Crate perfgate_app

Expand description

Application layer for perfgate.

The app layer coordinates adapters and domain logic. It does not parse CLI flags and it does not do filesystem I/O.

Modules§

baseline_resolve: Baseline path resolution logic.
comparison_logic: Logic for building budgets and metric statistics for comparisons.

CheckOutcome: Outcome of the check use case.
CheckRequest: Request for the check use case.
CheckUseCase: Use case for running a config-driven check.
CompareExportRow: Row structure for CompareReceipt export.
CompareRequest
CompareResult: Result from CompareUseCase including host mismatch information.
CompareUseCase
ExportUseCase: Use case for exporting receipts to different formats.
PairedRunOutcome
PairedRunRequest
PairedRunUseCase
PromoteRequest: Request for promoting a run receipt to become a baseline.
PromoteResult: Result of a promote operation.
PromoteUseCase: Use case for promoting run receipts to baselines.
ReportRequest: Request for generating a report from a compare receipt.
ReportResult: Result of a report generation operation.
ReportUseCase: Use case for generating perfgate reports.
RunBenchOutcome
RunBenchRequest
RunBenchUseCase
RunExportRow: Row structure for RunReceipt export.
SensorCheckOptions: Options for run_sensor_check.
SensorReportBuilder: Builder for constructing a SensorReport from a PerfgateReport.
SystemClock

BenchOutcome: A single bench’s outcome for aggregation into a sensor report.
ExportFormat: Supported export formats.

classify_error: Classify an error into (stage, error_kind) for structured error reporting.
default_engine_capability: Build a default engine capability based on the current platform.
direction_str: Return a human-readable label for a budget Direction.
format_metric: Return the canonical string key for a Metric.
format_metric_with_statistic: Format a metric key, appending the statistic name when it is not the default (median).
format_pct: Format a fractional change as a percentage string.
format_value: Format a metric value for display.
github_annotations: Produce GitHub Actions annotation strings from a CompareReceipt.
markdown_template_context: Build the JSON context object used by render_markdown_template.
metric_status_icon: Return an emoji icon for a MetricStatus.
metric_status_str: Return a lowercase string label for a MetricStatus.
parse_reason_token: Parse a verdict reason token like "wall_ms_warn" into its metric and status.
render_markdown: Render a CompareReceipt as a Markdown table for PR comments.
render_markdown_template: Render a CompareReceipt using a custom Handlebars template.
render_reason_line: Render a single verdict reason token as a human-readable bullet line.
run_sensor_check: Run a sensor check and return a SensorReport directly.
sensor_fingerprint: Build a fleet-standard fingerprint from semantic parts.