less verbose node hop if not in debug mode

This commit is contained in:
Dobromir Popov
2026-06-30 17:44:21 +03:00
parent 6e4f755e71
commit 5fe471d8ca
5 changed files with 1533 additions and 2 deletions

View File

@@ -38,6 +38,7 @@ class _TorchHTTPServer(http.server.HTTPServer):
tracker_mode: bool = False,
tracker_url: str | None = None,
route_timeout: float = 30.0,
debug: bool = False,
):
super().__init__(addr, handler)
self.backend = backend
@@ -46,6 +47,7 @@ class _TorchHTTPServer(http.server.HTTPServer):
self.tracker_mode = tracker_mode
self.tracker_url = tracker_url
self.route_timeout = route_timeout
self.debug = debug
self.total_requests: int = 0
self.failed_requests: int = 0
self.queue_depth: int = 0
@@ -406,7 +408,8 @@ class _TorchHandler(http.server.BaseHTTPRequestHandler):
current_pos = pos_ids
for hop_index, (node_url, start_layer) in enumerate(route):
print(f" [node] pipeline hop {hop_index}: {node_url} start_layer={start_layer}", flush=True)
if server.debug:
print(f" [node] pipeline hop {hop_index}: {node_url} start_layer={start_layer}", flush=True)
headers: dict[str, str] = {
"Content-Type": "application/octet-stream",
"X-Meshnet-Wire": _WIRE_VERSION,
@@ -440,7 +443,8 @@ class _TorchHandler(http.server.BaseHTTPRequestHandler):
try:
data = json.loads(resp_body)
text = str(data.get("text", ""))
print(f" [node] pipeline hop {hop_index} returned text={text!r}", flush=True)
if server.debug:
print(f" [node] pipeline hop {hop_index} returned text={text!r}", flush=True)
return text
except json.JSONDecodeError:
return resp_body.decode("utf-8", errors="replace")
@@ -611,6 +615,7 @@ class TorchNodeServer:
tracker_mode: bool | None = None,
tracker_url: str | None = None,
route_timeout: float = 30.0,
debug: bool = False,
) -> None:
self._host = host
self._requested_port = port
@@ -624,6 +629,7 @@ class TorchNodeServer:
self._tracker_mode = tracker_mode if tracker_mode is not None else (shard_start == 0)
self._tracker_url = tracker_url
self._route_timeout = route_timeout
self._debug = debug
self._server: _TorchHTTPServer | None = None
self._thread: threading.Thread | None = None
self.port: int | None = None
@@ -666,6 +672,7 @@ class TorchNodeServer:
self._tracker_mode,
self._tracker_url,
self._route_timeout,
self._debug,
)
self.port = self._server.server_address[1]
self._thread = threading.Thread(target=self._server.serve_forever, daemon=True)