#!/usr/bin/env python3
"""Lightweight local web demo (standard library only): paste an LLM-rewritten passage, get the HTR reconstruction,
a word-level diff and preservation warnings. Run: .venv/bin/python demo/app.py --model <release_dir> --port 8765"""
from __future__ import annotations

import argparse
import difflib
import html
import json
import sys
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
from pathlib import Path

ROOT = Path(__file__).resolve().parents[1]
sys.path.insert(0, str(ROOT / "src"))
from htr.inference.pipeline import NOTICE, HTRReconstructor  # noqa: E402

PAGE = (Path(__file__).parent / "index.html").read_text()


def word_diff(a: str, b: str) -> str:
    sm = difflib.SequenceMatcher(None, a.split(), b.split())
    out = []
    for op, i1, i2, j1, j2 in sm.get_opcodes():
        if op == "equal":
            out.append(html.escape(" ".join(a.split()[i1:i2])))
        else:
            if i2 > i1:
                out.append('<del>' + html.escape(" ".join(a.split()[i1:i2])) + "</del>")
            if j2 > j1:
                out.append('<ins>' + html.escape(" ".join(b.split()[j1:j2])) + "</ins>")
    return " ".join(out)


class H(BaseHTTPRequestHandler):
    rec: HTRReconstructor = None  # type: ignore[assignment]

    def _send(self, code: int, body: bytes, ctype: str) -> None:
        self.send_response(code)
        self.send_header("Content-Type", ctype)
        self.send_header("Content-Length", str(len(body)))
        self.end_headers()
        self.wfile.write(body)

    def do_GET(self):  # noqa: N802
        if self.path in ("/", "/index.html"):
            self._send(200, PAGE.replace("{{MODEL}}", html.escape(self.rec.meta.get("name", "htr"))).encode(), "text/html; charset=utf-8")
        elif self.path == "/health":
            self._send(200, json.dumps({"ok": True, "model": self.rec.meta.get("name")}).encode(), "application/json")
        else:
            self._send(404, b"not found", "text/plain")

    def do_POST(self):  # noqa: N802
        if self.path != "/api/reconstruct":
            return self._send(404, b"not found", "text/plain")
        n = int(self.headers.get("Content-Length", "0"))
        req = json.loads(self.rfile.read(n) or b"{}")
        text = (req.get("text") or "").strip()
        if not text:
            return self._send(400, json.dumps({"error": "empty text"}).encode(), "application/json")
        meta = {k: req.get(k) for k in ("discipline", "pub_year", "section_role") if req.get(k)}
        if "pub_year" in meta:
            meta["pub_year"] = int(meta["pub_year"])
        res = self.rec.reconstruct(text[:6000], meta, beam=int(req.get("beam", 1)))
        body = {**res.to_dict(), "diff_html": word_diff(text, res.reconstruction), "notice": NOTICE}
        self._send(200, json.dumps(body, ensure_ascii=False).encode(), "application/json")

    def log_message(self, *a):  # quiet
        pass


def main() -> None:
    ap = argparse.ArgumentParser()
    ap.add_argument("--model", required=True)
    ap.add_argument("--port", type=int, default=8765)
    ap.add_argument("--host", default="127.0.0.1")
    ap.add_argument("--device", default="auto")
    a = ap.parse_args()
    H.rec = HTRReconstructor.from_pretrained(a.model, device=a.device)
    print(f"HTR demo on http://{a.host}:{a.port}  (model {H.rec.meta.get('name')})")
    ThreadingHTTPServer((a.host, a.port), H).serve_forever()


if __name__ == "__main__":
    main()