| """A small local HTTP server for Dewpoint, using only the standard library. |
| |
| python serve.py # ensemble, port 8000 |
| python serve.py --single --port 9000 # the 307M member only |
| python serve.py --backend onnx # no torch needed |
| |
| curl -s localhost:8000/restore -H "content-type: application/json" \\ |
| -d '{"text": "what time is it in tokyo right now", "lang": "en"}' |
| -> {"text": "What time is it in Tokyo right now?"} |
| |
| POST /restore {"text": str | [str], "lang": "en", "labels": false} |
| GET /health {"ok": true, "backend": ..., "members": [...]} |
| |
| It serves one request at a time, which is right for a single model on one |
| machine; put it behind a process manager for anything larger. |
| """ |
| import argparse |
| import json |
| import os |
| import sys |
| from http.server import BaseHTTPRequestHandler, HTTPServer |
|
|
| sys.path.insert(0, os.path.dirname(os.path.abspath(__file__))) |
| from dewpoint import Punctuator, split_words |
|
|
|
|
| def make_handler(p): |
| class H(BaseHTTPRequestHandler): |
| def _send(self, code, obj): |
| body = json.dumps(obj, ensure_ascii=False).encode("utf-8") |
| self.send_response(code) |
| self.send_header("content-type", "application/json; charset=utf-8") |
| self.send_header("content-length", str(len(body))) |
| self.end_headers() |
| self.wfile.write(body) |
|
|
| def do_GET(self): |
| if self.path.rstrip("/") == "/health": |
| self._send(200, {"ok": True, "backend": p.backend, |
| "members": len(p.members)}) |
| else: |
| self._send(404, {"error": "try POST /restore or GET /health"}) |
|
|
| def do_POST(self): |
| if self.path.rstrip("/") != "/restore": |
| return self._send(404, {"error": "POST /restore"}) |
| try: |
| n = int(self.headers.get("content-length", 0)) |
| req = json.loads(self.rfile.read(n) or b"{}") |
| text, lang = req.get("text", ""), req.get("lang", "en") |
| texts = [text] if isinstance(text, str) else list(text) |
| out = [] |
| for t in texts: |
| item = {"text": p.restore(t, lang)} |
| if req.get("labels"): |
| r = p.predict(split_words(t, lang), lang) |
| item.update(words=r["words"], punct=r["punct"], case=r["case"]) |
| out.append(item) |
| self._send(200, out[0] if isinstance(text, str) else out) |
| except Exception as e: |
| self._send(400, {"error": f"{type(e).__name__}: {e}"}) |
|
|
| def log_message(self, fmt, *args): |
| pass |
|
|
| return H |
|
|
|
|
| def main(): |
| ap = argparse.ArgumentParser(description="Serve Dewpoint over HTTP.") |
| ap.add_argument("--host", default="127.0.0.1") |
| ap.add_argument("--port", type=int, default=8000) |
| ap.add_argument("--model", default=os.path.dirname(os.path.abspath(__file__))) |
| ap.add_argument("--single", action="store_true") |
| ap.add_argument("--backend", default="auto", choices=["auto", "torch", "onnx"]) |
| a = ap.parse_args() |
| p = Punctuator.from_pretrained(a.model, backend=a.backend, |
| members=["mmbert-base"] if a.single else None) |
| print(f"Dewpoint ({p.backend}, {len(p.members)} member(s)) on http://{a.host}:{a.port}") |
| HTTPServer((a.host, a.port), make_handler(p)).serve_forever() |
|
|
|
|
| if __name__ == "__main__": |
| main() |
|
|
