expose scales #425
expose scales #425
Conversation
Fil
added
enhancement
|
Same concept (legend drawn with Plot) for a radius legend |
accepts a function that receives the scale; if that function returns a DOM node, adds it to the figure.
|
Now we can add |
… be used to recreate the scale in another plot. In particular it doesn't show the (D3) scale anymore. Instead, a convenience function Plot.scale(options) is provided, that will turn a scale *options* object into a D3 scale, ready to use in another context. (This also clears the question of mutability.)
|
For anyone following this, 228565b exposes Plot.scale(options), allowing to create a D3 scale from Plot options. |
Co-authored-by: Mike Bostock <[email protected]>
Co-authored-by: Mike Bostock <[email protected]>
ensure we don't compute the scale options twice always return label, clamp, type
|
Color legends in #432 ! |
|
I have to work again on this, since some options are missing (like exponent: 0.3 for type="exp" scales), and we want all options including those that are not consumed by the scales, but could be used by the legend (like "columns" for swatches). |
…", "identity")
and type ("utc", "time" / "linear", "sqrt", "log"… / "categorical" / "identity")
except the following which are part of the definition of: scale.domain() // domain, nice, scale.range() // range, inset, reverse, scale.interpolate() // interpolate, clamp, round paddingInner & paddingOuter // padding
|
I plan on looking at this next. Exposing (materialized) scales will be great for debugging, too. I’m leaning towards just exposing plot.scales to start, and not the plot.scale method which returns a D3 scale instance. But mostly that’s so I can focus on getting this landed sooner and minimizing new API. 🙂 |
|
deferring to #484 |
The scales are exposed on the figure through a .scales property, a function that lazily computes scale option objects that can be used to reproduce the same scales in a different plot.
The function takes a scale name as parameter, and returns a scale definition that matches the given scale as instanciated by the plot's data. (If no scale name is given, it returns all the scales.)
With a lot of unit tests.
A (very messy) notebook with the build is https://observablehq.com/@fil/expose-scales-425
TODO: