.. _section_szybki_releasenotes_171:

Szybki TK 1.7.1
================================================================================

New features
--------------------------------------------------------------------------------
*  :oe:meth:`OESz::OESzybki::GetEntropy` added which takes a non-const multiconformer
   molecule and a reference to a new object :ref:`OESzybkiEnsembleResults <OESz_OESzybkiEnsembleResults>`.
   The method evaluates ligand entropy in different environments as other existing
   methods, however in addition provides essential data on the ensemble of ligand conformations.
   Specifically, it returns the optimized ensemble of unique ligand conformations in its first
   parameter, and populates the :ref:`OESzybkiEnsembleResults <OESz_OESzybkiEnsembleResults>` objects with entropy
   components data for the ensemble. The :ref:`OESzybkiEnsembleResults <OESz_OESzybkiEnsembleResults>` also provides
   the data for individual unique optimized conformations in the ensemble; in particular
   the free energy of isolating a specific conformation from the entire ensemble. An example
   code explaining the usage of the new :oe:meth:`OESz::OESzybki::GetEntropy` method and the
   :ref:`OESzybkiEnsembleResults <OESz_OESzybkiEnsembleResults>` class has been added.

*  :ref:`OESzybkiEnsembleResults <OESz_OESzybkiEnsembleResults>` class added that is mentioned above has been added.

*  Two new API methods have been added which control the temperature in the entropy estimation.
   These are: `SetTemperature` and `GetTemperature`. 
   The default temperature is 298K.

Bug fixes
-----------------------------------------------------------------------------------------------------
*  :oe:meth:`OESz::OESzybki::SetHarmonicConstraints` did not impose harmonic constraints on 
   flexible part of the protein. The problem has been fixed.

*  Trying to reuse an :ref:`OESzybki <OESz_OESzybki>` instance after an entropy calculation sometimes
   (depending on the subsequent methods call order) caused a crash. To work around that problem 
   it was necessary to use a new instance of the :ref:`OESzybki <OESz_OESzybki>` class.
   This bug has been fixed.

*  Method :oe:meth:`OESz::OESzybki::ClearFixAtoms` did not clear the unnecessary evaluation of the complete 
   ligand energy at the very end of optimization with a subset of atoms frozen. The problem has been
   fixed. 

*  Protection against exhausting memory upon the usage of too many Coulomb atom pairs (for example
   in the attempt to optimize a large protein) has been added. Also, a confusing message upon 
   exhausting the memory with too many vdW interactions has been removed.

*  A warning is now issued in the case where one of the three methods setting the protein flexibility
   is used (`SetSideChainsFlexibility`, `SetPolarHFlexibility` or `SetResidueFlexibility`)
   when protein is not set. Also, a request of single point calculation (using 
   `SetRunType`) properly clears any previous settings of partial 
   protein flexibility.

*  In the default optimization of a ligand inside the protein receptor the convergence criteria
   set by the method `SetOptimizerType` were not consistent with the 
   intrinsic criteria of one the optimizers used. This inconsistency has been fixed.

*  Using conjugate gradient optimizer (set with the `SetOptimizerType`) in
   combination with the method `SetExactVdWProteinLigand` did not report
   the number of cycles made by the optimizer. This problem has been fixed.

