WIP refactor of the profile generator

[kirberich]

Hi,

I've been spending some time over the last two weeks trying to refactor the profile generator with a slightly more modular approach - this is currently only done for Simplify3D profiles, but the idea is that the same logic can be reused for Cura as well, cutting down on a lot of duplication in the code. It brings in a dependency on Jinja2 to render profiles, keeping the rendering separate from the logic.

I've made some successful single extrusion prints with results from the refactored generator, but there are some differences in some calculations that make the results difficult to compare to the original.

I'm putting this up here mostly as a starting point for a discussion - it'd be quite difficult to actually merge this as it is, as there are a lot of potential regressions. So mostly I wanted show this around to see how you feel about this and to ask for some help concerning some of the calculations that might currently be wrong because I'm lacking some context.

Some notes:

• My code doesn't use the synchronised heating sequence - the code is very complicated and I didn't want to get too distracted. I'd be curious to know the reasons behind trying to bring everything to temperature at the same time (my code just heats up everything at the same time)
• The original code sometimes uses layer height references specific to one extruder, even if that layer height isn't used (as the smaller layer height always gets picked). Because of this, my code currently sometimes produces invalidly high flow values for the lower-quality settings.
• I've only tested the code with my current printer configuration (0.4mm left, 0.6mm right)
• I think some of my speed calculations are off - dual extrusion prints slow down to a crawl probably because of some wrong maths on my part. Any feedback on the speed/temperature/etc calcs would be hugely appreciated.
• I made the mistake of basing my work on master rather than development. I'll try to update it for the new changes in development, but I think this might turn out to be quite tricky
• I'm not sure I understand the logic for fan speeds - they seem to go up linearly with the print temperature, but my understanding was that for materials like PLA, more fan is almost always better.

Please let me know what you think! Again I don't think this should be merged as it is, but hopefully if I can get some more context I can make my code behave just like the old code, integrate changes from the development branch, and add cura profile generation, to bring it to a state where it can be merged.