forte2.helpers.diis
===================

.. py:module:: forte2.helpers.diis




Module Contents
---------------

.. py:class:: DIIS

   
   A class that implements the direct inversion in the iterative subspace (DIIS) method.


   :Parameters:

       **diis_start** : int, optional, default=3
           Start saving DIIS vectors after this many iterations.
           A value less than 0 means no DIIS

       **diis_nvec** : int, optional, default=8
           The number of vectors to keep in the DIIS.

       **diis_min** : int, optional, default=3
           The minimum number of vectors to perform extrapolation.

       **do_diis** : bool, optional
           If True, DIIS is performed. If False, DIIS is not performed.
           If None, DIIS is performed unless `diis_start` or `diis_nvec` is less than 1.

   :Attributes:

       **status** : str
           Status of the last DIIS update. It could be one of the following:
           - "": No DIIS update was performed.
           - "S": DIIS vectors are saved.
           - "S/E": DIIS vectors are saved and extrapolation is performed.










   .. rubric:: Notes

   The currently implemented DIIS method is Pulay's original method (https://doi.org/10.1016/0009-2614(80)80396-4), which is also known as CDIIS.
   The error vector is the AO-basis (we use the orthonormal basis in actual implementation) orbital gradient: FDS-SDF (https://doi.org/10.1080/00268976900100941).



   ..
       !! processed by numpydoc !!

   .. py:attribute:: diis_start
      :type:  int
      :value: 4



   .. py:attribute:: diis_nvec
      :type:  int
      :value: 8



   .. py:attribute:: diis_min
      :type:  int
      :value: 3



   .. py:attribute:: do_diis
      :type:  bool
      :value: None



   .. py:method:: update(p: numpy.typing.NDArray, e: numpy.typing.NDArray) -> numpy.typing.NDArray

      
      Update the DIIS object and return extrapolated parameters


      :Parameters:

          **p** : NDArray
              The current set of parameters

          **e** : NDArray
              The current error vector



      :Returns:

          NDArray
              The extrapolated parameters











      ..
          !! processed by numpydoc !!