Jan90 commited on
Commit
c6158bd
·
verified ·
1 Parent(s): d837fec

Create app.py

Browse files
Files changed (1) hide show
  1. app.py +118 -0
app.py ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, AutoModelForSequenceClassification, TextClassificationPipeline
2
+ import torch
3
+ import gradio as gr
4
+ from openpyxl import load_workbook
5
+ from numpy import mean
6
+
7
+ # Load tokenizers and models
8
+ tokenizer = AutoTokenizer.from_pretrained("suriya7/bart-finetuned-text-summarization")
9
+ model = AutoModelForSeq2SeqLM.from_pretrained("suriya7/bart-finetuned-text-summarization")
10
+
11
+ tokenizer_keywords = AutoTokenizer.from_pretrained("transformer3/H2-keywordextractor")
12
+ model_keywords = AutoModelForSeq2SeqLM.from_pretrained("transformer3/H2-keywordextractor")
13
+
14
+ device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
15
+ new_model = AutoModelForSequenceClassification.from_pretrained('roberta-rating')
16
+ new_tokenizer = AutoTokenizer.from_pretrained('roberta-rating')
17
+
18
+ classifier = TextClassificationPipeline(model=new_model, tokenizer=new_tokenizer, device=device)
19
+
20
+ label_mapping = {1: '1/5', 2: '2/5', 3: '3/5', 4: '4/5', 5: '5/5'}
21
+
22
+ # Function to parse Excel file
23
+ def parse_xl(file_path):
24
+ cells = []
25
+
26
+ workbook = load_workbook(filename=file_path)
27
+ for sheet in workbook.worksheets:
28
+ for row in sheet.iter_rows():
29
+ for cell in row:
30
+ if cell.value != None:
31
+ cells.append(cell.value)
32
+
33
+ return cells
34
+
35
+ # Function to evaluate reviews from Excel file
36
+ def evaluate(file):
37
+ reviews = parse_xl(file)
38
+ ratings = []
39
+ text = ""
40
+ sentiments = []
41
+
42
+ for review in reviews:
43
+ rating = int(classifier(review)[0]['label'].split('_')[1])
44
+ ratings.append(rating)
45
+ text += review
46
+ text += " "
47
+
48
+ sentiment = classifier(review)[0]['label']
49
+ sentiment_label = "Positive" if sentiment == "LABEL_4" or sentiment == "LABEL_5" else "Negative" if sentiment == "LABEL_1" or sentiment == "LABEL_2" else "Neutral"
50
+ sentiments.append(sentiment_label)
51
+
52
+ overall_sentiment = "Positive" if sentiments.count("Positive") > sentiments.count("Negative") else "Negative" if sentiments.count("Negative") > sentiments.count("Positive") else "Neutral"
53
+
54
+ inputs = tokenizer([text], max_length=1024, truncation=True, return_tensors="pt")
55
+ summary_ids = model.generate(inputs["input_ids"], num_beams=2, min_length=10, max_length=50)
56
+ summary = tokenizer.batch_decode(summary_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
57
+
58
+ # Modify the summary to third person
59
+ summary = summary.replace("I", "He/She").replace("my", "his/her").replace("me", "him/her")
60
+
61
+ inputs_keywords = tokenizer_keywords([text], max_length=1024, truncation=True, return_tensors="pt")
62
+ summary_ids_keywords = model_keywords.generate(inputs_keywords["input_ids"], num_beams=2, min_length=0, max_length=100)
63
+ keywords = tokenizer_keywords.batch_decode(summary_ids_keywords, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
64
+
65
+ return round(mean(ratings), 2), summary, keywords, overall_sentiment
66
+
67
+ # Function to test a single text input
68
+ def test_area(text):
69
+ inputs = tokenizer([text], max_length=1024, truncation=True, return_tensors="pt")
70
+ summary_ids = model.generate(inputs["input_ids"], num_beams=2, min_length=10, max_length=50)
71
+ summary = tokenizer.batch_decode(summary_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
72
+
73
+ # Modify the summary to third person
74
+ summary = summary.replace("I", "He/She").replace("my", "his/her").replace("me", "him/her")
75
+
76
+ inputs_keywords = tokenizer_keywords([text], max_length=1024, truncation=True, return_tensors="pt")
77
+ summary_ids_keywords = model_keywords.generate(inputs_keywords["input_ids"], num_beams=2, min_length=0, max_length=100)
78
+ keywords = tokenizer_keywords.batch_decode(summary_ids_keywords, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
79
+
80
+ sentiment = classifier(text)[0]['label']
81
+ sentiment_label = "Positive" if sentiment == "LABEL_4" or sentiment == "LABEL_5" else "Negative" if sentiment == "LABEL_1" or sentiment == "LABEL_2" else "Neutral"
82
+
83
+ rating = int(classifier(text)[0]['label'].split('_')[1])
84
+
85
+ return rating, summary, keywords, sentiment_label
86
+
87
+ # Main interface
88
+ main_interface = gr.Interface(
89
+ fn=evaluate,
90
+ inputs=gr.File(label="Reviews"),
91
+ outputs=[gr.Textbox(label="Rating"), gr.Textbox(label="Summary"), gr.Textbox(label="Keywords"), gr.Textbox(label="Overall Sentiment")],
92
+ title='Summarize Reviews',
93
+ description="Evaluate and summarize collection of reviews. Reviews are submitted as an Excel file, where each review is in its own cell."
94
+ )
95
+
96
+ # Testing area interface
97
+ testing_interface = gr.Interface(
98
+ fn=test_area,
99
+ inputs=gr.Textbox(label="Input Text"),
100
+ outputs=[gr.Textbox(label="Rating"), gr.Textbox(label="Summary"), gr.Textbox(label="Keywords"), gr.Textbox(label="Sentiment")],
101
+ title='Testing Area',
102
+ description="Test the summarization, keyword extraction, sentiment analysis, and rating on custom text input."
103
+ )
104
+
105
+ # Combine interfaces into a tabbed interface with a sidebar
106
+ with gr.Blocks() as demo:
107
+ with gr.Row():
108
+ with gr.Column(scale=1):
109
+ gr.Markdown("## Sidebar")
110
+ gr.Button("Button 1")
111
+ gr.Button("Button 2")
112
+ with gr.Column(scale=4):
113
+ iface = gr.TabbedInterface(
114
+ [main_interface, testing_interface],
115
+ ["Summarize Reviews", "Testing Area"]
116
+ )
117
+
118
+ demo.launch(share=True)