Support LISDF Flattening

[vacancy]

Currently the parser does not "flatten" the SDFInclude/URDFInclude data structures. Although the included files will be parsed and stored in include.content, they won't be merged into the main file. The primary reason for not doing that is because of the pose/scaling settings in <include>. Specifically, when we include a sub-sdf file with a non-identity scaling/translation, when we flatten it, we need to modify the pose and size of all links/joints.

In the future, we should support two arguments: flatten_sdf and flatten_urdf in the parser interface to allow users to decide how they want us to process the files. If flatten_sdf is set to true, we should recursively update the pose/scaling of all links/joints. This can be done either as a post-processing or by adding a transformation stack during parsing.