b1.ticks — b1 0.3.3

coverage clj/s

(coverage d-min d-max l-min l-max)

Objective function based on distances between extreme data and extreme labels

density clj/s

(density r rt)

Objective function for a candidate density r and desired density rt (e.g. labels-per-cm)

index-of clj/s

(index-of x coll)

Index of x in coll

label-range-contains-zero? clj/s

(label-range-contains-zero? l-min l-max l-step)

max-coverage clj/s

(max-coverage d-min d-max span)

When the label range is centered on the data range

max-density clj/s

(max-density r rt)

max-simplicity clj/s

(max-simplicity q j)

Q clj/s

Preference-ordered list of nice step sizes

search clj/s

(search [d-min d-max]
        &
        {:keys [target-density length clamp?]
         :or {target-density 0.01 length 500 clamp? false}})

Find best ticks for the data range [d-min, d-max]. Returns a map with {:min :max :step :extent :ticks} of optimal labeling (if one is found). Returns an empty map if no labelings can be found.

Kwargs:

:target-density labels per length, defaults to 0.01 (one label per 100 units)

:length available label spacing

:clamp? don't return ticks outside of data range, defaults to false.

Since there are no test input/output datasets for the labeling algorithm, I played it safe and copied the imperative algorithm from the paper. If you rewrite it in an understandable and performant functional style, I'll accept a pull request and buy you a bottle of whiskey.

simplicity clj/s

(simplicity q j label-range-contains-zero)

Objective function modeling niceness of step sizes and whether a range includes zero