[hytop] Add `--wait-idle-seconds`, move `--timeout` to the root command

projects/hytop/README.md

@@ -4,6 +4,7 @@
 
 ```bash
 uv pip install -e .
+hytop --help
 hytop gpu --help
 ```
 
@@ -11,7 +12,23 @@ hytop gpu --help
 
 - Python >= 3.10
 - Python packages: `rich`, `typer`
-- Passwordless SSH for remote monitoring
+- Passwordless SSH for remote 
+
+## `hytop`
+
+```bash
+# Show the version number
+hytop --version
+
+# Specify a timeout for the subcommand
+hytop --timeout 300 [COMMAND]
+
+# 0.5-second interval and 5-second rolling window for the subcommand
+hytop -n 0.5 --window 5 [COMMAND]
+
+# Specify a list of nodes for the subcommand
+hytop -H node01,node02 [COMMAND]
+```
 
 ## `hytop gpu`
 
@@ -22,15 +39,18 @@ A lightweight script for live `hy-smi` polling with rolling averages across mult
 Simple examples:
 
 ```bash
-# Local node, all GPUs, 5-second rolling window
-hytop -n 1 --window 5 gpu
+# Local node, all GPUs
+hytop gpu
 
-# Two nodes, monitor only GPU 0 and 1
-hytop -H node01,node02 -n 1 gpu --devices 0,1
+# Two nodes, 0.5-second interval
+hytop -H node01,node02 -n 0.5 gpu
 
 # Exit with code 0 when all monitored GPUs are available
 hytop gpu --devices 0,1 --wait-idle
 
+# Wait for GPUs to be idle for 30 seconds before exiting
+hytop gpu --devices 0,1 --wait-idle --wait-idle-seconds 30
+
 # Wait at most 300s for availability (exit 124 on timeout)
 hytop gpu --devices 0,1 --wait-idle --timeout 300
 
@@ -42,7 +62,7 @@ hytop gpu --showpower --showtemp
 Queue jobs in shared environments:
 
 ```bash
-if hytop -H node01,node02 gpu --wait-idle --timeout 300; then
+if hytop -H node01,node02 gpu --timeout 300 --wait-idle; then
   echo "GPUs available, starting workload..."
   # YOUR COMMAND HERE (e.g., python train.py)
 else

projects/hytop/src/hytop/gpu/cli.py

@@ -42,7 +42,12 @@ def gpu(
     wait_idle: bool = typer.Option(
         False,
         "--wait-idle",
-        help="Exit 0 when all monitored GPUs have zero VRAM/HCU avg in the configured window.",
+        help="Exit 0 when all monitored GPUs have zero VRAM/HCU avg.",
+    ),
+    wait_idle_seconds: float = typer.Option(
+        10.0,
+        "--wait-idle-seconds",
+        help="How long GPUs must stay idle before exiting. Effective only with --wait-idle.",
     ),
     showtemp: bool = typer.Option(
         False,
@@ -74,11 +79,6 @@ def gpu(
         callback=remember_show_flag_callback,
         help="Display GPU utilization.",
     ),
-    timeout: Optional[float] = typer.Option(
-        None,
-        "--timeout",
-        help="Max runtime in seconds. Effective only with --wait-idle.",
-    ),
 ) -> None:
     """GPU monitoring commands."""
 
@@ -90,6 +90,7 @@ def gpu(
         host_list = ctx.obj["hosts"]
         interval = ctx.obj["interval"]
         window_value = ctx.obj["window"]
+        timeout_value = ctx.obj.get("timeout")
         selected_show_flags = {
             "showtemp": showtemp,
             "showpower": showpower,
@@ -111,11 +112,6 @@ def gpu(
         parsed_device_filter: Optional[Set[int]] = None
         if device_filter:
             parsed_device_filter = set(parse_csv_ints(device_filter, "--devices"))
-        timeout_value = (
-            float(timeout)
-            if timeout is not None
-            else None
-        )
     except ValueError as exc:
         typer.echo(f"argument error: {exc}", err=True)
         raise typer.Exit(code=2) from exc
@@ -127,6 +123,7 @@ def gpu(
         window=window_value,
         interval=interval,
         wait_idle=wait_idle,
+        wait_idle_duration=max(wait_idle_seconds, interval),
         timeout=timeout_value,
     )
     raise typer.Exit(code=code)

projects/hytop/src/hytop/gpu/service.py

@@ -262,6 +262,7 @@ def run_monitor(
     interval: float,
     wait_idle: bool,
     timeout: Optional[float],
+    wait_idle_duration: float = 10.0,
 ) -> int:
     """Run the asynchronous collector + periodic renderer monitor loop.
 
@@ -271,6 +272,7 @@ def run_monitor(
         window: Rolling window length in seconds.
         interval: Sampling interval in seconds.
         wait_idle: Whether to exit when all monitored GPUs become idle.
+        wait_idle_duration: How long GPUs must stay idle before exiting.
         timeout: Optional timeout for wait-idle mode.
 
     Returns:
@@ -334,7 +336,7 @@ def run_monitor(
                         refresh=True,
                     )
                     elapsed_since_start = time.monotonic() - started
-                    warmup_done = elapsed_since_start >= state.max_window
+                    warmup_done = elapsed_since_start >= wait_idle_duration
                     if (
                         wait_idle
                         and warmup_done

projects/hytop/src/hytop/main.py

 from __future__ import annotations
 
+from typing import Optional
+
 import typer
 
 from hytop import __version__
@@ -50,11 +52,17 @@ def root(
         "--window",
         help="Single rolling window in seconds. Default: 5.0",
     ),
+    timeout: Optional[float] = typer.Option(
+        None,
+        "--timeout",
+        help="Max runtime in seconds.",
+    ),
 ) -> None:
     """Root callback that parses global options and stores them in context."""
     try:
         host_list = parse_csv_strings(hosts, "--hosts")
         window_value = parse_positive_float(str(window), "--window")
+        timeout_value = float(timeout) if timeout is not None else None
     except ValueError as exc:
         typer.echo(f"argument error: {exc}", err=True)
         raise typer.Exit(code=2) from exc
@@ -63,6 +71,7 @@ def root(
         "hosts": host_list,
         "interval": interval,
         "window": window_value,
+        "timeout": timeout_value,
     }