File size: 5,701 Bytes
818bad6
 
 
 
 
7a25c1c
818bad6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7a25c1c
818bad6
 
 
291f44e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
833bbde
818bad6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7a25c1c
 
 
818bad6
7a25c1c
 
818bad6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
291f44e
 
833bbde
291f44e
818bad6
291f44e
 
 
 
833bbde
291f44e
833bbde
291f44e
 
 
 
833bbde
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
163
164
165
166
167
import streamlit as st
import PyPDF2
import openai
import faiss
import os
import numpy as np
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity

# Function to extract text from a PDF file
def extract_text_from_pdf(pdf_file):
    reader = PyPDF2.PdfReader(pdf_file)
    text = ""
    for page in reader.pages:
        text += page.extract_text()
    return text

# Function to generate embeddings for a piece of text
def get_embeddings(text, model="text-embedding-ada-002"):
    response = openai.Embedding.create(input=[text], model=model)
    return response['data'][0]['embedding']

# Function to search for similar content
def search_similar(query_embedding, index, stored_texts, top_k=3):
    distances, indices = index.search(np.array([query_embedding]), top_k)
    results = [(stored_texts[i], distances[0][idx]) for idx, i in enumerate(indices[0])]
    return results

# Function to generate HTML with nice styling
def generate_html(response_content):
    html_template = f"""
    <!DOCTYPE html>
    <html lang="en">
    <head>
        <meta charset="UTF-8">
        <meta name="viewport" content="width=device-width, initial-scale=1.0">
        <title>Course Query Response</title>
        <style>
            body {{
                font-family: Arial, sans-serif;
                margin: 0;
                padding: 0;
                background-color: #f4f4f9;
                color: #333;
            }}
            .container {{
                width: 80%;
                margin: 30px auto;
                background-color: white;
                padding: 20px;
                border-radius: 8px;
                box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
            }}
            h1 {{
                color: #2C3E50;
                font-size: 2em;
                text-align: center;
            }}
            .response {{
                background-color: #ecf0f1;
                border-left: 5px solid #3498db;
                padding: 20px;
                font-size: 1.2em;
                margin-top: 20px;
                border-radius: 5px;
            }}
            footer {{
                text-align: center;
                margin-top: 30px;
                font-size: 0.9em;
                color: #7f8c8d;
            }}
        </style>
    </head>
    <body>
        <div class="container">
            <h1>Course Query Response</h1>
            <div class="response">
                <h3>Answer:</h3>
                <p>{response_content}</p>
            </div>
            <footer>
                <p>Generated by Course Query Assistant</p>
            </footer>
        </div>
    </body>
    </html>
    """
    return html_template

# Streamlit app starts here
st.title("Course Query Assistant")

# Input OpenAI API key
openai_api_key = st.text_input("Enter your OpenAI API key:", type="password")

if openai_api_key:
    openai.api_key = openai_api_key

    # Upload course materials
    uploaded_files = st.file_uploader("Upload Course Materials (PDFs)", type=["pdf"], accept_multiple_files=True)

    if uploaded_files:
        st.write("Processing uploaded course materials...")

        # Extract text and generate embeddings for all uploaded PDFs
        course_texts = []
        for uploaded_file in uploaded_files:
            text = extract_text_from_pdf(uploaded_file)
            course_texts.append(text)

        # Combine all course materials into one large text
        combined_text = " ".join(course_texts)

        # Split combined text into smaller chunks for embedding (max tokens ~1000)
        chunks = [combined_text[i:i+1000] for i in range(0, len(combined_text), 1000)]

        # Generate embeddings for all chunks
        embeddings = [get_embeddings(chunk) for chunk in chunks]

        # Convert the list of embeddings into a NumPy array (shape: [num_chunks, embedding_size])
        embeddings_np = np.array(embeddings).astype("float32")

        # Create a FAISS index for similarity search
        index = faiss.IndexFlatL2(len(embeddings_np[0]))  # Use the length of the embedding vectors for the dimension
        index.add(embeddings_np)

        st.write("Course materials have been processed and indexed.")

        # User query
        query = st.text_input("Enter your question about the course materials:")

        if query:
            # Generate embedding for the query
            query_embedding = get_embeddings(query)

            # Search for similar chunks in the FAISS index
            results = search_similar(query_embedding, index, chunks)

            # Create the context for the GPT prompt
            context = "\n".join([result[0] for result in results])
            modified_prompt = f"Context: {context}\n\nQuestion: {query}\n\nProvide a detailed answer based on the context."

            # Get the GPT-3.5-turbo response
            response = openai.ChatCompletion.create(
                model="gpt-3.5-turbo",
                messages=[{"role": "user", "content": modified_prompt}]
            )

            # Get the response content
            response_content = response['choices'][0]['message']['content']

            # Display the response in Streamlit
            st.write("### Intelligent Reply:")
            st.write(response_content)

            # Generate HTML content
            html_content = generate_html(response_content)

            # Provide the download button for the HTML file
            st.download_button(
                label="Download Response as HTML",
                data=html_content,
                file_name="course_query_response.html",
                mime="text/html"
            )