Update app.py

app.py

@@ -4,7 +4,7 @@ import nltk
 from nltk import sent_tokenize
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 import json
-
+import numpy as np
 from sentence_transformers import SentenceTransformer
 
 nltk.download('punkt')
@@ -38,6 +38,14 @@ def infer(input_ids, max_length, temperature, top_k, top_p):
     return output_sequences
 
 
+def cos_similarity(v1, v2):
+    dot_product = np.dot(v1, v2)
+    l2_norm = (np.sqrt(sum(np.square(v1))) * np.sqrt(sum(np.square(v2))))
+    similarity = dot_product / l2_norm     
+    
+    return similarity
+
+
 @st.cache_data
 def chunking(book_text):
     sentences = sent_tokenize(book_text)