Advanced html snapshot

I have to include multiple K3D plots into a sphinx documentation page. I thought I could insert the snapshot html file with:

.. raw:: html
	:file: figs/tut/fig-01.html

But it doesn’t work, in the console i see this error:

Uncaught ReferenceError: pako is not defined

Uncaught Error: Mismatched anonymous define() module: function(e){"use strict";var t=(e,t,i,n)=>{let a=65535&e|0,r=e>>>16&65535|0,s=0;for(;0!==i;){s=i>2e3?2e3:i,i-=s;do{a=a+t[n++]|0,r=r+a|0}while(--s);a%=65521,r%=65521}return a|r<<16|0};const i=new Uint32Array((()=>{let e,t=[];for(var i=0;i<256;i++){e=i;for(var n=0;n<8;n++)e=1&e?3988292384^e>>>1:e>>>1;t[i]=e}return t})());var n=(e,t,n,a)=>{const r=i,s=a+n;e^=-1;for(let i=a;i<s;i++)e=e>>>8^r[255&(e^t[i])];return-1^e};var a=function(e,t){let i,n,a,r,s,o,l,d,f,h,c,u,w,b,k,m,_,g,v,p,y,x,E,R;const A=e.state;i=e.next_in,E=e.input,n=i+(e.avail_in-5),a=e.next_out,R=e.output,r=a-(t-e.avail_out),s=a+(e.avail_out-257),o=A.dmax,l=A.wsize,d=A.whave,f=A.wnext,h=A.window,c=A.hold,u=A.bits,w=A.lencode,b=A.distcode,k=(1<<A.lenbits)-1,m=(1<<A.distbits)-1;e:do{u<15&&(c+=E[i++]<<u,u+=8,c+=E[i++]<<u,u+=8),_=w[c&k];t:for(;;){if(g=_>>>24,c>>>=g,u-=g,g=_>>>16&255,0===g)R[a++]=65535&_;else{if(!(16&g)){if(0==(64&g)){_=w[(65535&_)+(c&(1<<g)-1)];continue t}if(32&g){A.mode=12;bre…
    makeError https://cdnjs.cloudflare.com/ajax/libs/require.js/2.3.4/require.min.js:1
    ...

I think the requirejs included with the snapshot is in conflict with the requirejs of the documentation page.

Would it be possible to create an advanced snapshot that:

1. separates the different javascript libraries into their own files. For example, pako.js, k3d.js… In doing so, when multiple plots need to be included in the same document, the overall file size would be (potentially) much smaller.
2. It assigns a unique ID to the wrapper div, maybe some kind of hash based on the data, or an option passed in by the user. In doing so, it minimizes the overall chances of conflicts when including multiple plots in the same document.
3. Allows the possibility to create an html file without the html/head/body tags. This will create a file ready to be inserted into other documents.
4. It minimizes all the output files.

For reference, Plotly does something similar with the Interactive HTML export, by setting fig.write_html(path, include_plotlyjs="require", full_html=False). The html files created with that command are ready to be inserted into a documentation page: it works like a charm.

[CagtayFabry]

hi @Davide-sd @artur-trzesiok

Since I found this issue when looking into integrating k3d into a sphinx documentation I just wanted to let you know that as of k3d=2.12 the interactive output now also works in jupyter notebooks when using nbsphinx to convert/integrate the notebooks into the documentation

I simply have a notebook containing the k3d plot code that is normally displayed in JupyterLab when I work on the notebooks - super easy! have a look: https://weldx.readthedocs.io/en/latest/tutorials/01_03_geometry.html#Export-3d-geometry-into-a-CAD-file

[Davide-sd]

Thanks @artur-trzesiok for your work, really appreciated.