chore: DO NOT MERGE - benchmark tsam develop-v4 vs v3.4.x

[FBumann]

DO NOT MERGE — measurement experiment, not a dependency change proposal.

Pins tsam to the develop-v4 git branch via [tool.uv.sources] so the Benchmarks workflow's head-vs-base comparison measures tsam v4 against released v3.4.x on the identical suite (no lockfile is committed, so base resolves the v3 release line while head resolves develop-v4). Includes the #109 benchmark suite commits since the workflow and cases must exist on the head.

Early local signal (90-day config cases, min of 5):

• dist_global: ~87 ms → ~53 ms (~1.6× faster — consistent with the v4 rework around durationRepresentation)
• medoid / mean / dist_cluster: ~35 ms → ~19-21 ms (~1.7× faster)
• v4's result object now holds lazy _accuracy_metrics / _reconstructed_data internals, so the eager accuracy cost (62% of a v3 run) may have moved or vanished — the CI table will show where.

All 542 wrapper tests pass against develop-v4 locally (v4-only features like preserve_n_clusters from FZJ-IEK3-VSA/tsam#410 are not yet on that branch — the wrapper's feature detection correctly reports them absent).

🤖 Generated with Claude Code

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 8e92b137-9d9c-490c-a653-97f40c5777c9

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch chore/tsam-v4-bench

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Benchmarks

Δ% vs base, one table sorted by biggest change first
(min walltime and memray peak). Peak memory is
deterministic — any delta there is real; walltime on GitHub
runners is noisy, so only large time deltas are meaningful.
The interactive plot is attached as the benchmark-plot artifact.

Full table

benchmarks/test_bench_aggregate.py

name	metric	base	head
test_wrapper_to_dataframe	time	0.0003684s	+17.2%
	peak	8.55 MiB	0.0%
test_user_disaggregate	time	0.01331s	+3.4%
	peak	26.7 MiB	0.0%
test_wrapper_concat_results	time	0.0223s	-0.3%
	peak	8.72 MiB	0.0%
test_e2e_wide[dist_global]	time	2.034s	-1.7%
	peak	104 MiB	-0.9%
test_e2e_full	time	0.1646s	-3.4%
	peak	6.6 MiB	-2.1%
test_e2e_wide[medoid]	time	0.557s	-3.8%
	peak	113 MiB	-8.2%
test_user_apply	time	0.4492s	-4.0%
	peak	104 MiB	-0.9%
test_config_representation[dist_global]	time	0.1399s	-5.6%
	peak	1.78 MiB	-4.1%
test_config_representation[dist_global_minmax]	time	0.1427s	-6.1%
	peak	1.78 MiB	-4.1%
test_e2e_slices	time	1.327s	-6.1%
	peak	14.3 MiB	0.0%
test_wrapper_result_conversion	time	0.04064s	-10.1%
	peak	60.3 MiB	0.0%
test_e2e_default	time	0.08795s	-13.1%
	peak	7.11 MiB	-9.1%
test_config_representation[medoid]	time	0.05438s	-13.4%
	peak	1.9 MiB	-9.9%
test_config_extremes[replace]	time	0.05605s	-14.1%
	peak	1.9 MiB	-10.1%
test_config_representation[mean]	time	0.05148s	-14.3%
	peak	1.78 MiB	-4.1%
test_config_representation[dist_cluster]	time	0.05351s	-14.9%
	peak	1.78 MiB	-4.1%
test_config_extremes[append]	time	0.05628s	-14.9%
	peak	1.9 MiB	-10.2%
test_e2e_multidim	time	0.2545s	-20.7%
	peak	26.7 MiB	-8.9%