import os 
#!pip install -q gradio langchain pypdf chromadb
import gradio as gr
from dotenv import load_dotenv
from PyPDF2 import PdfReader
from langchain.vectorstores import Chroma
from langchain.vectorstores import FAISS
from langchain.document_loaders import PyPDFLoader
from langchain.text_splitter import CharacterTextSplitter
from langchain.embeddings import HuggingFaceInferenceAPIEmbeddings
from langchain.embeddings import HuggingFaceBgeEmbeddings
from langchain.memory import ConversationBufferMemory
from langchain.chains import ConversationalRetrievalChain
from htmlTemplates import css, bot_template, user_template
from langchain.llms import HuggingFaceHub


# Use Hugging Face Inference API embeddings
inference_api_key = os.environ['HF']
api_hf_embeddings = HuggingFaceInferenceAPIEmbeddings(
    api_key=inference_api_key,
    model_name="sentence-transformers/all-MiniLM-l6-v2"
)

# Load and process the PDF files
loader = PyPDFLoader("./new_papers/ALiBi.pdf")
documents = loader.load()
print("-----------")
print(documents[0])
print("-----------")

# Split the documents into chunks and embed them using the HfApiEmbeddingTool
text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=0)
vdocuments = text_splitter.split_documents(documents)




model = "BAAI/bge-base-en-v1.5"
encode_kwargs = {
    "normalize_embeddings": True
}  # set True to compute cosine similarity
embeddings = HuggingFaceBgeEmbeddings(
    model_name=model, encode_kwargs=encode_kwargs, model_kwargs={"device": "cpu"}
)
api_db = FAISS.from_texts(texts=vdocuments, embedding=embeddings)
api_db.as_retriever.similarity("What is ICD?")


# Extract the embedding arrays from the PDF documents
#embeddings = []
#for doc in vdocuments:
#    embeddings.extend(api_hf_embeddings.get_embeddings(doc))

# Create Chroma vector store for API embeddings
#api_db = Chroma.from_documents(vdocuments, HfApiEmbeddingRetriever, collection_name="api-collection")

# Define the PDF retrieval function
def pdf_retrieval(query):
    # Run the query through the retriever
    response = api_db.similarity_search(query)
    return response

# Create Gradio interface for the API retriever
# Create Gradio interface for the API retriever
api_tool = gr.Interface(
    fn=pdf_retrieval,
    inputs=[gr.Textbox()],
    outputs=gr.Textbox(),
    live=True,
    title="API PDF Retrieval Tool",
    description="This tool indexes PDF documents and retrieves relevant answers based on a given query (HF Inference API Embeddings).",
)

# Launch the Gradio interface
api_tool.launch()