Releases: JuliaPlots/PlotlyJS.jl
Releases · JuliaPlots/PlotlyJS.jl
Close plot windows
reinstated function to close plot window.
Julia 1.0!
TODO: will write later. Big news is:
- julia 0.7+ compatible (including 1.0)
- Uses WebIO for all display code
- uses ORCA.jl for savefig
Proper support for plotly mimetype
Enhancements
- forward implementation of plotly mimetype for PlotlyJS.SyncPlot on to PlotlyBase.Plot.
- Update REQUIRE to reflect this new dependency
IJulia and build fixes
Bug fixes:
- Only start an IJulia comm if IJulia has been inited. Closes #181 see 99b98f9
- Fix paths in build.jl file. Thanks @EricForgy ref #183
Enhancements
- Added an auto-generated CHANGELOG.md file (ref 03207e1)
Persistent plots in Jupyter notebook (when connected to internet)
Plotly Base
Changes
- All plot creation routines (including the
Plot
,Layout
,GenericTrace
, and associated types) have been moved to the PlotlyBase.jl (ref #171 )
Improvements
- For the electron display, we have CI tests for plotly.js API methods (
restyle
,relayout
,update
,addtraces
,deletetraces
,movetraces
,extendtraces
,to_image
, andpurge
) (see 0cfcb4c) - Also for electron display, we are testing the ability of the
savefig
routine to run for files with extensions[".plotly.json", "json", "html", "svg", "pdf", "png", "eps"]
(see 0cfcb4c)
Bug fixes
- A docstring error in the DataFrames API has been fixed (that code now lives inside PlotlyBase.jl) (ref #169)
New traces, examples, and DataFrame api functionality
New
- New trace types
violin
,scatterpolar
, andscatterpolargl
(2c909b7) - Docs for violin and table
- In DataFrame api, can apply group logic to trace functions instead of just plots (037e84)
- When using DataFrame api, can now pass a function as a value for a keyword argument. This function is evaluated with the DataFrame and the return value from the function is used as the trace attribute (037e84)
Improvements
avoid race condition with Blink.jl v0.6.0 using up to date electron
Bug fixes
- Blink.jl v0.6.0 updated to a recent version of electron. This had unintended consequences and created race conditions for Julia code trying to load resources into a blink window immediately after the window is created. To get around this, we ask the window to evaluate
1+1
in a while loop. If the window is ready we get back2
and move on. If the window is not ready we get back and empty dict, sleep for 0.01 seconds, and try again. We do this for 2 seconds (up to 200 retries). In my testing I get that we usually get success after < 10 tries (see 35c8c77).
Bug in DataFrames api
3d subplots
Bug fixes
- Fix subplot builder code so that it handles putting multiple 3d plots into subplots