Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
* Initial CST push, not close to being ready to run/home/eytan/miniconda3/envs/foil/bin/python * Added tests * Removed plotting from test * Ran black * Parameterized more airfoils in the test * Added more (and fixed) derivative tests * First draft of most of the needed functions in DVGeoCST have been added (except derivative stuff) * Fixed point set so it works with only the points assigned to the current proc * Added separate N1 and N2 values for upper and lower surfaces and associated DVs * Added chord DV * Fixed default DV errors * Ran black * First pass (certainly incorrect so far) at derivatives for CST * Cleaned up a bit and added more (still incorrect) derivatives for use with DVCon...still needs testing and some other TODOs * Supposedly added CST parameter fitting * Initialized CST parameter and chord DVs to the fit values * Added some tests for totalSensitivityProd, still need more for N1 and N2 DVs * Added TODO for better complex handling * Much cleaner handling of compelx numbers for derivatives debugging via isComplex in __init__ * Derivatives tests and some related corrections to DVGeoCST * Fixed bug in trailing edge that messed up derivatives * Ran black * Removed old TODO about derivatives with scaling * Getting there on the fixes, need to figure out how to set the default DVs to the airfoil fit in the pyOpt initialization and fix the bug when there are no DVGeoCST DVs (only alpha) in an opt problem * All good except initializing optimization problem with fitted CST params * Changed the interface yet again to take in a dat file, fixes all existing problems * Added individual xMax, xMin, and thicknessTE to each point set (fixing another bug) * Added dummy comm to total sensitivities * Added allreduce to totalSensitivity function * Added debug mode and plotting * Ran black and removed unused imports * fix prefoil import * Updated prefoil import in DVGeoCST * Updated imports in test for DVGeometryCST * Switched airfoil splitting to use preFoil utilities, added printDesignVariables, added new airfoil for testing * Ran black * Minor version bump to 1.12.0 * Hoping to assuage our flake8 overlords * I failed * Fixed test failures (I hope) * Added prefoil to autodoc_mock_imports * Added ESP and CST to the API docs * Hopefully fixed the math * Added pyOCSM to autodoc_mock_imports * Added DVGeometryVSP to the API docs * Added cstDV * Sorry black and flake8 * Refined docstrings * Added more tests * Fixed some math in docstrings * More tests * Fixed the silver baboon * Credentials no longer required * No more drivatives * Addressed most of Anil's comments * Fixed new flake8 complaint * Moved number of CST coeffs to init from addDV * Added improved trailing edge detection to identify non-vertical trailing edges and sharp and closed cases * Added metric to measure goodness of airfoil CST fit * Undid change to try to fix flake8 * Modified tolTE input so the angle is provided by the user * Moved prints only on root and changed upper/lower bound name in addDV * Perform fit on only one proc and broadcast the result to the rest * Another attempt to fix flake8 * Ran black * Improved debug plot so it gathers data into a single plot on the root * Fixed complex warnings in tests * Ran black Co-authored-by: Alasdair Gray <alachris@umich.edu>
- Loading branch information