Loading SVG files for Icons (annotations).

[achary]

SVGs are permitted image types (as per Modelica specification). Support for SVG as a valid image type usable in annotation is added.

Related Issues

#10102

Purpose

as in the issue.

Approach

1. Adding support for new file type (extension).
2. Checking base64 encoded content (in case image is stored inline in the icon annotation itself) does represent the original SVG file, and not a rendered pixmap.
   The app decides, based on the decoded base64 content, whether it deals with one of binary formats (PNG, JPG) that can be simply pushed to QPixmap and alike, without worrying about them, or whether it's SVG and needs slightly different rendering path. For that it constructs one of appropriate image renderers.
3. This in turn, ensures that SVG images are rendered on demand according to their actual display size, so we get a good visual effect - no zoomed pixels, no edge blurs.

This PR where it stands, makes SVG files first-class citizens and improves quality of their renderings in the UI (can be zoomed in without pixelization).

Screens

The SVG file doc/images/logo.svg was used here:

and with larger zoom:

[achary]

FYI: the "bitmapFileTypes" image filter has its usage for selecting bitmap textures somewhere in 3D animation viewer, so not extending this one with SVG as it's not the intention of this PR to alter this accidentally.

[casella]

https://stackoverflow.com/questions/3312607/php-binary-image-data-checking-the-image-type
https://stackoverflow.com/questions/64656124/determine-if-file-is-svg

[achary]

@casella, are there any formatting standards you stick to in this project?
I've only found info about trailing whitespaces in the
How about intentation, max line length, ordering includes, brackets etc?
Clang-format etc? no?

Writing because my editor settings (almost) by default makes plenty of changes (like include order sorting/grouping). Are you accepting such formatting changes mixed with functional changes? What's your policy here?

Sorry for plenty of question at once. Better to clarify upfront...

[adeas31]

@achary read this https://github.com/OpenModelica/OpenModelica/tree/master/OMEdit#coding-style

For includes we first include our own files and then the Qt headers.
There are no standard settings for max line length. For brackets, use the standard GNU style.

[achary]

@adeas31 Thanks. I think it only scratches the surface, but I'll try to stick.

[casella]

@achary I guess you can also get an idea by looking at the existing codebase.

[achary]

FYI: This is the root place where the change actually needed to be applied at. Instead of always drawing a bitmap image, now, we now delegate the painting operation to whatever the mpRenderer (runtime polymorphic type) has been constructed. Bitmap images render pretty much the same way (see BitmapRenderer class) and SVG images use QSvgRenderer under the hood instead (see `SvgRenderer class).

[achary]

The change ended up being slightly more involving than I've imagined prior seeing the code.
There was a bit of refactoring made of the ShapeAnnotation and BitmapAnnotation classes.

IMO the ShapeAnnotation class is much waaaay too heavy, as it contains data fields for all possible occasions (like for instance font details, which apply only to TextAnnotation only etc.) - worth considering refactoring that too in the future.

[achary]

BTW: Some CI on Jenkins jobs looks unstable, giving false negatives.

The job testsuite-matlab-translator gave two different results (first fail, then pass) on two different runs over the same code (a trivially amended Git commit).

[casella]

@adeas31 can you please review and merge if you think this is OK? Thanks!

[sjoelund]

BTW: Some CI on Jenkins jobs looks unstable, giving false negatives.

The job testsuite-matlab-translator gave two different results (first fail, then pass) on two different runs over the same code (a trivially amended Git commit).

You need to check the logs to see what is wrong:

ryzen-5950x-2-3 does not seem to be running inside a container
$ docker run -t -d -u 1001:1001 -w /var/lib/jenkins3/ws/OpenModelica_PR-10104 -v /var/lib/jenkins3/ws/OpenModelica_PR-10104:/var/lib/jenkins3/ws/OpenModelica_PR-10104:rw,z -v /var/lib/jenkins3/ws/OpenModelica_PR-10104_tmp:/var/lib/jenkins3/ws/OpenModelica_PR-10104_tmp:rw,z -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** -e ******** docker.openmodelica.org/build-deps:v1.16.3 cat
 ---> Running in 685d52f784a9
ERROR: Timeout after 180 seconds

Sometimes jobs simply timeout... Or crash inside of Jenkins itself and it has nothing to do with the code. Sometimes it's GitHub that is slow.

[adeas31]

I didn't look at the implementation details yet.

I compiled the code and it is not working as expected. For example, look at Modelica.Mechanics.MultiBody.Examples.Systems.RobotR3.FullRobot. Also create a new model and try adding a bitmap to it.

[adeas31]

FYI I also get message,

QIODevice::read (QFile, "modelica:\Modelica\Resources\Images\Mechanics\MultiBody\Examples\Systems\RobotR3\robot_kr15.png"): device not open

[achary]

Oh, that's worrying. I'll see if I can reproduce it.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}