app.py

"""
app.py — Gradio interface for the Multi-Agent RAG Research System.

Agent 1 (Research Summarizer) — runs automatically when a paper is submitted.
Agent 2 (Code Explainer)      — runs on-demand via a button click.

Set OPENAI_API_KEY in Space → Settings → Variables and secrets.
Optionally set GITHUB_TOKEN for higher GitHub API rate limits.
"""

from __future__ import annotations

import os
import tempfile

import gradio as gr
from dotenv import load_dotenv
from langchain_core.messages import AIMessage, HumanMessage

load_dotenv()

_MISSING_KEY_MSG = (
    "**GROQ_API_KEY is not set.**\n\n"
    "Go to **Settings → Variables and secrets** and add your key."
)

def _key_is_set() -> bool:
    return bool(os.getenv("GROQ_API_KEY"))

_SUMMARY_COMPLETE = False
_SUMMARY_RUNNING = False

def reset_summary_state():
    global _SUMMARY_COMPLETE, _SUMMARY_RUNNING
    _SUMMARY_COMPLETE = False
    _SUMMARY_RUNNING = False

def _invoke_agent(agent, query: str) -> str:
    result = agent.invoke({"messages": [HumanMessage(content=query)]})
    messages = result.get("messages", [])
    last = next(
        (m for m in reversed(messages) if isinstance(m, AIMessage)),
        AIMessage(content="(no response)"),
    )
    return last.content


# ---------------------------------------------------------------------------
# Agent 1 — Research Summarizer
# Returns: (summary_out update, explain_btn update)
# ---------------------------------------------------------------------------

def run_summarizer(paper_url: str, pdf_file):
    global _SUMMARY_COMPLETE, _SUMMARY_RUNNING

    if not _key_is_set():
        return gr.update(value=_MISSING_KEY_MSG, visible=True), gr.update(visible=False)

    has_url = bool(paper_url and paper_url.strip())
    has_pdf = pdf_file is not None

    if not has_url and not has_pdf:
        return (
            gr.update(value="Please provide a paper URL / arXiv ID or upload a PDF.", visible=True),
            gr.update(visible=False),
        )

    if _SUMMARY_RUNNING:
        return (
            gr.update(value="⏳ Summary is already running.", visible=True),
            gr.update(visible=False),
        )

    if _SUMMARY_COMPLETE:
        return (
            gr.update(visible=True),
            gr.update(visible=True),
        )

    _SUMMARY_RUNNING = True

    from vector_store import reset_store
    reset_store()

    try:
        from agents import build_paper_summariser

        if has_pdf:
            from tools import fetch_pdf_paper
            pdf_path = pdf_file if isinstance(pdf_file, str) else pdf_file.name
            fetch_pdf_paper.invoke({"pdf_path_or_url": pdf_path})
            query = (
                "The research paper has already been indexed from an uploaded PDF. "
                "Summarise it now — do NOT call fetch_arxiv_paper or fetch_pdf_paper again."
            )
        else:
            query = f"Summarise this research paper: {paper_url.strip()}"

        agent = build_paper_summariser()
        summary = _invoke_agent(agent, query)
        _SUMMARY_COMPLETE = True
        _SUMMARY_RUNNING = False
        return gr.update(value=summary, visible=True), gr.update(visible=True)

    except Exception as exc:
        import traceback
        traceback.print_exc()
        return gr.update(value=f"Error: {exc}", visible=True), gr.update(visible=False)

    finally:
        _SUMMARY_RUNNING = False


# ---------------------------------------------------------------------------
# Agent 2 — Code Explainer
# Returns: (code_out update, download_btn update)
# ---------------------------------------------------------------------------

def run_code_explainer():
    if not _key_is_set():
        return gr.update(value=_MISSING_KEY_MSG, visible=True), gr.update(visible=False)

    query = (
        "Find all GitHub repository links mentioned in the indexed paper. "
        "Fetch and index each repository. "
        "Extract and explain every algorithm and pseudocode block from the paper. "
        "Then produce a complete how-to-run guide with exact terminal commands "
        "starting from git clone."
    )

    try:
        from agents import build_code_explainer
        print("[DEBUG] Starting code explainer...")
        agent = build_code_explainer()
        explanation = _invoke_agent(agent, query)
        print("[DEBUG] Code explainer finished.")

