#!/usr/bin/env python3
"""Ollama shim for nightmare — binds 100.103.34.0:11440 (tailnet only), proxies /api/generate
to localhost:11434 and returns ONE plain, non-chunked response. Binds tailnet IP only so no
new exposure beyond the tailnet."""
import json
import http.server, socketserver, urllib.request

LOCAL = "http://127.0.0.1:11434"
BIND = ("100.103.34.0", 11440)

class H(http.server.BaseHTTPRequestHandler):
    protocol_version = "HTTP/1.0"          # no keep-alive, no chunked
    def do_GET(self):
        if self.path == "/api/tags":
            body = urllib.request.urlopen(LOCAL + "/api/tags", timeout=10).read()
            self.send_response(200); self.send_header("Content-Length", str(len(body)))
            self.end_headers(); self.wfile.write(body)
        else:
            self.send_response(404); self.end_headers()
    def do_POST(self):
        n = int(self.headers.get("Content-Length", 0))
        payload = json.loads(self.rfile.read(n) or b"{}")
        payload["stream"] = False
        req = urllib.request.Request(LOCAL + "/api/generate",
                                     data=json.dumps(payload).encode(),
                                     headers={"Content-Type": "application/json"})
        with urllib.request.urlopen(req, timeout=600) as r:
            body = r.read()
        self.send_response(200)
        self.send_header("Content-Type", "application/json")
        self.send_header("Content-Length", str(len(body)))
        self.end_headers()
        self.wfile.write(body)
    def log_message(self, *a):
        pass

class TS(socketserver.ThreadingTCPServer):
    allow_reuse_address = True
    daemon_threads = True

if __name__ == "__main__":
    TS(BIND, H).serve_forever()