Crate perfgate_domain 

Domain logic for perfgate.

Pure, I/O-free business logic: statistics computation, budget policy evaluation, host mismatch detection, and regression analysis. All data comes in via function arguments; no filesystem, network, or process access.

Part of the perfgate workspace.

Structs

BinaryBlame

Result of a binary blame analysis.

BudgetResult

Result of evaluating a single metric against a budget.

Comparison

DependencyChange

Information about a dependency change.

Finding

A single finding in a report.

FindingData

Data for a single finding in a report.

PairedComparison

Result of comparing paired statistics, including significance testing.

Report

Report derived from a CompareReceipt.

SignificancePolicy

TrendAnalysis

Result of trend analysis for a single metric.

TrendConfig

Parameters controlling drift classification thresholds.

Enums

BudgetError

Errors that can occur during budget evaluation.

DependencyChangeType

DomainError

DriftClass

Classification of metric drift direction and severity.

StatsError

Functions

aggregate_verdict

Aggregates multiple metric statuses into a final verdict.

analyze_trend

Perform a full trend analysis on a sequence of metric values.

calculate_regression

Calculates the regression percentage between baseline and current values.

classify_drift

Classify the drift of a metric given regression parameters.

compare_lockfiles

Compares two lockfiles and returns the differences.

compare_paired_stats

Compare paired statistics and compute a confidence interval.

compare_runs

Compare full run receipts under the provided budgets.

compare_stats

Compare stats under the provided budgets.

compute_headroom_pct

Compute headroom as a percentage: how far the current value is from the threshold.

compute_paired_cv

Compute the coefficient of variation (CV) of the wall-time differences from a set of paired samples (excluding warmups).

compute_paired_stats

Compute summary statistics from paired benchmark samples.

compute_significance

Compute statistical significance using Welch’s t-test.

compute_stats

Compute perfgate stats from samples.

derive_report

Derives a report from a CompareReceipt.

detect_host_mismatch

Detect host mismatches between baseline and current runs.

determine_status

Determines the metric status based on regression and thresholds.

evaluate_budget

Evaluates a metric against a budget threshold.

evaluate_budgets

Evaluates multiple metrics against their budgets.

linear_regression

Fit a simple linear regression: y = slope * x + intercept.

mean_and_variance

Compute sample mean and unbiased variance (Bessel’s correction).

median_f64_sorted

median_u64_sorted

metric_value

parse_lockfile

Parses a Cargo.lock string and returns a map of package name to version.

predict_breach_run

Predict the run index at which the regression line crosses threshold.

reason_token

Generates a reason token for a metric status.

spark_chart

Render a mini ASCII spark chart for a series of values.

summarize_f64

Compute min, max, and median for an f64 slice.

summarize_u64

Compute min, max, and median for a u64 slice.