Meta-Eval/EvalHarness
3
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Output polish: 1-based phase indices everywhere, drop [1/1] prefix on single-bench runs, slim progress bar (drop Waiting/Last columns)

Browse Source 
Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
sora 2026-09-10 09:16:33 +00:00
parent 2dfb60cdf2
commit b6c473ac26
2 changed files with 8 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
evalharness/cli.py
View File

@ -191,7 +191,9 @@ def _print_run_plan(console, args, model_spec):
def _print_phase(console, index, total, name, message): def _print_phase(console, index, total, name, message):
text = f'[{index}/{total}] {name}: {message}' # single-benchmark runs: the [1/1] tag is noise, drop it
prefix = f'[{index}/{total}] ' if total > 1 else ''
text = f'{prefix}{name}: {message}'
if console is not None: if console is not None:
console.print(f'[cyan]{text}[/cyan]') console.print(f'[cyan]{text}[/cyan]')
else: else:
@ -302,11 +304,12 @@ def _cmd_eval_run(args) -> int:
# writer during the live bar interleave incorrectly # writer during the live bar interleave incorrectly
progress_reporter = RichTerminalProgress(console=console) progress_reporter = RichTerminalProgress(console=console)
def status_callback(msg, _idx=i, _name=name, def status_callback(msg, _idx=i + 1, _name=name,
_reporter=progress_reporter, _reporter=progress_reporter,
_console=console): _console=console):
if _reporter is not None: if _reporter is not None:
_reporter.log(f'[{_idx}/{total_runs}] {_name}: {msg}') tag = f'[{_idx}/{total_runs}] ' if total_runs > 1 else ''
_reporter.log(f'{tag}{_name}: {msg}')
else: else:
_print_phase(_console, _idx, total_runs, _name, msg) _print_phase(_console, _idx, total_runs, _name, msg)
report = asyncio.run(run_eval( report = asyncio.run(run_eval(

6
evalharness/progress/rich_terminal.py
View File

@ -26,10 +26,8 @@ class RichTerminalProgress:
BarColumn(complete_style="green", finished_style="bold green"), BarColumn(complete_style="green", finished_style="bold green"),
TaskProgressColumn(), TaskProgressColumn(),
TextColumn("• Completed {task.completed}/{task.total}"), TextColumn("• Completed {task.completed}/{task.total}"),
TextColumn("• In flight {task.fields[inflight]}"), TextColumn("• in-flight {task.fields[inflight]}"),
TextColumn("• Waiting {task.fields[waiting]}"), TextColumn("• {task.fields[rate]}/s"),
TextColumn("• Last {task.fields[last_result]}"),
TextColumn("• {task.fields[rate]} sample/s"),
TextColumn("•"), TextColumn("•"),
TimeElapsedColumn(), TimeElapsedColumn(),
TextColumn("•"), TextColumn("•"),