Upload 4 files

app.py

@@ -0,0 +1,58 @@
+import streamlit as st
+import pandas as pd
+from safetensors import safe_open
+from sentence_transformers import SentenceTransformer, CrossEncoder, util
+import pickle
+
+st.title('Search offers')
+
+bi_encoder = SentenceTransformer('multi-qa-MiniLM-L6-cos-v1')
+cross_encoder = CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2')
+
+tensors = {}
+with safe_open("embeddings.safetensors", framework="pt") as f : 
+	for k in f.keys():
+		tensors[k] = f.get_tensor(k)
+corpus_embeddings = tensors['embedding']
+
+with open('corpus.pickle', 'rb') as f:
+	passages = pickle.load(f)
+
+
+def search(query, top_k):
+
+	query_embedding = bi_encoder.encode(query, convert_to_tensor=True)
+	query_embedding = query_embedding #.cuda()
+	hits = util.semantic_search(query_embedding, corpus_embeddings, top_k=top_k)
+	hits = hits[0]
+
+	cross_inp = [[query, passages[hit['corpus_id']]] for hit in hits]
+	cross_scores = cross_encoder.predict(cross_inp)
+
+	# Sort results by the cross-encoder scores
+	for idx in range(len(cross_scores)):
+	    hits[idx]['cross-score'] = cross_scores[idx]
+
+	hits = sorted(hits, key=lambda x: x['cross-score'], reverse=True)
+	score_list, output_list = [],[]
+	for hit in hits[:10]:
+	  score_list.append("{:.3f}".format(hit['cross-score']))
+	  temp_output = passages[hit['corpus_id']].replace("\n", " ")
+	  temp_output = list(temp_output.rsplit('{'))[0].strip()
+	  output_list.append(temp_output)
+
+	dataframe = pd.DataFrame({'score': score_list, 'offers': output_list})
+	dataframe.drop_duplicates(subset=['offers'], keep='first', inplace=True)
+	return dataframe
+	
+	
+with st.form("my_form"):
+	query = st.text_input("Enter the brand name, category or  retailer name to search \
+							for relevant offers 👇",
+							placeholder = "Enter the text here")
+	num = st.number_input('Manximum number of offers to display', min_value=1, max_value=10)
+	
+	submitted = st.form_submit_button("Submit")
+	if submitted:
+		df = search(query, num)
+		st.dataframe(df, use_container_width=True)

corpus.pickle

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5347f3cfe8e90681a71e27f3a77bbb906e1c05eb9f5cbe9c1f28eeafecdd94ab
+size 72475

embeddings.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17f262c76dd1865aac7f4c486d7c951e1f0adfb7bc5e044b13268e67ead5a494
+size 1274968

requirements.txt

@@ -0,0 +1,4 @@
+sentence-transformers==2.2.2
+safetensors==0.3.3
+pandas==1.5.3
+pickle==4.0