Update app.py
Browse files
app.py
CHANGED
@@ -1,13 +1,19 @@
|
|
1 |
-
|
2 |
-
from user_utils import *
|
3 |
-
|
4 |
import streamlit as st
|
5 |
from dotenv import load_dotenv
|
|
|
6 |
|
7 |
-
|
8 |
-
|
9 |
-
load_dotenv()
|
10 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
st.header("Automatic Ticket Classification Tool")
|
12 |
|
13 |
# Capture user input
|
@@ -16,7 +22,7 @@ def main():
|
|
16 |
|
17 |
if user_input:
|
18 |
try:
|
19 |
-
#
|
20 |
embeddings = create_embeddings()
|
21 |
|
22 |
# Fetch the Pinecone index using the API key and environment info
|
@@ -37,7 +43,7 @@ def main():
|
|
37 |
# Access the document content directly
|
38 |
content = getattr(doc, "page_content", "No content available.") # Safely access content
|
39 |
|
40 |
-
# Apply fine-tuned extraction model
|
41 |
relevant_info = fine_tune_extraction(content, user_input)
|
42 |
|
43 |
if relevant_info:
|
|
|
|
|
|
|
|
|
1 |
import streamlit as st
|
2 |
from dotenv import load_dotenv
|
3 |
+
from transformers import pipeline
|
4 |
|
5 |
+
# Load environment variables
|
6 |
+
load_dotenv()
|
|
|
7 |
|
8 |
+
# Initialize the question-answering pipeline with a pre-trained model
|
9 |
+
qa_pipeline = pipeline("question-answering", model="distilbert-base-uncased-distilled-squad")
|
10 |
+
|
11 |
+
def fine_tune_extraction(content, query):
|
12 |
+
# Use the pipeline to answer the question using the document content
|
13 |
+
result = qa_pipeline(question=query, context=content)
|
14 |
+
return result['answer']
|
15 |
+
|
16 |
+
def main():
|
17 |
st.header("Automatic Ticket Classification Tool")
|
18 |
|
19 |
# Capture user input
|
|
|
22 |
|
23 |
if user_input:
|
24 |
try:
|
25 |
+
# Assuming create_embeddings and pull_from_pinecone are defined elsewhere
|
26 |
embeddings = create_embeddings()
|
27 |
|
28 |
# Fetch the Pinecone index using the API key and environment info
|
|
|
43 |
# Access the document content directly
|
44 |
content = getattr(doc, "page_content", "No content available.") # Safely access content
|
45 |
|
46 |
+
# Apply fine-tuned extraction model to extract relevant information
|
47 |
relevant_info = fine_tune_extraction(content, user_input)
|
48 |
|
49 |
if relevant_info:
|