Recalculate SVG viewbox after combining

[dmadison]

Currently the combined SVG output from generate_2d.py uses the dimensions and viewbox from the first merged file. This causes the combined SVG to render incorrectly in browsers and most viewing software, as the content extends outside of the viewing bounds. To fix it, this PR adds a few functions to the Python scripts which recalculate and set the new dimension and viewboxes for the combined SVG based off of the imported files.

Before:

After:

(screen captures from Inkscape 1.0.1, 3bc2e813f5, 2020-09-07 - Using View->Zoom->Drawing)

As files are added to the combined SVG (import_paths()), the imported viewbox is compared against the current one. If the imported viewbox is larger in any dimension, the current viewbox is expanded to include it. The output is then saved to the SVG attributes (viewbox and dimensions).

Originally I started down this road when I noticed that the SVG dimensions were wrong on Windows. This was caused by the fix_dimens function, which added 'mm' to the combined output. It looks as though OpenSCAD started adding 'mm' to the SVG dimension output a few years ago, after the 'fix' function was written. On Windows the combined SVG had dimensions in "mmmm" which caused all sorts of weird sizing issues. This fixes that bug by proxy, since the width and height dimensions (including the 'mm' annotation) are replaced wholesale based off of the viewbox.

To demonstrate the updating viewbox here's a gif showing each stage in the combining process. Each SVG was exported during the render_svgs loop in projection_renderer.py, opened in Chrome, then screenshotted and combined into a gif in Photoshop:

As this change doesn't touch the paths themselves it shouldn't have any effect on the final parts. It just makes the combined SVG easier to view and work with.

[scottbez1]

This is great, I'll try to review in the next few days!