mdolab/dafoam: v3.0.6

Changes Added a new feature to call Python functions in the C++ layer. 1f680066294c46abde303659ea0a7ba39b570f37 Use TensorFlow Python APIs for predicting and calculating gradients. The old TensorFlow C APIs are deprecated. 1f680066294c46abde303659ea0a7ba39b570f37 Fixed a bug for multi-point setup caused by a recent update. fd32b8106687b218db94a3273592d3a710105cf6 Updated the CoDiPack and MediPack versions and used the latest APIs for AD tapes 6e9cd144f66305a25862dd35d9f99d13cab91dec Added an option to use uniform fvSource as a design variables to replace the uniformPressureGradient option 106f05466d6d4cac303e99955494cf2ccb3b1855 Allowed using non-geo design variable setup for topology optimization or field inversion. Now we don't need to set up FFD or meshOptions for these cases. fb38ae153dd46e01b4fcc9578a22a3115df6e3b6 Check the new tutorial run script from here Showed errors in the GitHub Actions tests by removing the 2> errorLog.txt compilation argument in the test scripts. 3c477e2e4401403f014086fbaa1547c9f1941f8d Dependency update OpenFOAM-AD upgraded to v1.3.0 with the latest CoDiPack and MeDiPack versions. Deprecated the TensorFlow C APIs and used the Python APIs instead. Backward compatibility Non