AxisAnchor AxisReference

[schlichtanders]

the current AxisAnchor / AxisReference implementation looks unnecessarily limited.
Putting both together, currently the only options for axisanchor are "x", "x1", ..., "x3", "y", ..., "y4", "free"

for instance there can only 5 subplots right now

Maybe this is a plotly limitation, I don't know, but if not, please consider taking off this limiting factor of subplots.

[schlichtanders]

again run into this problem that the number of subplots is limited by having this sealed abstract class AxisReference with only a couple of instances.

is there a need to make this sealed?

look for instance https://plot.ly/python/subplots/#multiple-custom-sized-subplots
there they use at least up to x6,y6

the syntax definitely suggest that it should seamlessly work for arbitrary many AxisReferences

Please, any comment is highly apprecitated
(I am currently thinking about writing my own thinner plotly.js wrapper which does not have such restrictions)

[sbrunk]

The current restriction might just be for historical reasons because plotly upstream is constantly evolving (I haven't checked though).

Do you want to look into it and perhaps create a PR with a more generic solution for AxisAnchor and AxisReference instead of creating your own wrapper?

[schlichtanders]

This is unfortunately not trivial because the AxisReference is interpreted as Enum

plotly-scala/render/shared/src/main/scala/plotly/internals/ArgonautCodecsInternals.scala

Lines 122 to 123 in 6a8f629

	implicit val axisAnchorIsEnum = IsEnum.instance[AxisAnchor](_.label)
	implicit val axisReferenceIsEnum = IsEnum.instance[AxisReference](_.label)

However, I would need a plain String...
Haven't worked with Argonaut yet

[schlichtanders]

for those who stumble upon this and need a quick solution from scala:

• use import shapeless.record._ and val trace = Record(x = Seq(1,2,3,5), y = Seq(1,3,14), name = "title", `type` = "scatter")
• together with import io.circe.shapes._ circe can transform arbitrary Record to the correct json string
• paste everything together in your custom plotly html file

super tiny, but with this you can already access all plotly features however recent they may be

Remarks on Type Support:

• this will give you type support in the sense of that you can still access types correctly,
• however this won't give you type support in the sense that if you use a wrong attribute name or attribute type, things will break only at html/javascript level