labnotes 0.0.107__tar.gz → 0.0.109__tar.gz

{labnotes-0.0.107 → labnotes-0.0.109}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: labnotes
-Version: 0.0.107
+Version: 0.0.109
 Summary: Browse experiment designs, notes, and run data from a folder.
 Project-URL: Repository, https://github.com/thekevinscott/labnotes
 Author-email: Kevin Scott <me@thekevinscott.com>
@@ -122,7 +122,8 @@ line per event instead.
 - With runs going, Ctrl-C asks first. Enter (or a second Ctrl-C) leaves them
   running. `s` stops the running scripts and records their end, then kills any
   still going 5 seconds later. Escape goes back. SIGTERM stops them like `s`.
-- A second `run` of the same experiment fails while the first holds the lock.
+- A second `run` of the same experiment fails with "already running" while the
+  daemon still has the first going.
 - It exits 0 even if runs fail; the summary line counts failures. If it can't
   start or reach the daemon, it prints the cause on one line and exits 1.
 

{labnotes-0.0.107 → labnotes-0.0.109}/README.md

@@ -98,7 +98,8 @@ line per event instead.
 - With runs going, Ctrl-C asks first. Enter (or a second Ctrl-C) leaves them
   running. `s` stops the running scripts and records their end, then kills any
   still going 5 seconds later. Escape goes back. SIGTERM stops them like `s`.
-- A second `run` of the same experiment fails while the first holds the lock.
+- A second `run` of the same experiment fails with "already running" while the
+  daemon still has the first going.
 - It exits 0 even if runs fail; the summary line counts failures. If it can't
   start or reach the daemon, it prints the cause on one line and exits 1.
 

{labnotes-0.0.107 → labnotes-0.0.109}/src/labnotes/context.md

@@ -77,8 +77,9 @@ Each param is appended as `--name value`, with the name as written (use
 and tables as JSON. labnotes writes `run.json` (params, created, ended, exit
 code, `ended_by`, `signal`), `stdout.log` and `stderr.log` to the
 run folder. `ended_by` is `"timeout"`, `"stop"` (labnotes stopped it),
-`"signal"` (a signal labnotes didn't send, like the out-of-memory killer) or
-null (the script exited itself). `signal` is the signal number read from the
+`"signal"` (a signal labnotes didn't send, like the out-of-memory killer),
+`"lost"` (its process died unseen; the daemon marks these when it starts, with
+`ended` set and `exit_code` null) or null (the script exited itself). `signal` is the signal number read from the
 exit code (negative, or 128 + N from the shell), else null.
 A run counts when it exited 0 and its `run.json` doesn't hold
 `"excluded": true`, so excluding a run by hand means adding that key; the next

labnotes-0.0.109/src/labnotes/daemon/finalize_lost_runs.py

@@ -0,0 +1,24 @@
+import json
+import time
+
+from labnotes.list_runs import UNFINISHED
+from labnotes.models import LabnotesConfig
+from labnotes.runner.write_run_json import write_run_json
+from labnotes.tables.iso_time import iso_time
+from labnotes.tables.open_tables import open_tables
+from labnotes.utils.process_alive import process_alive
+
+
+async def finalize_lost_runs(config: LabnotesConfig) -> None:
+    rows = await open_tables(config).query(
+        f"SELECT folder, run FROM runs WHERE {UNFINISHED}"
+    )
+    for row in rows:
+        run = json.loads(str(row["run"]))
+        pid, created = run.get("pid"), run.get("pid_created")
+        if isinstance(pid, int) and process_alive(
+            pid, created if isinstance(created, int | float) else None
+        ):
+            continue
+        run |= {"ended": iso_time(time.time()), "ended_by": "lost"}
+        write_run_json(config.root / str(row["folder"]), run)

{labnotes-0.0.107 → labnotes-0.0.109}/src/labnotes/daemon/run_daemon.py

@@ -6,11 +6,13 @@ import uvicorn
 from labnotes.daemon.daemon_app import daemon_app
 from labnotes.daemon.daemon_server import DaemonServer
 from labnotes.daemon.exit_when_idle import exit_when_idle
+from labnotes.daemon.finalize_lost_runs import finalize_lost_runs
 from labnotes.daemon.socket_path import socket_path
 from labnotes.models import LabnotesConfig
 
 
 async def run_daemon(config: LabnotesConfig) -> None:
+    await finalize_lost_runs(config)
     path = socket_path(config.root)
     app = daemon_app(config)
     server = DaemonServer(

labnotes-0.0.109/src/labnotes/utils/process_alive.py

@@ -0,0 +1,13 @@
+import psutil
+
+from labnotes.utils.tree_rss import SAME_START_SECONDS
+
+
+def process_alive(pid: int, created: float | None) -> bool:
+    try:
+        started = psutil.Process(pid).create_time()
+    except psutil.NoSuchProcess:
+        return False
+    except psutil.AccessDenied:
+        return True
+    return created is None or abs(started - created) <= SAME_START_SECONDS