Update app.py

app.py

@@ -3,17 +3,15 @@ from minivectordb.embedding_model import EmbeddingModel
 from minivectordb.vector_database import VectorDatabase
 from multiprocessing import cpu_count
 from functools import lru_cache
-import fasttext, random, nltk, tiktoken, os
+import fasttext, random, tiktoken, os, pickle
 import concurrent.futures
-nltk.download('stopwords')
-from nltk.corpus import stopwords
 
 os.environ['TOKENIZERS_PARALLELISM'] = 'true'
 
 langdetect_model = fasttext.load_model('lid.176.ftz')
 embedding_model = EmbeddingModel(onnx_model_cpu_core_count=1)
-en_stop_words = stopwords.words('english')
-pt_stop_words = stopwords.words('portuguese')
+en_stop_words = pickle.load(open("en_stopwords.pkl", "rb"))
+pt_stop_words = pickle.load(open("pt_stopwords.pkl", "rb"))
 tokenizer = tiktoken.encoding_for_model("gpt-4")
 
 def count_tokens_tiktoken(text):