        with tempfile.NamedTemporaryFile(
            mode="w", suffix=".md", delete=False, encoding="utf-8"
        ) as f:
            f.write(explanation)
            tmp_path = f.name

        return gr.update(value=explanation, visible=True), gr.update(value=tmp_path, visible=True)

    except Exception as exc:
        import traceback
        traceback.print_exc()
        return gr.update(value=f"Error: {exc}", visible=True), gr.update(visible=False)


# ---------------------------------------------------------------------------
# Gradio UI
# ---------------------------------------------------------------------------

def show_summarizer_working():
    if _SUMMARY_COMPLETE:
        return gr.update(value="", visible=False)
    if _SUMMARY_RUNNING:
        return gr.update(value="⏳ Summary is already running.", visible=True)
    return gr.update(value="⏳ Summarizing paper... this may take a minute.", visible=True)

def show_code_working():
    return gr.update(
        value=(
            "⏳ Searching for repositories, indexing code evidence, and generating "
            "the explanation. This can take a few minutes."
        ),
        visible=True,
    )

TITLE_HTML = """
<div style="text-align:center; padding:28px 0 4px 0;">
  <h1 style="font-size:2rem; font-weight:700; margin:0; letter-spacing:-0.5px;">
    Multiagent Research Paper Summarizer &amp; Code Extractor
  </h1>

  <hr style="margin:20px auto; width:60%; border:none; border-top:1px solid #e5e7eb;">
</div>
"""

with gr.Blocks(
    title="Multiagent Research Paper Summarizer & Code Extractor",
    theme=gr.themes.Soft(),
    css="footer { display:none !important; }",
) as demo:
    gr.HTML(TITLE_HTML)

    with gr.Row():
        with gr.Column(scale=3):
            paper_url = gr.Textbox(
                label="Paper URL or arXiv ID",
                placeholder="https://arxiv.org/abs/1706.03762  or  1706.03762",
                lines=1,
            )
        with gr.Column(scale=2):
            pdf_upload = gr.File(
                label="Or upload a PDF",
                file_types=[".pdf"],
                type="filepath",
            )

    summarize_btn = gr.Button("Summarize Paper", variant="primary", size="lg")

    summary_out = gr.Markdown(visible=False)
    summary_status = gr.Markdown(visible=False)

    explain_btn = gr.Button(
        "Explain Code & Algorithms from this Paper",
        variant="secondary",
        size="lg",
        visible=False,
    )

    code_out = gr.Markdown(visible=False)
    code_status = gr.Markdown(visible=False)

    download_btn = gr.File(
        label="Download explanation as Markdown",
        visible=False,
        interactive=False,
    )

    # Agent 1: 2 outputs → summary_out, explain_btn
    summarize_btn.click(
        fn=show_summarizer_working,
        inputs=[],
        outputs=[summary_status],
    ).then(
        fn=run_summarizer,
        inputs=[paper_url, pdf_upload],
        outputs=[summary_out, explain_btn],
    ).then(
        fn=lambda: gr.update(visible=False),
        inputs=[],
        outputs=[summary_status],
    )

    # Enter key or click summarize_btn to trigger summarizer
    paper_url.submit(
        fn=show_summarizer_working,
        inputs=[],
        outputs=[summary_status],
    ).then(
        fn=run_summarizer,
        inputs=[paper_url, pdf_upload],
        outputs=[summary_out, explain_btn],
    ).then(
        fn=lambda: gr.update(visible=False),
        inputs=[],
        outputs=[summary_status],
    )

    # Agent 2: 2 outputs → code_out, download_btn
    explain_btn.click(
        fn=show_code_working,
        inputs=[],
        outputs=[code_status],
    ).then(
        fn=run_code_explainer,
        inputs=[],
        outputs=[code_out, download_btn],
    ).then(
        fn=lambda: gr.update(visible=False),
        inputs=[],
        outputs=[code_status],
    )

    # Reset summary state when paper URL changes
    paper_url.change(
        fn=reset_summary_state,
        inputs=[],
        outputs=[],
    )

    # Reset summary state when PDF file changes
    pdf_upload.change(
        fn=reset_summary_state,
        inputs=[],
        outputs=[],
    )

if __name__ == "__main__":
    demo.launch(ssr_mode=False)