A request has come in for an updated version of the old area type. As mentioned by @etpinard in #2255:
How to replace legacy polar's area trace type? I'd vote for replacing it with polarbar or wedge. Most of the bar logic could be reused here.
Yes, it seems like this trace type is precisely bar transformed onto polar axes. For consistency with other trace type names perhaps it should be barpolar though?
Looks to me as though essentially all the existing bar trace attributes will translate to the new type straightforwardly. In fact, there's even an analog of orientation, though instead of (x|y) we'd have (radial|angular). For the request at hand only radial (which matches the legacy area type) is necessary, but some day someone will want angular as well, like https://peltiertech.com/time-is-on-my-side/
The layout attributes (barmode, barnorm, bargap, bargroupgap) to the extent that we want them at all, we can make into attributes of the polar<N> subplot. I would suggest omitting (at least initially) barmode: 'group' and 'relative', and therefore also bargroupgap. Probably also omit barnorm, that only enhances the confusion over whether it's the radius or area that encodes value.
A request has come in for an updated version of the old
areatype. As mentioned by @etpinard in #2255:Yes, it seems like this trace type is precisely
bartransformed onto polar axes. For consistency with other trace type names perhaps it should bebarpolarthough?Looks to me as though essentially all the existing
bartrace attributes will translate to the new type straightforwardly. In fact, there's even an analog oforientation, though instead of(x|y)we'd have(radial|angular). For the request at hand onlyradial(which matches the legacyareatype) is necessary, but some day someone will wantangularas well, like https://peltiertech.com/time-is-on-my-side/The
layoutattributes (barmode,barnorm,bargap,bargroupgap) to the extent that we want them at all, we can make into attributes of thepolar<N>subplot. I would suggest omitting (at least initially)barmode: 'group'and'relative', and therefore alsobargroupgap. Probably also omitbarnorm, that only enhances the confusion over whether it's the radius or area that encodes value.