app.py

import os
import streamlit as st
import arxiv
import networkx as nx
import matplotlib.pyplot as plt
import datetime

# -------------------------------
# Groq API Client
# -------------------------------
from groq import Groq

client = Groq(
    api_key=os.environ.get("GROQ_API_KEY"),
)

# -------------------------------
# Helper Functions (Groq-based)
# -------------------------------
def groq_summarize(text: str) -> str:
    """
    Summarize the given text using Groq's chat completion API.
    """
    response = client.chat.completions.create(
        messages=[
            {"role": "user", "content": f"Summarize the following text in detail:\n\n{text}"}
        ],
        model="llama-3.3-70b-versatile",
    )
    return response.choices[0].message.content.strip()

def groq_simplify(text: str) -> str:
    """
    Explain Like I'm 5 (ELI5) version of the summary.
    """
    response = client.chat.completions.create(
        messages=[
            {"role": "user", "content": f"Explain the following like I'm 5 years old:\n\n{text}"}
        ],
        model="llama-3.3-70b-versatile",
    )
    return response.choices[0].message.content.strip()

def groq_generate_key_takeaways(text: str) -> str:
    """
    Generate key takeaways from the paper.
    """
    response = client.chat.completions.create(
        messages=[
            {"role": "user", "content": f"Provide key takeaways from this research paper:\n\n{text}"}
        ],
        model="llama-3.3-70b-versatile",
    )
    return response.choices[0].message.content.strip()

def retrieve_papers(query, max_results=5):
    """Retrieve academic papers from arXiv, including DOI and tools for relevance/trust scoring."""
    search = arxiv.Search(query=query, max_results=max_results)
    papers = []
    for result in search.results():
        paper = {
            "title": result.title,
            "summary": result.summary,
            "url": result.pdf_url,
            "authors": [author.name for author in result.authors],
            "published": result.published,
            "doi": result.doi if hasattr(result, "doi") else f"https://doi.org/10.48550/arXiv.{result.entry_id.split('/')[-1]}",
            "litmaps": f"https://app.litmaps.com/preview/{result.entry_id.split('/')[-1]}",
            "connected_papers": f"https://www.connectedpapers.com/main/{result.entry_id.split('/')[-1]}",
            "scite_ai": f"https://scite.ai/reports/{result.entry_id.split('/')[-1]}",
            "biblio_explorer": f"https://arxiv.org/bib_explorer/{result.entry_id.split('/')[-1]}",
        }
        papers.append(paper)
    return papers

def get_cached_summary(paper_id, text):
    """Retrieve or generate summaries, ELI5 explanations, and key takeaways."""
    if 'summaries' not in st.session_state:
        st.session_state.summaries = {}
    if paper_id not in st.session_state.summaries:
        st.session_state.summaries[paper_id] = {
            "summary": groq_summarize(text),
            "eli5": groq_simplify(text),
            "key_takeaways": groq_generate_key_takeaways(text),
        }
    return st.session_state.summaries[paper_id]

st.title("📚 PaperPilot – Intelligent Academic Navigator")

with st.sidebar:
    st.header("🔍 Search Parameters")
    query = st.text_input("Research topic or question:")
    if st.button("🚀 Find Articles"):
        if query.strip():
            with st.spinner("Searching arXiv..."):
                papers = retrieve_papers(query)
                if papers:
                    st.session_state.papers = papers
                    st.success(f"Found {len(papers)} papers!")
                    st.session_state.active_section = "review"
                else:
                    st.error("No papers found. Try different keywords.")
        else:
            st.warning("Please enter a search query")

if 'active_section' not in st.session_state:
    st.session_state.active_section = "none"

if 'papers' in st.session_state and st.session_state.papers:
    papers = st.session_state.papers

    if st.session_state.active_section == "review":
        st.header("📚 Literature Review & Summary")
        for idx, paper in enumerate(papers, 1):
            with st.expander(f"Summary: {paper['title']}"):
                with st.spinner(f"Analyzing {paper['title']}..."):
                    paper_id = f"paper_{idx}"
                    summary_data = get_cached_summary(paper_id, paper['summary'])
                    st.markdown(f"**Summary:** {summary_data['summary']}")
                    st.markdown(f"**ELI5:** {summary_data['eli5']}")
                    st.markdown("**Key Takeaways:**")
                    st.write(summary_data['key_takeaways'])
                    st.markdown(f"**DOI:** [Link]({paper['doi']})")
                    st.markdown(f"**Bibliographic Explorer:** [View]({paper['biblio_explorer']})")
                    st.markdown(f"**Connected Papers:** [View]({paper['connected_papers']})")
                    st.markdown(f"**Litmaps:** [View]({paper['litmaps']})")
                    st.markdown(f"**Scite.ai Citations:** [View]({paper['scite_ai']})")
    
else:
    st.info("Enter a query in the sidebar and click 'Find Articles' to get started.")

st.caption("Built with ❤️ using AI")