app.py

import gradio as gr
from paperqa import Docs, SentenceTransformerEmbeddingModel
from langchain_anthropic import ChatAnthropic

MODEL_NAME = "claude-3-5-sonnet-20240620"
class MyChatAnthropic(ChatAnthropic):
    model_name = MODEL_NAME
llm = MyChatAnthropic(
    model=MODEL_NAME,
    temperature=0.2,
    max_tokens=4096,)

class MyEmb(SentenceTransformerEmbeddingModel):
    async def aembed_documents(self, texts):
        return await self.embed_documents(None, texts)

emb = MyEmb(model_name="mixedbread-ai/mxbai-embed-large-v1")
docs = Docs(llm="langchain",
            embedding="langchain",
            embedding_client=emb,
            client=llm)
docs.max_concurrent = 1
docs.add('knowledge_extraction.csv', disable_check=True)

def respond(message):
    return docs.query(messages).answer

"""
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
"""
demo = gr.ChatInterface(respond)


if __name__ == "__main__":
    demo.launch()