from langchain_community.document_loaders import PyPDFLoader
import gradio as gr
from langchain.chains.summarize import load_summarize_chain
from huggingface_hub import InferenceClient
from langchain_huggingface import HuggingFaceEndpoint
import os
# Set your Hugging Face token securely
HUGGINGFACEHUB_API_TOKEN = os.environ["HUGGINGFACEHUB_API_TOKEN"] 

# Create the LLM
#llm = HuggingFaceHub(
#    repo_id="facebook/bart-large-cnn",  # Summarization-capable model
#    model_kwargs={"temperature": 0.7, "max_length": 512}
#)
repo_id = "google/gemma-3-12b-it"

llm = HuggingFaceEndpoint(
    repo_id=repo_id,
    temperature=0.5,
    huggingfacehub_api_token=HUGGINGFACEHUB_API_TOKEN,
    provider="featherless-ai",  # set your provider here hf.co/settings/inference-providers
    # provider="hyperbolic",
    # provider="nebius",
    # provider="together",
)


#TEXT_MODEL_NAME = "google/gemma-3-270m"
loader = PyPDFLoader("http://arxiv.org/pdf/2508.13246v1") 
documents = loader.load()
#llm = OpenAI(temperature=0)

def summarize_pdf (pdf_file_path, custom_prompt=""):
    loader = PyPDFLoader(pdf_file_path)
    docs = loader.load_and_split()
    chain = load_summarize_chain(llm, chain_type="map_reduce")
    summary = chain.invoke(docs)
    
    return summary


input_pdf_path = gr.Textbox(label="Enter the PDF file path")
output_summary = gr.Textbox(label="Summary")
    
interface = gr.Interface(
    fn = summarize_pdf,
    inputs = input_pdf_path,
    outputs = output_summary,
    title = "PDF Summarizer",
    description = "This app allows you to summarize your PDF files.",
)
#demo.launch(share=True)

if __name__ == "__main__":
    interface.launch(share=True)