From f028aea536a08acb873b5d1cf76aa4dae8570509 Mon Sep 17 00:00:00 2001
From: Derek Feichtinger <derek.feichtinger@psi.ch>
Date: Tue, 30 Jun 2026 15:25:52 +0200
Subject: [PATCH] more preset formats, improve help texts

---
 slurm-eff-tool.py | 72 ++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 58 insertions(+), 14 deletions(-)

diff --git a/slurm-eff-tool.py b/slurm-eff-tool.py
index b3d7e87..dc94043 100755
--- a/slurm-eff-tool.py
+++ b/slurm-eff-tool.py
@@ -118,6 +118,45 @@ PRESET_COLUMNS = {
         "Mem_Eff",
         "Time_Eff",
         "jobname",],
+    "time": [
+        "username",
+        "JobID",
+        "state",
+        "Count",
+        "NTasks",
+        "CPUs",
+        "Nodes",
+        "ReqWalltime",
+        "Walltime",
+        "Walltime_max",
+        "Time_Eff",
+        "jobname",],
+    "cpu": [
+        "username",
+        "JobID",
+        "state",
+        "Count",
+        "NTasks",
+        "CPUs",
+        "Nodes",
+        "CPU_Eff",
+        "waste_CPU",
+        "jobname",],
+    "mem": [
+        "username",
+        "JobID",
+        "state",
+        "Count",
+        "NTasks",
+        "CPUs",
+        "Nodes",
+        "AllocMem",
+        "UsedMem",
+        "MaxRSS_max",
+        "MemPerCPU",
+        "Mem_Eff",
+        "waste_Mem",
+        "jobname",],
 }
 
 # One-character aliases for sorting and output format specifications.
@@ -801,8 +840,8 @@ def aggregate_records(records: list[JobRecord], args: argparse.Namespace) -> lis
     """Aggregate records according to given grouping instructions."""
 
     global default_mempercpu_gb
-    if args.deflt_mpcpu:
-        default_mempercpu_gb = float(args.deflt_mpcpu)
+    if args.dflt_mpcpu:
+        default_mempercpu_gb = float(args.dflt_mpcpu)
 
     if args.aggr_regexp:
         compiled = [(pat, re.compile(pat)) for pat in args.aggr_regexp]
@@ -1112,32 +1151,32 @@ def parse_args(argv: list[str]) -> argparse.Namespace:
                    default=None,
                    help="sacct state filter, e.g. COMPLETED,FAILED,TIMEOUT")
 
-    p.add_argument("-O", "--output-raw", help="write raw sacct output cache to this file")
-    p.add_argument("-F", "--from-raw", help="read raw sacct output cache from this file instead of running sacct")
-    p.add_argument("-B", "--write-binary-cache", help="write a binary cache file in msgpack format")
-    p.add_argument("-L", "--load-binary-cache", help="load a binary cache file in msgpack format")
-    p.add_argument("-i", "--info", help="show information for the given binary cache file")
+    p.add_argument("-O", "--output-raw", help="write raw sacct output cache to this file (text format, large).")
+    p.add_argument("-F", "--from-raw", help="read raw sacct output cache from this file.")
+    p.add_argument("-B", "--write-binary-cache", help="write a binary cache file in gzipped msgpack format")
+    p.add_argument("-L", "--load-binary-cache", help="load a binary cache file in gzipped msgpack format")
+    p.add_argument("-i", "--info", help="show metadata information for the given binary cache file")
     p.add_argument("-U", "--aggr-user", action="store_true", help="aggregate jobs by user, CPUs, nodes, ReqMem, and timelimit")
     p.add_argument(
         "-R",
         "--aggr-regexp",
         action="append",
-        help="aggregate jobs matching regexp by regexp, CPUs, nodes, ReqMem, and timelimit; may be repeated",
+        help="aggregate jobs based on a regexp applied to the jobnames. Option may be repeated multiple times.",
     )
 
-    p.add_argument("--deflt-mpcpu", help=f"Default memory to CPU ratio of cluster ({default_mempercpu_gb} GB/cpu)")
+    p.add_argument("--dflt-mpcpu", help=f"Default memory/CPU ratio of cluster [{default_mempercpu_gb} GB/cpu]. Used in memory waste calculation.")
     p.add_argument("--sdev", action="store_true", help="after each efficiency average, add sdev, max, and min columns")
     p.add_argument("--json", action="store_true", help="emit JSON instead of an ASCII table")
-    p.add_argument("-s", "--sort", help="comma-separated numeric sort columns or aliases; prefix with - for descending")
+    p.add_argument("-s", "--sort", help="comma-separated numeric sort columns or short aliases; prefix with - for descending")
     fmthelp = ", ".join([f"{k}:{ALIASES[k]}" for k in ALIASES])
     p.add_argument(
         "-o",
         "--format",
-        help= f"String of comma separated column names or short names defining the output format:\n{fmthelp}",
+        help= f"String of comma separated column names or short aliases defining the output format [{fmthelp}]",
     )
-    p.add_argument("--expr", help="arithmetic expression using column names",
+    p.add_argument("--expr", help="filter using an arithmetic expression using column names as variables (case insensitive)",
                    default=None)
-    p.add_argument("-p", "--preset", help=f"use one of several preset output column formats ({','.join(PRESET_COLUMNS.keys())})",
+    p.add_argument("-p", "--preset", help=f"use one of several preset output column formats [{','.join(PRESET_COLUMNS.keys())}]",
                    default=None)
 
     args = p.parse_args(argv)
@@ -1160,7 +1199,12 @@ def main(argv: list[str] | None = None) -> int:
 
     output_columns = PRESET_COLUMNS["default"]
     if args.preset:
-        output_columns = PRESET_COLUMNS[args.preset]
+        try:
+            output_columns = PRESET_COLUMNS[args.preset]
+        except KeyError:
+            sys.stderr.write(f"ERROR: no such preset format ({args.preset})\n")
+            sys.exit(1)
+
     if args.format:
         output_columns = columns_from_fmtstr(args.format)