criterium.viewer.common.regression — org.hugoduncan/criterium 0.5.245-ALPHA

Domain regression data preparation functions.

Provides functions to prepare regression model data for table and chart rendering, including model fit data, log-log analysis, and an orchestration function for rendering regression views.

format-log-log-slope clj

(format-log-log-slope slope)

Format log-log slope as complexity class estimate. Returns string like 'O(n^1.02)' or 'O(n)' for integer slopes.

prepare-log-log-fit-line clj

(prepare-log-log-fit-line log-log-data {:keys [_axis _impl-axis]})

Generate fit line points for log-log plot. Returns vector of point maps with x, y, and optionally impl key.

Uses the pre-computed slope and intercept: y = slope * x + intercept where x = log(n), y = log(metric).

prepare-log-log-points clj

(prepare-log-log-points log-log-data {:keys [axis _impl-axis]})

Prepare log-log transformed data points for scatter plot. Returns {:points [...] :axis-name string} or nil.

Points have keys: x (log(n)), y (log(metric)), and optionally yLower, yUpper for log-transformed error bounds. For multi-impl mode, points also have :impl key.

The log-log-data comes from the :regressions map of a :criterium/domain-log-log-regression result.

prepare-log-log-residuals clj

(prepare-log-log-residuals log-log-data {:keys [_axis _impl-axis]})

Compute residual points for log-log plot. Returns vector of point maps with x (log(n)), residual, and optionally impl.

Residuals are in log space: residual = log(y) - (slope * log(x) + intercept)

prepare-regression-fit-lines clj

(prepare-regression-fit-lines {:keys [x-vals total-scale]}
                              {:keys [models by-impl impl-keys]})

Generate fit line points for plotting. For single-impl mode, models is a seq of model maps. For multi-impl mode, by-impl is a map of impl-key -> {:models [...] :best-fit id}. Returns vector of point maps with x, y, and model or impl key.

prepare-regression-model-table clj

(prepare-regression-model-table
  {:keys [models best-fit]}
  {:keys [best-fit-marker plotted-marker tolerance]
   :or {best-fit-marker "✓" plotted-marker "" tolerance 0.01}})

Prepare model table rows for single-impl regression display. Returns vector of row maps with :model :r-squared :aic :bic :equation :best-fit keys. AIC and BIC values may be nil when insufficient data points. Options: :best-fit-marker - string to show for best fit (default "✓") :plotted-marker - string to show for plotted but not best (default "")

:tolerance - fraction within best r-squared to mark as plotted (default 0.01)

prepare-regression-model-table-multi-impl clj

(prepare-regression-model-table-multi-impl by-impl impl-keys options)

Prepare model table rows for multi-impl regression display. Returns vector of row maps with :implementation :model :r-squared :aic :bic :equation :best-fit.

Options same as prepare-regression-model-table.

prepare-regression-points clj

(prepare-regression-points extract-data
                           {:keys [axis impl-axis has-error-bounds? metric]})

Prepare data points for regression scatter plot. Returns {:points [...] :total-scale number :unit string :x-vals [...]} or nil. Points have keys: x, y, and optionally yLower, yUpper for error bounds. For multi-impl mode, points also have :impl key.

prepare-regression-residuals clj

(prepare-regression-residuals {:keys [valid-data total-scale]}
                              {:keys [axis impl-axis has-error-bounds? models
                                      by-impl impl-keys]})

Compute residual points for plotting. Returns vector of point maps with x, residual, and model or impl key.

with-domain-regression-data clj

(with-domain-regression-data data-map opts handlers)

Orchestrate domain-regression rendering by iterating over metrics and implementations, calling the provided render functions at appropriate points.

The opts map must contain: :regression-id - key for regression in data-map (default :regression) :extract-id - key for extract in data-map (default :extract) :log-log-id - key for log-log in data-map (default :log-log) :tolerance - tolerance for selecting plotted models (default 0.01) :table-options - options for prepare-regression-model-table functions

The handlers map must contain render functions: :render-log-log-charts - (fn [{:keys [title axis metric impl-axis points line-pts residual-pts chart-opts]}] ...) :render-model-heading - (fn [{:keys [title axis metric impl-axis]}] ...) :render-model-table - (fn [{:keys [table-rows multi-impl?]}] ...) :render-regression-charts - (fn [{:keys [points line-pts residual-pts y-title residual-title axis chart-opts multi-impl?]}] ...)

For multi-impl mode, the impl-key is included in callback maps. For single-impl mode, slope/r-squared from log-log are included for display.