color: #ca8a04;
-            padding: 3px;
-            border-radius: 8px;
-            display: inline-block;
-            margin-bottom: 6px;
-        }
+# Create a classification pipeline
+classifier = TextClassificationPipeline(model=new_model, tokenizer=new_tokenizer, device=device)
 
-        /* Box for stars */
-        .star-box {
-            border: 1px transparent;
-            background-color: #374151;
-            padding: 10px;
-            border-radius: 5px;
-            width: fit-content;
-            margin: 0;
-        }
-    </style>
+# Add label mapping for sentiment analysis
+label_mapping = {1: '1/5', 2: '2/5', 3: '3/5', 4: '4/5', 5: '5/5'}
 
-    <div class="rating-box">
-        <div class="title-box">
-            <p class="title-text">Star rating</p>
-        </div>
-        <div class="star-box">
-            <div class="star-rating" data-rating="4">
-                <span class="star">★</span>
-                <span class="star">★</span>
-                <span class="star">★</span>
-                <span class="star">★</span>
-                <span class="star">★</span>
-            </div>
-        </div>
-    </div>
+def parse_xl(file_path):
+    cells = []
 
-    <script>
-        document.addEventListener('DOMContentLoaded', function() {
-            const ratingElement = document.querySelector('.star-rating');
-            const rating = parseInt(ratingElement.getAttribute('data-rating'));
-            const stars = ratingElement.querySelectorAll('.star');
-            for (let i = 0; i < rating; i++) {
-                stars[i].classList.add('filled');
-            }
-        });
-    </script>
-    """
-    return stars_html
+    workbook = load_workbook(filename=file_path)
+    for sheet in workbook.worksheets:
+        for row in sheet.iter_rows():
+            for cell in row:
+                if cell.value != None:
+                    cells.append(cell.value)
 
-# Function to handle feedback submission -Piitu
-def handle_feedback(feedback, name):
-    return f"Thank you for your feedback, {name}!"
+    return cells
 
-# Function to update placeholder text -Piitu
-def update_placeholder(name):
-    if name:
-        return f"Enter your feedback here, {name}..."
-    else:
-        return "Enter your feedback here..."
+def evaluate(file):
+    reviews = parse_xl(file)
+    ratings = []
+    text = ""
 
-# Random data for testing, actual data added later
-df = pd.DataFrame({
-    'Year': np.random.randint(2000, 2024, 25),
-    'Reviews': np.random.randint(120, 320, 25),
-})
+    for review in reviews:
+        ratings.append(int(classifier(review)[0]['label'].split('_')[1]))
+        text += review
+        text += " "
+            
+    inputs = tokenizer([text], max_length=1024, truncation=True, return_tensors="pt")
+    summary_ids = model.generate(inputs["input_ids"], num_beams=2, min_length=50, max_length=1000)
+    summary = tokenizer.batch_decode(summary_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
 
-# Theme
-theme = gr.themes.Soft(
-    primary_hue="yellow",
-    secondary_hue="amber",
-    spacing_size="sm",
-    radius_size="lg",
-)
-
-with gr.Blocks(theme=theme) as demo:
-
-    # Basic user interface for company's view -Janika
-    with gr.Tab("User Interface"):
-        with gr.Row():
-            with gr.Column(scale=2, min_width=300):
-                # Overall summary from all feedback
-                com_summary_output = gr.Textbox(label="Summary")
-
-            with gr.Column(scale=1, min_width=300):
-                # Average star rating across all feedback
-                com_star_rating = gr.HTML(value=show_stars())
-
-                # The most common keywords across all feedback
-                com_keywords_output = gr.Textbox(label="Keywords")
-
-
-     # Testing Area -Piitu
-    with gr.Tab("Testing Area"):
-        with gr.Row():
-            with gr.Column(scale=1, min_width=300):
-                name_input = gr.Textbox(label="Enter your name", placeholder="Enter your name here...")
+    inputs_keywords = tokenizer_keywords([text], max_length=1024, truncation=True, return_tensors="pt")
+    summary_ids_keywords = model_keywords.generate(inputs_keywords["input_ids"], num_beams=2, min_length=0, max_length=100)
+    keywords = tokenizer_keywords.batch_decode(summary_ids_keywords, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]   
 
-                text_input = gr.Textbox(label="Please give us feedback!", 
-                                        placeholder="Enter your feedback here...", 
-                                        max_length=5000)
+    return round(mean(ratings), 2), summary, keywords
 
-                # Update placeholder based on name input -Piitu
-                name_input.change(fn=update_placeholder, inputs=name_input, outputs=text_input)
-
-                # Send button -Piitu
-                send_button = gr.Button("Send")
-
-                # Output for feedback submission confirmation -Piitu
-                feedback_output = gr.Textbox(label="Submission Result", interactive=False)
-
-                # Link button to function that handles feedback submission -Piitu
-                send_button.click(handle_feedback, inputs=[text_input, name_input], outputs=feedback_output)
-
-            with gr.Column(scale=2, min_width=300):
-                # Estimated star rating from customer's feedback -Janika
-                cus_star_rating = gr.HTML(value=show_stars())
-
-                # Estimated sentiment from customer's feedback -Janika
-                cus_sentiment = gr.Textbox(label="Sentiment")
-                
-                # Summary from customer's feedback -Janika
-                cus_summary_output = gr.Textbox(label="Summary")
-
-                # The most common keywords from customer's feedback -Janika
-                cus_keywords_output = gr.Textbox(label="Keywords")  
+iface = gr.Interface(
+    fn=evaluate,
+    inputs=gr.File(label="Reviews", file_types=[".xlsx", ".xlsm", ".xltx", ".xltm"]),
+    outputs=[gr.Textbox(label="Rating"), gr.Textbox(label="Summary"), gr.Textbox(label="Keywords")],
+    title='Summarize Reviews',
+    description="Evaluate and summarize collection of reviews. Reviews are submitted as an Excel file, where each reviews is in its own cell."
+)
 
-demo.launch()
+iface.launch(share=True)