Add 'bencher' console-script entry point for the render/compare CLI

[blooop]

What

Register a bencher console-script entry point so the render/compare CLI is a real command instead of only python -m bencher.render.

[project.scripts]
bencher = "bencher.render:main"

After install:

bencher <result.pkl> <out_dir> [--json result.json]   # render (+ optional JSON export)
bencher compare <a.pkl> <b.pkl> --json comparison.json # machine-readable diff

Why

python -m bencher.render is non-obvious and easy to mistype — the module name render doesn't hint that it also does JSON export and result comparison. A first-time user (or an agent told to "use the bencher CLI") has to go discover the module path. A named command matches the mental model and makes bencher --help self-documenting.

Non-breaking

This is purely additive:

• python -m bencher.render … is unchanged and keeps working — same main().
• No argument names, subcommand structure, or exit codes changed.
• The only behaviour change is help/usage text, which is now invocation-aware via a small _prog() helper: it reads bencher under the console script and python -m bencher.render from a source checkout, so the displayed command always matches how the tool was run.

Verification

• test/test_render.py — added test_prog_name_is_invocation_aware; full module green (22 passed). Existing CLI tests drive render.main([...]) directly and are unaffected.
• Built + installed the wheel in an isolated venv and confirmed the bencher script resolves and dispatches:
  • bencher --help → usage: bencher …
  • bencher compare --help → usage: bencher compare …
  • no-arg → Usage: bencher <result_path> <output_dir> [--json PATH], exit 2
• ruff check / ruff format / prek run all clean.

Version bumped 1.107.0 → 1.108.0 (minor — new feature) with a CHANGELOG entry, per the auto-publish-on-version-bump convention.

Summary by Sourcery

Introduce a named bencher CLI entry point while keeping the existing python -m bencher.render invocation working unchanged.

New Features:

• Add a bencher console-script entry point that invokes the existing render/compare CLI main entry.

Enhancements:

• Make CLI usage and help text invocation-aware so it reflects whether the tool is run via the bencher script or python -m bencher.render.

Build:

• Register the bencher console script in pyproject.toml and bump the project version to 1.108.0.

Documentation:

• Document the new bencher CLI entry and behavior in the changelog.

Tests:

• Add a test ensuring the computed program name matches the actual invocation for usage/help output.

[sourcery-ai]

Reviewer's Guide

Adds a new bencher console-script entry point wired to bencher.render:main, and makes CLI help/usage text dynamically reflect how the command was invoked, while keeping the existing python -m bencher.render entrypoint fully backward compatible. Also bumps the project version and documents the change in the changelog, plus adds tests to pin the new invocation-aware behaviour.

Sequence diagram for bencher CLI invocation paths

sequenceDiagram
    actor User
    participant Shell
    participant bencher_script as bencher_console_script
    participant python_module as bencher.render
    participant main as main
    participant prog as _prog

    User->>Shell: bencher ...
    Shell->>bencher_script: execute entry point
    bencher_script->>python_module: import bencher.render
    bencher_script->>main: main(argv)
    main->>prog: _prog()
    prog-->>main: bencher
    main-->>User: usage/help with bencher

    alt python -m bencher.render
        User->>Shell: python -m bencher.render ...
        Shell->>python_module: python -m bencher.render
        python_module->>main: main(argv)
        main->>prog: _prog()
        prog-->>main: python -m bencher.render
        main-->>User: usage/help with python -m bencher.render
    end

Loading

File-Level Changes

Change	Details	Files
Introduce an invocation-aware program-name helper and update CLI parsers/usage to use it so help text matches the actual command used (console script vs python -m).	Add a private _prog() helper that inspects sys.argv[0] and returns bencher when invoked via the console script, otherwise python -m bencher.render. Switch _render_parser() and _compare_parser() to use _prog() for their prog value, with the compare parser building f"{_prog()} compare". Update main() docstring examples to reference bencher instead of python -m bencher.render and describe both invocation forms. Change the no-argument usage error message to print a Usage: line built from _prog() so it tracks the invocation form.	bencher/render.py
Add tests to cover the new invocation-aware _prog() behaviour.	Add test_prog_name_is_invocation_aware to assert _prog() returns bencher when sys.argv[0] looks like the installed console script and python -m bencher.render otherwise, using mock.patch on bencher.render.sys.argv. Keep existing CLI tests unchanged so they continue to exercise render.main([...]) directly.	test/test_render.py
Register a bencher console-script entry point and bump project version with corresponding changelog entry.	Add a [project.scripts] section mapping bencher to "bencher.render:main" so installers generate a bencher console script. Increment the project version from 1.107.0 to 1.108.0 to reflect the new feature. Document the new bencher CLI entry point, its non-breaking nature, and the invocation-aware _prog() helper in a new 1.108.0 changelog entry.	pyproject.toml CHANGELOG.md

