update to plotly 1.11.0

[timelyportfolio]

update to plotly 1.11.0

[cpsievert]

The message below was automatically generated after build https://travis-ci.org/ropensci/plotly/builds/131021508

Detected 3 differences ->

http://cpsievert.github.io/plotly-test-table/cb04782/jitter-basic
http://cpsievert.github.io/plotly-test-table/cb04782/scatterplots-geom_jitter
http://cpsievert.github.io/plotly-test-table/cb04782/scatterplots-jitter

[cpsievert]

@timelyportfolio do you think it makes sense to reflect plotly.js' semantic versioning in the R package? In other words, since plotly.sj bumped from 1.10.1 to 1.11.0, should we bump the R package from 3.6.0 to 3.7.0?

[timelyportfolio]

@cpsievert I was just wondering that. It makes sense, but compare the big subplot bump to this little puny one. To the R user, little changes. I vote for just 0.0.1 bump, but I could argue either way :)

[cpsievert]

yea, that has been my approach thus far, so let's stick with that.

Go ahead and bump the version, update news, and you have my 💃

[cpsievert]

The message below was automatically generated after build https://travis-ci.org/ropensci/plotly/builds/131028280

Detected 3 differences ->

http://cpsievert.github.io/plotly-test-table/6684abc/jitter-basic
http://cpsievert.github.io/plotly-test-table/6684abc/scatterplots-geom_jitter
http://cpsievert.github.io/plotly-test-table/6684abc/scatterplots-jitter

[talgalili]

My two cents (which basically supports your positions):

I follow http://semver.org/ where it says that given a version number MAJOR.MINOR.PATCH, increment the:
MAJOR version when you make incompatible API changes,
MINOR version when you add functionality in a backwards-compatible manner, and
PATCH version when you make backwards-compatible bug fixes.

I like this distinction between functionality and bug fixes.
So, from my perspective, if plotly.js upgrade mainly dealt with bug fixes it should be a PATCH bump, but if it dealt with new functionality (such as with the recent subplot) - then MINOR bump should be made (with update to NEWS on what is new for the R user).

Thank you both for the great support of this package!

Tal

[talgalili]

Following my previous note. I am looking at the release notes of plotly 1.11.0:
https://github.com/plotly/plotly.js/releases/tag/v1.11.0
And it says:
Add top-level methods Plotly.toImage to convert a plotly graph to an image data URL (svg, png, jpg, and webp are supported) and Plotly.downloadImage to download a plotly graph as an image [#446]
Add the ability to add arbitrary images loaded from a url to a plot's layout [#525]
Add the option of making legend span horizontally [#535]

If this is also reflected in the R options, then I would not consider this a patch addition, this is new (and useful!) functionality.

Tal

[timelyportfolio]

Good point @talgalili on bumping minor on support of the new functionality in R. None have been implemented on the R side yet. Any in particular appeal to you?

[talgalili]

Thanks @timelyportfolio
Yes, the following (by order) are (IMHO) important to add:

• Add top-level methods Plotly.toImage to convert a plotly graph to an image data URL (svg, png, jpg, and webp are supported) and Plotly.downloadImage to download a plotly graph as an image [Pass plot configuration using ... to avoid conflicts in defaults/documentation #446] - this is VERY useful. (people will interact, find an angle of the data they care about - and save a file with it)
• Add the ability to add arbitrary images loaded from a url to a plot's layout [Legend Scroll issues. #525] -

this is nice, it enable implementing "annotation_custom".
For example:

# source: http://stackoverflow.com/questions/9917049/inserting-an-image-to-ggplot2
library(png)
library(grid)
library(ggplot2)
img <- readPNG(system.file("img", "Rlogo.png", package="png"))
g <- rasterGrob(img, interpolate=TRUE)

p <- qplot(1:10, 1:10, geom="blank") +
  annotation_custom(g, xmin=-Inf, xmax=Inf, ymin=-Inf, ymax=Inf) +
  geom_point()

p

library(plotly)
ggplotly(p) # currently doesn't show the figure

But also annotation_raster:

# source: http://stackoverflow.com/questions/27637455/display-custom-image-as-geom-point
library(png)
library(ggplot2)

img <- readPNG(system.file("img", "Rlogo.png", package="png"))

p <- ggplot(mtcars, aes(mpg, wt)) + 
       mapply(function(xx, yy) 
          annotation_raster(img, xmin=xx-1, xmax=xx+1, ymin=yy-0.2, ymax=yy+0.2),
          mtcars$mpg, mtcars$wt) 


p

library(plotly)
ggplotly(p) # doesn't work

• Add the option of making legend span horizontally [Use transparent scatter lines for errorbars. Fixes #513. #535]
  This is nice to have