Update app.py

app.py

@@ -7,6 +7,9 @@ import requests
 import random
 import urllib.parse
 import spacy
+import nltk
+from nltk.tokenize import sent_tokenize
+from typing import List, Dict
 from tempfile import NamedTemporaryFile
 from typing import List, Dict
 from bs4 import BeautifulSoup
@@ -22,26 +25,28 @@ from langchain_core.documents import Document
 
 huggingface_token = os.environ.get("HUGGINGFACE_TOKEN")
 
+# Download necessary NLTK data
+nltk.download('punkt')
+
 class Agent1:
     def __init__(self):
-        self.nlp = spacy.load("en_core_web_sm")
+        pass
 
     def rephrase_and_split(self, user_input: str) -> List[str]:
-        doc = self.nlp(user_input)
-        
         # Identify question words
         question_words = set(["what", "when", "where", "who", "whom", "which", "whose", "why", "how"])
         
         # Split sentences
-        sentences = list(doc.sents)
+        sentences = sent_tokenize(user_input)
         
         # Identify questions
         questions = []
         for sent in sentences:
-            if sent[0].text.lower() in question_words or sent[-1].text == "?":
-                questions.append(sent.text)
-            elif any(token.text.lower() in question_words for token in sent):
-                questions.append(sent.text)
+            words = sent.lower().split()
+            if words[0] in question_words or sent.strip().endswith('?'):
+                questions.append(sent)
+            elif any(word in question_words for word in words):
+                questions.append(sent)
         
         # If no questions identified, return the original input
         if not questions: