Update app.py

app.py

@@ -9,13 +9,11 @@ from langchain.chains import ConversationalRetrievalChain
 from ctransformers import AutoModelForCausalLM
 from langchain_g4f import G4FLLM
 from g4f import Provider, models
-import unicodedata
+# import spacy
 import requests
 # Define the path for generated embeddings
 DB_FAISS_PATH = 'vectorstore/db_faiss'
-def is_japanese_character(character):
-    return 'CJK UNIFIED' in unicodedata.name(character, '')
-    
+
 # Load the model of choice
 def load_llm():
     # url = "https://huggingface.co/TheBloke/Llama-2-7B-Chat-GGML/blob/main/llama-2-7b-chat.ggmlv3.q4_K_M.bin"  # 2.87G
@@ -54,7 +52,7 @@ st.markdown(hide_streamlit_style, unsafe_allow_html=True)
 # Set the title for the Streamlit app
 st.title("Zendo美女チャットボックス")
 
-csv_url = "https://huggingface.co/spaces/uyen13/chatgirl2/raw/main/testchatdata.csv"
+csv_url = "https://huggingface.co/spaces/uyen13/chatzendo/raw/main/testchatdata.csv"
 # csv_url="https://docs.google.com/uc?export=download&id=1fQ2v2n9zQcoi6JoOU3lCBDHRt3a1PmaE"
 
 # Define the path where you want to save the downloaded file
@@ -92,14 +90,6 @@ chain = ConversationalRetrievalChain.from_llm(llm=llm, retriever=db.as_retriever
 def conversational_chat(query):
     query = "提供されたデータに基づいて,"+query
     result = chain({"question": query, "chat_history": st.session_state['history']})
-    # i = 0
-    # while i < len(result["answer"]):
-    #     character = input_string[i]
-    #     if is_japanese_character(character):
-    #         break
-    #     else:
-    #         result = chain({"question": query, "chat_history": st.session_state['history']})
-    #     i += 1
     st.session_state['history'].append((query, result["answer"]))
     return result["answer"]