Simple piece of code to query local document using LLAMA2 in replicate and OpenAI

 USE_REPLICATE = True
if USE_REPLICATE:
    print('Using replicate')
    import os
    os.environ["REPLICATE_API_TOKEN"] = "r8_Undampori9Pazhampory10SavalaVAda"
    
    from llama_index.core import Settings, VectorStoreIndex, SimpleDirectoryReader
    from llama_index.embeddings.huggingface import HuggingFaceEmbedding
    from llama_index.llms.replicate import Replicate
    from transformers import AutoTokenizer
    print("Finished importing everything")
    # set the LLM
    llama2_7b_chat = "meta/llama-2-7b-chat:8e6975e5ed6174911a6ff3d60540dfd4844201974602551e10e9e87ab143d81e"
    Settings.llm = Replicate(
        model=llama2_7b_chat,
        temperature=0.01,
        additional_kwargs={"top_p": 1, "max_new_tokens": 300},
    )
    print("Settings finished")
    # set tokenizer to match LLM
    Settings.tokenizer = AutoTokenizer.from_pretrained(
        "NousResearch/Llama-2-7b-chat-hf"
    )
    print("Initialized Settings tokenizer")
    # set the embed model
    Settings.embed_model = HuggingFaceEmbedding(
        model_name="BAAI/bge-small-en-v1.5"
    )
    print("Loaded Settings embedded model")
    documents = SimpleDirectoryReader("docs").load_data()
    index = VectorStoreIndex.from_documents(
        documents,
    )
    print("Loaded index")
    query_engine = index.as_query_engine()
    print("initialized query engine")
    response = query_engine.query(" What is auto-negotiation in a switch")
    print('response is ',response)
else:
    print('Running without replicate')
    import os
    os.environ['OPENAI_API_KEY'] = "sk-_Undampori9Pazhampory10SavalaVAda"
    from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
    documents = SimpleDirectoryReader("docs").load_data()
    index = VectorStoreIndex.from_documents(documents)
    query_engine = index.as_query_engine()
    response = query_engine.query(" What is auto-negotiation in a switch")
    print('response is ',response)
x