WLJS Notebook is modular project.
Please, contact @JerryI before working on code and etc.
Here is some to-do list grouped by repos
server, IO, UI, cells manager, render process and client app
- Electron app auto-updater. Use Github releases to update client app automatically (probably won't work for linux)
- tutorials/examples for new users (no WL background) on Wolfram Language
- autotests in a format of WLJS notebook, which go though most features
- more example on differnet topics (for data-scientists, scientists, students, ...)
Main repo is stable and there are no urget needs in any refactoring
-
Inset
works incorrectly onImage
and otherGraphics
objects. -
PlotLegends
are not rendered inside aGraphics
object (acts like a wrapper). Exported figures do not show legends.
- misssing implementations for
FilledCurve
, ... -
GraphicsComplex
with manyPolygon
is too slow. Especially it can be seen onContourPlot
, since it produces tons of SVG paths.
-
AxesLabel
,Ticks
, does not have any effect.PlotRange
is absent. -
Text
is not implemented in the context ofGraphics3D
-
RTX
feature shows incorrect faces of polygons (black triangles) on OSX (probably also depends on a browser)
-
TableForm
does not support styling -
Dataset
does not support nested sets with asscoiations or other datasets inside. -
Entity
is not supported
- get rid of double backslashes from LaTeX somehow (single backslash causes problems with encoding)
- get rid of double backslashes from LaTeX somehow (single backslash causes problems with encoding)
- video preview is extremely slow, decoding from video to a set of images is a main bottleneck.
- no sound (need to be used together with PCM streamer from wljs-audio)
-
Sound
does not works with nested list ofSoundNote
-
Sound[SoundNote[...]]
needs a proper syntax sugar for showing piano roll and possible note names
- Wolfram Language Documentation (frontend agnostic and in a Markdown format)
- support for syntax sugar for
NeuralNetwork
package of Wolfram Language. For now it is not clear, how to clean up Mathematica'sMakeBoxes
definitons for various neural network object. There is a possibility to use wrapper function, see more in Issue. - write a fast implementation of parser/encoder of WXF in Javascript. For now we use
ExpressionJSON
for communication, which does not use features ofNumericArray
andByteArray
and adds a large overhead in serializaing/deserializing on WL's and JS's sides. There is a preliminaly work posted already.