File size: 6,408 Bytes
c1ff2ef
 
 
 
 
 
 
 
 
 
 
 
 
 
bfe5bbd
c1ff2ef
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bfe5bbd
 
c1ff2ef
 
bfe5bbd
 
 
c1ff2ef
bfe5bbd
c1ff2ef
 
 
bfe5bbd
 
c1ff2ef
 
 
bfe5bbd
 
 
 
 
 
 
 
c1ff2ef
 
 
 
 
bfe5bbd
c1ff2ef
bfe5bbd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c1ff2ef
 
bfe5bbd
c1ff2ef
bfe5bbd
 
c1ff2ef
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
import os
import logging
import json
import gradio as gr
import pandas as pd
from datasets import load_dataset
import random
from openai import OpenAI
from typing import List, Tuple, Dict
from dotenv import load_dotenv
from transformers import pipeline
import asyncio

# Import the required functions from the pipeline file
from pipeline_gradio_experimental import generate_single_question, rank_questions_with_details

# Set up logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# Load environment variables
load_dotenv()

# Initialize OpenAI client
client = OpenAI(api_key=os.environ.get("OPENAI_API_KEY"))

# Load the SQuAD dataset
dataset = load_dataset("squad")

# Initialize the question answering pipeline
qa_pipeline = pipeline("question-answering", model="deepset/roberta-base-squad2")

def get_random_entry():
    random_index = random.randint(0, len(dataset['train']) - 1)
    entry = dataset['train'][random_index]
    return entry['context'], entry['answers']['text'][0], entry['question']

def generate_answer(context: str, question: str) -> str:
    try:
        result = qa_pipeline(question=question, context=context)
        return result['answer']
    except Exception as e:
        logger.error(f"Error in generate_answer: {e}")
        return "Failed to generate answer"

def compare_questions(context: str, original_answer: str, question1: str, answer1: str, question2: str, answer2: str) -> Dict[str, any]:
    try:
        response = client.chat.completions.create(
            model="gpt-4o-2024-08-06",
            messages=[
                {"role": "system", "content": "You are an expert in evaluating question-answer pairs based on a given context."},
                {"role": "user", "content": f"""Compare the following two question-answer pairs based on the given context and original answer. Evaluate their quality and relevance.

Context: {context}
Original Answer: {original_answer}

Question 1: {question1}
Answer 1: {answer1}

Question 2: {question2}
Answer 2: {answer2}

Score each question-answer pair on a scale of 0 to 10 based on the quality and relevance of the question and answer. Provide an explanation for your evaluation. Focus on how well the new answer matches the old answer considering the context. Make sure to grade one higher than the other."""}
            ],
            response_format={
                "type": "json_schema",
                "json_schema": {
                    "name": "question_comparison_evaluator",
                    "strict": True,
                    "schema": {
                        "type": "object",
                        "properties": {
                            "question1_score": {"type": "number"},
                            "question2_score": {"type": "number"},
                            "explanation": {"type": "string"}
                        },
                        "required": ["question1_score", "question2_score", "explanation"],
                        "additionalProperties": False
                    }
                }
            }
        )
        return json.loads(response.choices[0].message.content)
    except Exception as e:
        logger.error(f"Error in comparing questions: {e}")
        return {"question1_score": 0, "question2_score": 0, "explanation": "Failed to compare questions"}

async def process_entry(context, answer, progress=gr.Progress()):
    progress(0, desc="Starting process...")
    await asyncio.sleep(1)
    
    progress(0.2, desc="Generating questions...")
    basic_question = generate_single_question(context, answer, [])
    _, _, enhanced_question = rank_questions_with_details(context, answer)
    
    progress(0.4, desc="Generating answers...")
    basic_answer = generate_answer(context, basic_question)
    enhanced_answer = generate_answer(context, enhanced_question)
    
    progress(0.6, desc="Comparing questions...")
    comparison_result = compare_questions(context, answer, basic_question, basic_answer, enhanced_question, enhanced_answer)
    
    winner = "Basic" if comparison_result["question1_score"] > comparison_result["question2_score"] else "Enhanced"
    
    progress(1.0, desc="Process complete!")
    return (
        f"Question: {basic_question}\nAnswer: {basic_answer}",
        f"Question: {enhanced_question}\nAnswer: {enhanced_answer}",
        f"Question 1 Score: {comparison_result['question1_score']}\n"
        f"Question 2 Score: {comparison_result['question2_score']}\n"
        f"Explanation: {comparison_result['explanation']}\n"
        f"Winner: {winner} Generation"
    )

# Create Gradio interface
with gr.Blocks(theme=gr.themes.Default()) as iface:
    gr.Markdown("# Question Generation and Comparison")
    gr.Markdown("Enter a context and answer, or click 'Random' to get a random entry from the SQuAD dataset.")
    
    with gr.Row():
        with gr.Column(scale=2):
            context_input = gr.Textbox(label="Context", lines=10)
            answer_input = gr.Textbox(label="Answer", lines=2)
            with gr.Row():
                submit_button = gr.Button("Submit")
                random_button = gr.Button("Random")
        
        with gr.Column(scale=3):
            original_question_output = gr.Textbox(label="Original Question from Dataset", lines=2)
            basic_generation_output = gr.Textbox(label="Basic Generation", lines=4)
            enhanced_generation_output = gr.Textbox(label="Enhanced Generation", lines=4)
            comparison_result_output = gr.Textbox(label="Comparison Result", lines=6)

    async def on_submit(context, answer):
        return await process_entry(context, answer)

    async def on_random():
        context, answer, question = get_random_entry()
        results = await process_entry(context, answer)
        return [context, answer, question] + list(results)

    submit_button.click(
        fn=on_submit,
        inputs=[context_input, answer_input],
        outputs=[basic_generation_output, enhanced_generation_output, comparison_result_output]
    )

    random_button.click(
        fn=on_random,
        outputs=[
            context_input,
            answer_input,
            original_question_output,
            basic_generation_output,
            enhanced_generation_output,
            comparison_result_output
        ]
    )

# Launch the app
if __name__ == "__main__":
    iface.launch()