Adaptive sampling

[Jeyhun1]

Added adaptive sampling to the framework and applied it in 1D Heat Equation and 1D Burgers Equation examples

[StillerPatrick]

My remarks:

• schrödinger equation is not in geometry fashion
• PDE loss should allocate the next points from the sampler
• adaptive PDE loss is ok
• Geometry should be a module such that I have PINNFramework.Geometry
• Geometry should be changed to an interface such that we can implement multiple ones it
• Geometry should have sampler insight
• Sampler should be a module such that I have PINNFramework.Sampler
• we need some standard samplers such as lhs, random etc.
• sampler should be implemented from an interface that allows implementing own sampling methods in the future.

[StillerPatrick]

@sutariyaraj please review the code as well to check if it fits your current implementation of the inference pipeline. Please communicate prober changes.

[StillerPatrick]

Approved. Nice work