Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -4,7 +4,7 @@ from pinecone import Pinecone
|
|
4 |
from sentence_transformers import SentenceTransformer
|
5 |
|
6 |
# Title of the Streamlit App
|
7 |
-
st.title("
|
8 |
|
9 |
# Initialize Pinecone globally
|
10 |
index = None
|
@@ -14,7 +14,8 @@ def initialize_pinecone():
|
|
14 |
api_key = os.getenv('PINECONE_API_KEY') # Get Pinecone API key from environment variable
|
15 |
if api_key:
|
16 |
# Initialize Pinecone client using the new class instance method
|
17 |
-
|
|
|
18 |
else:
|
19 |
st.error("Pinecone API key not found! Please set the PINECONE_API_KEY environment variable.")
|
20 |
return None
|
@@ -22,9 +23,10 @@ def initialize_pinecone():
|
|
22 |
# Function to connect to the 'pubmed-splade' index
|
23 |
def connect_to_index(pc):
|
24 |
index_name = 'pubmed-splade' # Hardcoded index name
|
|
|
25 |
if index_name in pc.list_indexes().names():
|
26 |
-
|
27 |
-
return
|
28 |
else:
|
29 |
st.error(f"Index '{index_name}' not found!")
|
30 |
return None
|
@@ -38,9 +40,10 @@ pc = initialize_pinecone()
|
|
38 |
|
39 |
# If Pinecone initialized successfully, proceed with index management
|
40 |
if pc:
|
|
|
41 |
index = connect_to_index(pc)
|
42 |
|
43 |
-
#
|
44 |
model = SentenceTransformer('msmarco-bert-base-dot-v5')
|
45 |
|
46 |
# Query input
|
@@ -50,25 +53,18 @@ if pc:
|
|
50 |
if st.button("Search Query"):
|
51 |
if query_text and index:
|
52 |
dense_vector = encode_query(model, query_text)
|
53 |
-
|
54 |
# Search the index
|
55 |
results = index.query(
|
56 |
vector=dense_vector,
|
57 |
-
top_k=
|
58 |
include_metadata=True
|
59 |
)
|
60 |
|
61 |
st.write("### Search Results:")
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
# Display score and context in a formatted way
|
68 |
-
st.markdown(f"**Score**: `{score}`")
|
69 |
-
st.markdown(f"**Context**: {context}")
|
70 |
-
st.markdown("---") # Divider for each result
|
71 |
-
else:
|
72 |
-
st.warning("No results found for this query.")
|
73 |
else:
|
74 |
st.error("Please enter a query and ensure the index is initialized.")
|
|
|
4 |
from sentence_transformers import SentenceTransformer
|
5 |
|
6 |
# Title of the Streamlit App
|
7 |
+
st.title("Medical Hybrid Search")
|
8 |
|
9 |
# Initialize Pinecone globally
|
10 |
index = None
|
|
|
14 |
api_key = os.getenv('PINECONE_API_KEY') # Get Pinecone API key from environment variable
|
15 |
if api_key:
|
16 |
# Initialize Pinecone client using the new class instance method
|
17 |
+
pc = Pinecone(api_key=api_key)
|
18 |
+
return pc
|
19 |
else:
|
20 |
st.error("Pinecone API key not found! Please set the PINECONE_API_KEY environment variable.")
|
21 |
return None
|
|
|
23 |
# Function to connect to the 'pubmed-splade' index
|
24 |
def connect_to_index(pc):
|
25 |
index_name = 'pubmed-splade' # Hardcoded index name
|
26 |
+
# Connect to the 'pubmed-splade' index
|
27 |
if index_name in pc.list_indexes().names():
|
28 |
+
index = pc.Index(index_name)
|
29 |
+
return index
|
30 |
else:
|
31 |
st.error(f"Index '{index_name}' not found!")
|
32 |
return None
|
|
|
40 |
|
41 |
# If Pinecone initialized successfully, proceed with index management
|
42 |
if pc:
|
43 |
+
# Connect directly to 'pubmed-splade' index
|
44 |
index = connect_to_index(pc)
|
45 |
|
46 |
+
# Model for query encoding
|
47 |
model = SentenceTransformer('msmarco-bert-base-dot-v5')
|
48 |
|
49 |
# Query input
|
|
|
53 |
if st.button("Search Query"):
|
54 |
if query_text and index:
|
55 |
dense_vector = encode_query(model, query_text)
|
56 |
+
|
57 |
# Search the index
|
58 |
results = index.query(
|
59 |
vector=dense_vector,
|
60 |
+
top_k=3,
|
61 |
include_metadata=True
|
62 |
)
|
63 |
|
64 |
st.write("### Search Results:")
|
65 |
+
for match in results.matches:
|
66 |
+
st.markdown(f"#### Score: **{match.score:.4f}**")
|
67 |
+
st.write(f"**Context:** {match.metadata.get('context', 'No context available.')}")
|
68 |
+
st.write("---")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
69 |
else:
|
70 |
st.error("Please enter a query and ensure the index is initialized.")
|