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey - I've found 1 issue, and left some high level feedback:

• The _prog() helper assumes the console script name is exactly bencher, which may not hold on some platforms or packaging setups (e.g., bencher.exe, bencher-3.12); consider a more robust detection (e.g., exe.startswith("bencher") or checking sys.argv[0] for a suffix) so usage/help stays accurate in those cases.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- The `_prog()` helper assumes the console script name is exactly `bencher`, which may not hold on some platforms or packaging setups (e.g., `bencher.exe`, `bencher-3.12`); consider a more robust detection (e.g., `exe.startswith("bencher")` or checking `sys.argv[0]` for a suffix) so usage/help stays accurate in those cases.

## Individual Comments

### Comment 1
<location path="bencher/render.py" line_range="142" />
<code_context>
+    to the explicit ``python -m bencher.render`` form otherwise, so the usage
+    line always names a command the reader can actually run.
+    """
+    exe = Path(sys.argv[0]).name
+    return "bencher" if exe == "bencher" else "python -m bencher.render"
+
</code_context>
<issue_to_address>
**issue (bug_risk):** Handle Windows console-script suffixes when deriving the program name.

On Windows, installed console scripts are usually invoked via `bencher.exe` or `bencher-script.py`, so `Path(sys.argv[0]).name` will rarely equal `"bencher"`. That means `_prog()` will still report `"python -m bencher.render"`, even when users actually run `bencher`, which makes the usage text misleading for Windows users.

To avoid this, make the comparison extension-agnostic, for example by stripping common suffixes before comparing, or simply using `Path(sys.argv[0]).stem` instead of `.name` when checking for `"bencher"`.
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[github-actions]

Performance Report for e01506e

Metric	Value
Total tests	1681
Total time	123.71s
Mean	0.0736s
Median	0.0020s

Top 10 slowest tests

Test	Time (s)
test.test_bench_examples.TestBenchExamples::test_example_meta	17.120
test.test_over_time_save_perf::test_save_faster_without_aggregated_tab	5.072
test.test_generated_examples::test_generated_example[regression/example_regression_tuning_drift.py]	4.519
test.test_generated_examples::test_generated_example[cartesian_animation/example_cartesian_animation.py]	2.965
test.test_split_render_examples::test_split_render_subprocess_media	2.953
test.test_hash_persistent.TestCrossProcessDeterminism::test_hash_stable_across_two_processes[ResultBool]	2.815
test.test_over_time_repeats.TestMaxSliderPoints::test_default_subsampling_caps_at_max	2.810
test.test_generated_examples::test_generated_example[regression/example_regression_tuning_noise.py]	2.787
test.test_generated_examples::test_generated_example[result_types/result_image/example_result_image_to_video.py]	2.777
test.test_generated_examples::test_generated_example[regression/example_regression_tuning_step.py]	2.491

Full report

Updated by Performance Tracking workflow

[github-actions]

Performance Report for 496ba5b

Metric	Value
Total tests	1681
Total time	124.41s
Mean	0.0740s
Median	0.0020s

Top 10 slowest tests

Test	Time (s)
test.test_bench_examples.TestBenchExamples::test_example_meta	17.703
test.test_over_time_save_perf::test_save_faster_without_aggregated_tab	4.969
test.test_generated_examples::test_generated_example[regression/example_regression_tuning_drift.py]	4.547
test.test_split_render_examples::test_split_render_subprocess_media	2.989
test.test_generated_examples::test_generated_example[cartesian_animation/example_cartesian_animation.py]	2.948
test.test_generated_examples::test_generated_example[regression/example_regression_tuning_noise.py]	2.785
test.test_generated_examples::test_generated_example[result_types/result_image/example_result_image_to_video.py]	2.782
test.test_hash_persistent.TestCrossProcessDeterminism::test_hash_stable_across_two_processes[ResultBool]	2.724
test.test_generated_examples::test_generated_example[regression/example_regression_tuning_step.py]	2.493
test.test_over_time_repeats.TestMaxSliderPoints::test_default_subsampling_caps_at_max	2.409

Full report

Updated by Performance Tracking workflow