|
from langchain_core.prompts import ChatPromptTemplate
|
|
from langchain_core.output_parsers import StrOutputParser
|
|
from langchain.llms import Ollama
|
|
|
|
|
|
import streamlit as st
|
|
import os
|
|
from dotenv import load_dotenv
|
|
|
|
load_dotenv()
|
|
|
|
os.environ["LANGCHAIN_TRACING_V2"]="true"
|
|
os.environ["LANGCHAIN_API_KEY"]=os.getenv("LANGCHAIN_API_KEY")
|
|
|
|
|
|
|
|
prompt=ChatPromptTemplate.from_messages(
|
|
[
|
|
("system","You are a helpful assistant. Please response to the user queries"),
|
|
("user","Question:{question}")
|
|
]
|
|
)
|
|
|
|
|
|
|
|
st.title('Langchain Demo With LLAMA2 API')
|
|
input_text=st.text_input("Search the topic u want")
|
|
|
|
|
|
llm=Ollama(model="llama2")
|
|
output_parser=StrOutputParser()
|
|
chain=prompt|llm|output_parser
|
|
|
|
if input_text:
|
|
st.write(chain.invoke({'question':input_text})) |