change local llm to huggingface inference API

requirements.txt

@@ -16,15 +16,14 @@ seaborn
 matplotlib
 
 # Hugging Face model integration
-huggingface-hub
+langchain-huggingface>=0.0.2
+huggingface-hub>=0.19.0
+transformers>=4.30.0
 
 # LangChain (for AI chat and LLM integration)
-langchain
-langchain-community
-
 langchain==1.0.2
 langchain-community
-langchain-huggingface
-transformers>=4.30.0
+
 accelerate
-safetensors
+safetensors
+

tabs/business.py

@@ -104,6 +104,6 @@ def render():
                 gr.HTML("<h3> Scenario Chat : What-If Analysis</h3>")
                 respond,title = ai_chat_factory("Business Insights & What-If Analysis")
                 # gr.ChatInterface(fn=respond, chatbot=gr.Chatbot(label=title,height=290))
-                gr.ChatInterface(fn=respond, chatbot=gr.Chatbot(type="messages",height=250))
+                gr.ChatInterface(fn=respond, type="messages", chatbot=gr.Chatbot(type="messages",height=250))
 
     # gr.HTML(f"<style>{css}</style>")

tabs/shared_ai.py

@@ -17,7 +17,7 @@ from pydantic import BaseModel, ValidationError
 
 # LangChain components
 from langchain_community.chat_models import ChatOllama, ChatHuggingFace
-from langchain_community.llms import HuggingFaceHub
+# from langchain_community.llms import HuggingFaceHub
 
 from langchain_community.chat_models import ChatOllama
 from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint
@@ -61,7 +61,7 @@ def get_llm():
             return None
         try:
             llm = ChatHuggingFace(
-                llm=HuggingFaceHub(
+                llm=HuggingFaceEndpoint(
                     repo_id="mistralai/Mistral-7B-Instruct-v0.2",
                     huggingfacehub_api_token=HF_TOKEN,
                     model_kwargs={"temperature": 0.3, "max_new_tokens": 512}