Update app.py
Browse files
app.py
CHANGED
|
@@ -7,6 +7,7 @@ import numpy as np
|
|
| 7 |
from sklearn.feature_extraction.text import TfidfVectorizer
|
| 8 |
from sklearn.metrics.pairwise import cosine_similarity
|
| 9 |
import html
|
|
|
|
| 10 |
|
| 11 |
# Function to extract text from a PDF file
|
| 12 |
def extract_text_from_pdf(pdf_file):
|
|
@@ -27,70 +28,22 @@ def search_similar(query_embedding, index, stored_texts, top_k=3):
|
|
| 27 |
results = [(stored_texts[i], distances[0][idx]) for idx, i in enumerate(indices[0])]
|
| 28 |
return results
|
| 29 |
|
| 30 |
-
# Function to generate
|
| 31 |
-
def
|
| 32 |
-
|
| 33 |
-
safe_content = html.escape(response_content)
|
| 34 |
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
background-color: #f4f4f9;
|
| 48 |
-
color: #333;
|
| 49 |
-
}}
|
| 50 |
-
.container {{
|
| 51 |
-
width: 80%;
|
| 52 |
-
margin: 30px auto;
|
| 53 |
-
background-color: white;
|
| 54 |
-
padding: 20px;
|
| 55 |
-
border-radius: 8px;
|
| 56 |
-
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
|
| 57 |
-
}}
|
| 58 |
-
h1 {{
|
| 59 |
-
color: #2C3E50;
|
| 60 |
-
font-size: 2em;
|
| 61 |
-
text-align: center;
|
| 62 |
-
}}
|
| 63 |
-
.response {{
|
| 64 |
-
background-color: #ecf0f1;
|
| 65 |
-
border-left: 5px solid #3498db;
|
| 66 |
-
padding: 20px;
|
| 67 |
-
font-size: 1.2em;
|
| 68 |
-
margin-top: 20px;
|
| 69 |
-
border-radius: 5px;
|
| 70 |
-
}}
|
| 71 |
-
footer {{
|
| 72 |
-
text-align: center;
|
| 73 |
-
margin-top: 30px;
|
| 74 |
-
font-size: 0.9em;
|
| 75 |
-
color: #7f8c8d;
|
| 76 |
-
}}
|
| 77 |
-
</style>
|
| 78 |
-
</head>
|
| 79 |
-
<body>
|
| 80 |
-
<div class="container">
|
| 81 |
-
<h1>Course Query Response</h1>
|
| 82 |
-
<div class="response">
|
| 83 |
-
<h3>Answer:</h3>
|
| 84 |
-
<p>{safe_content}</p>
|
| 85 |
-
</div>
|
| 86 |
-
<footer>
|
| 87 |
-
<p>Generated by Course Query Assistant</p>
|
| 88 |
-
</footer>
|
| 89 |
-
</div>
|
| 90 |
-
</body>
|
| 91 |
-
</html>
|
| 92 |
-
"""
|
| 93 |
-
return html_template
|
| 94 |
|
| 95 |
# Streamlit app starts here
|
| 96 |
st.title("Course Query Assistant")
|
|
@@ -145,9 +98,9 @@ if openai_api_key:
|
|
| 145 |
context = "\n".join([result[0] for result in results])
|
| 146 |
modified_prompt = f"Context: {context}\n\nQuestion: {query}\n\nProvide a detailed answer based on the context."
|
| 147 |
|
| 148 |
-
# Get the GPT-
|
| 149 |
response = openai.ChatCompletion.create(
|
| 150 |
-
model="gpt-4o-mini",
|
| 151 |
messages=[{"role": "user", "content": modified_prompt}]
|
| 152 |
)
|
| 153 |
|
|
@@ -158,13 +111,13 @@ if openai_api_key:
|
|
| 158 |
st.write("### Intelligent Reply:")
|
| 159 |
st.write(response_content)
|
| 160 |
|
| 161 |
-
# Generate
|
| 162 |
-
|
| 163 |
|
| 164 |
-
# Provide the download button for the
|
| 165 |
st.download_button(
|
| 166 |
-
label="Download Response as
|
| 167 |
-
data=
|
| 168 |
-
file_name="course_query_response.
|
| 169 |
-
mime="
|
| 170 |
)
|
|
|
|
| 7 |
from sklearn.feature_extraction.text import TfidfVectorizer
|
| 8 |
from sklearn.metrics.pairwise import cosine_similarity
|
| 9 |
import html
|
| 10 |
+
from docx import Document # New import for working with docx files
|
| 11 |
|
| 12 |
# Function to extract text from a PDF file
|
| 13 |
def extract_text_from_pdf(pdf_file):
|
|
|
|
| 28 |
results = [(stored_texts[i], distances[0][idx]) for idx, i in enumerate(indices[0])]
|
| 29 |
return results
|
| 30 |
|
| 31 |
+
# Function to generate a docx file with the response content
|
| 32 |
+
def generate_docx(response_content):
|
| 33 |
+
doc = Document()
|
|
|
|
| 34 |
|
| 35 |
+
# Adding title and response to the docx
|
| 36 |
+
doc.add_heading('Course Query Response', 0)
|
| 37 |
+
doc.add_heading('Answer:', level=1)
|
| 38 |
+
doc.add_paragraph(response_content)
|
| 39 |
+
|
| 40 |
+
# Save the document to a byte stream to allow for download in Streamlit
|
| 41 |
+
from io import BytesIO
|
| 42 |
+
doc_io = BytesIO()
|
| 43 |
+
doc.save(doc_io)
|
| 44 |
+
doc_io.seek(0)
|
| 45 |
+
|
| 46 |
+
return doc_io
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
|
| 48 |
# Streamlit app starts here
|
| 49 |
st.title("Course Query Assistant")
|
|
|
|
| 98 |
context = "\n".join([result[0] for result in results])
|
| 99 |
modified_prompt = f"Context: {context}\n\nQuestion: {query}\n\nProvide a detailed answer based on the context."
|
| 100 |
|
| 101 |
+
# Get the GPT-4 response
|
| 102 |
response = openai.ChatCompletion.create(
|
| 103 |
+
model="gpt-4o-mini", # Update to GPT-4 (or your desired model)
|
| 104 |
messages=[{"role": "user", "content": modified_prompt}]
|
| 105 |
)
|
| 106 |
|
|
|
|
| 111 |
st.write("### Intelligent Reply:")
|
| 112 |
st.write(response_content)
|
| 113 |
|
| 114 |
+
# Generate a docx file based on the exact response content
|
| 115 |
+
docx_file = generate_docx(response_content)
|
| 116 |
|
| 117 |
+
# Provide the download button for the .docx file
|
| 118 |
st.download_button(
|
| 119 |
+
label="Download Response as DOCX",
|
| 120 |
+
data=docx_file,
|
| 121 |
+
file_name="course_query_response.docx",
|
| 122 |
+
mime="application/vnd.openxmlformats-officedocument.wordprocessingml.document"
|
| 123 |
)
|