app.py

import streamlit as st
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
from langchain.memory import ConversationBufferMemory
from langchain.memory.chat_message_histories import StreamlitChatMessageHistory
from langchain.chains import LLMChain
from langchain.prompts import PromptTemplate
from langchain.embeddings import VoyageEmbeddings
from langchain.vectorstores import SupabaseVectorStore
from langchain.llms.huggingface_pipeline import HuggingFacePipeline
from st_supabase_connection import SupabaseConnection

msgs = StreamlitChatMessageHistory()
memory = ConversationBufferMemory(memory_key="history", chat_memory=msgs)

supabase_client = st.connection(
    name="orbgpt",
    type=SupabaseConnection,
    ttl=None,
)

embeddings = VoyageEmbeddings(model="voyage-01")
vector_store = SupabaseVectorStore(
    embedding=embeddings,
    client=supabase_client,
    table_name="documents",
    query_name="match_documents",
)


model_path = "01-ai/Yi-6B-Chat-8bits"
tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)
model = AutoModelForCausalLM.from_pretrained(
    model_path, device_map="auto", torch_dtype="auto"
).eval()
pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    max_new_tokens=10,
    use_fast=False,
)
hf = HuggingFacePipeline(pipeline=pipe)

template = """Question: {question}

Answer: Let's think step by step."""
prompt = PromptTemplate.from_template(template)

chain = prompt | hf

question = "What is electroencephalography?"

st.text(chain.invoke({"question": question}))

st.title("🪩🤖")

if len(msgs.messages) == 0:
    msgs.add_ai_message("Ask me anything about orb community projects!")

for msg in msgs.messages:
    st.chat_message(msg.type).write(msg.content)

if prompt := st.chat_input("Ask something"):
    st.chat_message("human").write(prompt)
    # Run
    st.chat_message("ai").write("hehe")