app.py

import asyncio
import re
from pathlib import Path

import streamlit as st
import pandas as pd
import plotly.express as px
from fpdf import FPDF
from streamlit_agraph import agraph

from mcp.orchestrator import orchestrate_search, answer_ai_question
from mcp.knowledge_graph import build_agraph
from mcp.graph_metrics import build_nx, get_top_hubs, get_density
from mcp.protocols import draft_protocol

# Streamlit configuration
st.set_page_config(page_title="MedGenesis AI", layout="wide")

# Initialize session state
if "res" not in st.session_state:
    st.session_state.res = None

# Header UI
st.title("🧬 MedGenesis AI")
llm = st.radio("LLM engine", ["openai", "gemini"], horizontal=True)
query = st.text_input("Enter biomedical question")

# PDF generation helper
def _make_pdf(papers):
    pdf = FPDF()
    pdf.add_page()
    pdf.set_font("Helvetica", size=12)
    pdf.cell(0, 10, "MedGenesis AI – Results", ln=True, align="C")
    pdf.ln(5)
    for i, p in enumerate(papers, 1):
        pdf.set_font("Helvetica", "B", 11)
        pdf.multi_cell(0, 7, f"{i}. {p.get('title','')}")
        pdf.set_font("Helvetica", size=9)
        body = f"""{p.get('authors','')}
{p.get('summary','')}
{p.get('link','')}"""
        pdf.multi_cell(0, 6, body)
        pdf.ln(3)
    return pdf.output(dest="S").encode("latin-1", errors="replace")

# Trigger search
if st.button("Run Search 🚀") and query.strip():
    with st.spinner("Gathering data…"):
        st.session_state.res = asyncio.run(orchestrate_search(query, llm))

# Retrieve results
res = st.session_state.res

# If no results yet, prompt user
if not res:
    st.info("Enter a question and press **Run Search 🚀** to begin.")
else:
    # Create tabs
    tabs = st.tabs(["Results", "Graph", "Clusters", "Variants", "Trials", "Metrics", "Visuals", "Protocols"])
    title_tab, graph_tab, clust_tab, var_tab, trial_tab, met_tab, vis_tab, proto_tab = tabs

    # Results Tab
    with title_tab:
        for i, p in enumerate(res["papers"], 1):
            st.markdown(f"**{i}. [{p['title']}]({p['link']})**")
            st.write(p["summary"])
        c1, c2 = st.columns(2)
        c1.download_button("CSV", pd.DataFrame(res["papers"]).to_csv(index=False),
                           "papers.csv", "text/csv")
        c2.download_button("PDF", _make_pdf(res["papers"]),
                           "papers.pdf", "application/pdf")
        st.subheader("AI summary")
        st.info(res["ai_summary"])

    # Graph Tab
    with graph_tab:
        nodes, edges, cfg = build_agraph(res["papers"], res["umls"], res.get("drug_safety", []), res.get("umls_relations", []))
        hl = st.text_input("Highlight node:", key="hl")
        if hl:
            pat = re.compile(re.escape(hl), re.I)
            for n in nodes:
                n.color = "#f1c40f" if pat.search(n.label) else n.color
        agraph(nodes, edges, cfg)

    # Clusters Tab
    with clust_tab:
        clusters = res.get("clusters", [])
        if clusters:
            df = pd.DataFrame({
                "title": [p['title'] for p in res['papers']],
                "cluster": clusters
            })
            st.write("### Paper Clusters")
            for c in sorted(set(clusters)):
                st.write(f"**Cluster {c}**")
                for t in df[df['cluster'] == c]['title']:
                    st.write(f"- {t}")
        else:
            st.info("No clusters to show.")

    # Variants Tab
    with var_tab:
        variants = res.get("variants", [])
        if variants:
            st.json(variants)
        else:
            st.warning("No variants found. Try a well-known gene like 'TP53'.")

    # Trials Tab
    with trial_tab:
        trials = res.get("clinical_trials", [])
        if trials:
            st.json(trials)
        else:
            st.warning("No trials found. Try a disease name or specific drug.")

    # Metrics Tab
    with met_tab:
        G = build_nx([n.__dict__ for n in nodes], [e.__dict__ for e in edges])
        st.metric("Density", f"{get_density(G):.3f}")
        st.markdown("**Top hubs**")
        for nid, sc in get_top_hubs(G):
            label = next((n.label for n in nodes if n.id == nid), nid)
            st.write(f"- {label}: {sc:.3f}")

    # Visuals Tab
    with vis_tab:
        years = [p.get("published", "")[:4] for p in res["papers"] if p.get("published")]
        if years:
            st.plotly_chart(px.histogram(pd.DataFrame({'year': years}), x='year', nbins=10, title="Publication Year"))

    # Protocols Tab
    with proto_tab:
        hyp = st.text_input("Enter hypothesis for protocol:", key="proto_q")
        if st.button("Draft Protocol") and hyp.strip():
            with st.spinner("Generating protocol…"):
                doc = asyncio.run(draft_protocol(hyp, context=res["ai_summary"], llm=llm))
            st.subheader("Experimental Protocol")
            st.write(doc)