Taking a look at the code now, it looks like the index.js is prepared to find images in a specific directory structure:
// Create scenes.
var scenes = data.scenes.map(function(data) {
var urlPrefix = "tiles";
var source = Marzipano.ImageUrlSource.fromString(
urlPrefix + "/" +
data.id + "/{z}/{f}/{y}/{x}.jpg",
{ cubeMapPreviewUrl: urlPrefix + "/" +
data.id + "/preview.jpg" });
var geometry = new Marzipano.CubeGeometry(data.levels);
the z/f/y/x.jpg will turn into "0-sala/1/b/0/0.jpg" as an example. Where you see 'b' there are the directories b, d, f, l, r and u, that stands for the back, down, front, left, right and upper view of a cube face image format. The '0-sala' directory in the example is inside the low level 'tiles' directory, which is probably represented by the '
data.id' in the code.
At the data.js file the code specifies the tile and image sizes, like shown below:
"scenes": [
{
"id": "0-sala",
"name": "Sala",
"levels": [
{
"tileSize": 256,
"size": 256,
"fallbackOnly": true
},
{
"tileSize": 512,
"size": 512
},
{
"tileSize": 512,
"size": 1024
},
{
"tileSize": 512,
"size": 2048
}
],
That variable format looks pretty much like json format, which is also equal to python variable format for lists and dictionaries. There the '0-sala' scene is defined with 4 resolution levels. To use a cube face without tiles (which I suggested as a first step) you probably should change that scene to have only one level and the tileSize would be equal to the size, and both equal to the cube face size.
Bests,