Update app_utils.py

app_utils.py

@@ -4,6 +4,10 @@ Financial Chatbot Utilities
 Core functionality for RAG-based financial chatbot
 """
 
+# mount
+import sys
+sys.path.append('/home/user/app')
+
 import os
 import re
 import nltk
@@ -26,20 +30,17 @@ from sentence_transformers import CrossEncoder
 from sklearn.metrics.pairwise import cosine_similarity
 
 # Initialize NLTK stopwords
-nltk.download('stopwords')
-stop_words = set(stopwords.words('english'))
-# nltk.data.path.append('./nltk_data')  # Point to local NLTK data
-# stop_words = set(nltk.corpus.stopwords.words('english'))
+# nltk.download('stopwords')
+# stop_words = set(stopwords.words('english'))
+nltk.data.path.append('./nltk_data')  # Point to local NLTK data
+stop_words = set(nltk.corpus.stopwords.words('english'))
 
-# mount
-import sys
-sys.path.append('/mount/src/gen_ai_dev')
 
 # Configuration
 DATA_PATH = "./Infy financial report/"
 DATA_FILES = ["INFY_2022_2023.pdf", "INFY_2023_2024.pdf"]
 EMBEDDING_MODEL = "sentence-transformers/all-MiniLM-L6-v2"
-LLM_MODEL = "gpt2"  # Or "distilgpt2"  # Or "HuggingFaceH4/zephyr-7b-beta" or "microsoft/phi-2"
+LLM_MODEL = "microsoft/phi-1.5" # or "gpt2"  # Or "distilgpt2"  # Or "HuggingFaceH4/zephyr-7b-beta" or "microsoft/phi-2"
 
 # Environment settings
 os.environ["TOKENIZERS_PARALLELISM"] = "false"