Replace boost hash usage with TfHash in pxr/base/gf

[nvmkuruc]

Description of Change(s)

This is a reconstruction #2175 which didn't get successfully merged.

Requires: #2173
To remove the dependency of pxr/base/gf on boost's hashing functions

• Copy constructor for python bindings added when missing to support test coverage (Gf.Rotation, Gf.Quaternion, Gf.Interval)
• __hash__ operator overloads are added when missing to support test coverage (Gf.BBox3d, Gf.Frustum, Gf.Quat{f,h,d}, Gf.Rect2i, Gf.Rotation)
• __hash__ test coverage added to testGfRotation, testGfVec, testGfRect2i, testGfRange, testGfQuarternion, testGfMatrix, testGfInterval, testGfFrustum, testGfDualQuaternion, testGfBBox3d, and testGfFrustum.
• Replaced boost::hash_combine with TfHash::Combine in hash_value implementations. Preserving hash_value allows boost::hash to still work with these types without the explicit dependency. TfHash will use these hash_value implementations when no TfHashAppend overload has been provided.

Fixes Issue(s)

-#2172 (more PRs forthcoming)

• I have verified that all unit tests pass with the proposed changes

• I have submitted a signed Contributor License Agreement

[sunyab]

Filed as internal issue #USD-8026