Wishlist for demos

[olexandr-konovalov]

Which other demos would be good to have here?

• Francy
• JupyterViz
• Meataxe64
• HPC-GAP

[stevelinton]

Add syntaxtree. It should be possible to demo some kind of very crude automatic parallelisation, turning

for i in [1..n] do
    y[i] := foo(x[i]);
od;

into

tasks := [];
for i in [1..n] do
   tasks[i] := RunTask(foo,x[i]);
od;
for i in [1..n] do
   y[i] := TaskResult(tasks[i]);
od;
Unbind(tasks);

or something like that.

Maybe also include something with an SCSCP server farm.

[olexandr-konovalov]

Add SemigroupViz demo - this is https://github.com/nathancarter/semigroupviz by @nathancarter, not SgpViz.

[nthiery]

At this rate, it will soon be nice to run planetaryum on this repo :-)

[olexandr-konovalov]

Suggested by @stevelinton in D5.15: MajoranaAlgebras by Markus Pfeiffer and Madeleine Whybrow, which constructs Majorana representations of finite groups (added in GAP4.10.1).

[olexandr-konovalov]

Pushed very much initial version of JupyterViz demo with an incomplete map of locations of GAP package authors and some heuristics implemented to determine countries if they are not stated in PackageInfo.g

[nathancarter]

Nice!!

[olexandr-konovalov]

Thanks @nathancarter! I will update the data cleaning script and offload most of it to a .g file, but if you have any suggestions how the map could be made better, they will be much appreciated!

[olexandr-konovalov]

@stevelinton for HPC-GAP in Jupyter see gap-packages/JupyterKernel#108. Getting it to work requires gap-system/gap#3710 and gap-packages/ZeroMQInterface#26. Print does not work though.

[olexandr-konovalov]

@nathancarter is there an option to put different marks on the line plot (instead of circles, use also triangles and squares)? We want it for the plot at the bottom of https://github.com/OpenDreamKit/gap-demos/blob/master/notebooks-master/Meataxe64%20Demo.ipynb

[nathancarter]

Yes. To specify detailed options, you have to dig inside a bit, but it goes like so:

LoadPackage( "jupyterviz" );

# let's assume you have your list of numbers to plot here...
numbers := [ 3, 5, 4, 6, 5, 7 ];

# we package it up for use by the Plotly graphing tool:
series := ConvertDataSeriesForTool.plotly( [
    JUPVIZMakePlotDataSeries( numbers ) ] );

# then we tweak one of Plotly's defaults, to make the markers different:
series.data[1].marker := rec( size := 20, symbol := "cross-dot" );
# (see https://plot.ly/javascript/reference/#scatter-marker-symbol
# for all the different types of markers you can use)

# then we have JupyterViz draw it, using Plotly:
CreateVisualization( rec( tool := "plotly", data := series ) );

Does that do what you want?