from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.document_loaders import GutenbergLoader
from langchain.vectorstores import Chroma
from langchain.text_splitter import CharacterTextSplitter
from langchain.chains.question_answering import load_qa_chain
from langchain.llms import OpenAI


docs = [
    doc
    for path in ['cache/epub/35688/pg35688.txt', 'files/23718/23718-readme.txt']
    for doc in GutenbergLoader(f'https://www.gutenberg.org/{path}').load()
]
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0, separator = "\n")
texts = text_splitter.split_documents(docs)

chain = None
embeddings = None
docsearch = None


def make_inference(openai_api_key, query):
    global chain, embeddings, docsearch

    if chain is None:
        chain = load_qa_chain(OpenAI(temperature=0, openai_api_key=openai_api_key.strip()), chain_type="stuff")
        embeddings = OpenAIEmbeddings(openai_api_key=openai_api_key)
        docsearch = Chroma.from_documents(texts, embeddings, metadatas=[{"source": str(i)} for i in range(len(texts))]).as_retriever()

    docs = docsearch.get_relevant_documents(query)
    return(chain.run(input_documents=docs, question=query))

if __name__ == "__main__":
    # make a gradio interface
    import gradio as gr

    gr.Interface(
        make_inference,
        [
            gr.inputs.Textbox(lines=1, label="OpenAI API Key"),
            gr.inputs.Textbox(lines=2, label="Query"),
        ],
        gr.outputs.Textbox(label="Response"),
        title="Multiple Document QA with LangChain",
        description="Ask me anything about Lewis Carroll's Alice in Wonderland and Through the Looking-Glass",
    ).launch()