pytest_park.models

source package pytest_park.models

Classes

• BenchmarkCase — A single benchmark case in a run.

• BenchmarkDelta — A comparison result for one benchmark case.

• BenchmarkRun — A full benchmark run loaded from one JSON artifact.

• BenchmarkStats — Core benchmark statistics from pytest-benchmark.

• GroupSummary — Aggregated comparison metrics for a logical group.

• ImprovementSummary — Aggregated improvement metrics across all methods.

• MethodHistoryComparison — A method mean observation compared against a reference run baseline.

• MethodHistoryPoint — A single mean observation for a method in one run.

• MethodImprovement — Aggregated improvement metrics for a method within a group.

• OverviewStatistics — Accumulated comparison statistics across all benchmark deltas.

• PriorRunComparison — Comparison of a candidate method mean against one prior reference run.

• SplitBarRow — Original vs new mean pair for one argument combination.

• TrendPoint — Time-series data for one case and run.

source dataclass BenchmarkCase(name: str, fullname: str, normalized_name: str, normalized_fullname: str, base_name: str, method_parameters: str | None, method_postfix: str | None, benchmark_group: str | None, marks: tuple[str, ...], params: dict[str, str], custom_groups: dict[str, str], stats: BenchmarkStats)

A single benchmark case in a run.

Attributes

• case_key : str — Build a deterministic key for cross-run comparisons.

source property BenchmarkCase.case_key: str

Build a deterministic key for cross-run comparisons.

source dataclass BenchmarkDelta(group_label: str, case_key: str, benchmark_name: str, params: dict[str, str], reference_run_id: str, candidate_run_id: str, reference_mean: float, candidate_mean: float, delta_pct: float, speedup: float)

A comparison result for one benchmark case.

source dataclass BenchmarkRun(run_id: str, source_file: str, created_at: datetime | None, tag: str | None, commit_id: str | None, machine: str | None, python_version: str | None, metadata: dict[str, Any] = field(default_factory=dict), cases: list[BenchmarkCase] = field(default_factory=list), profiler: dict[str, dict[str, Any]] = field(default_factory=dict))

A full benchmark run loaded from one JSON artifact.

source dataclass BenchmarkStats(mean: float, median: float, min: float, max: float, stddev: float, rounds: int, iterations: int, ops: float)

Core benchmark statistics from pytest-benchmark.

source dataclass GroupSummary(label: str, count: int, average_delta_pct: float, median_delta_pct: float, improvements: int, regressions: int, unchanged: int)

Aggregated comparison metrics for a logical group.

source dataclass ImprovementSummary(count: int, avg_vs_orig_time: float | None = None, avg_vs_orig_pct: float | None = None, med_vs_orig_time: float | None = None, med_vs_orig_pct: float | None = None, min_vs_orig_time: float | None = None, min_vs_orig_pct: float | None = None, max_vs_orig_time: float | None = None, max_vs_orig_pct: float | None = None, avg_vs_prev_time: float | None = None, avg_vs_prev_pct: float | None = None, med_vs_prev_time: float | None = None, med_vs_prev_pct: float | None = None, min_vs_prev_time: float | None = None, min_vs_prev_pct: float | None = None, max_vs_prev_time: float | None = None, max_vs_prev_pct: float | None = None)

Aggregated improvement metrics across all methods.

source dataclass MethodHistoryComparison(run_id: str, timestamp: str | None, method: str, distinct: str, mean: float, reference_mean: float, delta_pct: float, speedup: float)

A method mean observation compared against a reference run baseline.

source dataclass MethodHistoryPoint(run_id: str, timestamp: str | None, method: str, distinct: str, mean: float)

A single mean observation for a method in one run.

source dataclass MethodImprovement(group: str, method: str, current_benchmark_name: str | None = None, comparison_benchmark_name: str | None = None, original_benchmark_name: str | None = None, orig_arg_count: int = 0, ref_arg_count: int = 0, avg_vs_orig_time: float | None = None, avg_vs_orig_pct: float | None = None, med_vs_orig_time: float | None = None, med_vs_orig_pct: float | None = None, min_vs_orig_time: float | None = None, min_vs_orig_pct: float | None = None, max_vs_orig_time: float | None = None, max_vs_orig_pct: float | None = None, avg_vs_prev_time: float | None = None, avg_vs_prev_pct: float | None = None, med_vs_prev_time: float | None = None, med_vs_prev_pct: float | None = None, min_vs_prev_time: float | None = None, min_vs_prev_pct: float | None = None, max_vs_prev_time: float | None = None, max_vs_prev_pct: float | None = None)

Aggregated improvement metrics for a method within a group.

source dataclass OverviewStatistics(count: int, avg_delta_pct: float, median_delta_pct: float, avg_speedup: float, improved: int, regressed: int, unchanged: int)

Accumulated comparison statistics across all benchmark deltas.

source dataclass PriorRunComparison(method: str, candidate_run_id: str, reference_run_id: str, distinct: str, mean: float, reference_mean: float, delta_pct: float, speedup: float, reference_timestamp: str | None)

Comparison of a candidate method mean against one prior reference run.

source dataclass SplitBarRow(argument: str, original: float, new: float, delta_pct: float, speedup: float)

Original vs new mean pair for one argument combination.

source dataclass TrendPoint(run_id: str, timestamp: datetime | None, mean: float)

Time-series data for one case and run.