Working with a MAPDL Full File (full)
=====================================
The MAPDL full file is a FORTRAN formatted binary file containing the
mass and stiffness from an Ansys analysis.  Using pyansys it can be
loaded into memory as either a sparse or full matrix.


Reading a Full File
-------------------
This example reads in the mass and stiffness matrices associated with
the above example.  ``load_km`` sorts degrees of freedom such that the
nodes are ordered from minimum to maximum, and each degree of freedom
(i.e. X, Y, Z), are sorted within each node.  The matrices ``k`` and
``m`` are sparse by default, but if ``scipy`` is not installed, or if
the optional parameter ``as_sparse=False`` then they will be full
numpy arrays.

By default ``load_km`` outputs the upper triangle of both matrices.
The constrained nodes of the analysis can be identified by accessing
``fobj.const`` where the constrained degrees of freedom are True and
all others are False.  This corresponds to the degrees of reference in
``dof_ref``.

By default dof_ref is unsorted.  To sort these values, set
``sort==True``.  It is enabled for this example to allow for plotting
of the values later on.

.. code:: python

    from ansys.mapdl import reader as pymapdl_reader
    from ansys.mapdl.reader import examples
    
    # Create result reader object and read in full file
    full = pymapdl_reader.read_binary(examples.fullfile)
    dof_ref, k, m = full.load_km(sort=True)

ANSYS only stores the upper triangular matrix in the full file.  To
create the full matrix:

.. code:: python

    k += sparse.triu(k, 1).T
    m += sparse.triu(m, 1).T

If you have ``scipy`` installed, you can solve solve for the natural
frequencies and mode shapes of a system.

.. code:: python

    import numpy as np
    from scipy.sparse import linalg

    # condition the k matrix
    # to avoid getting the "Factor is exactly singular" error
    k += sparse.diags(np.random.random(k.shape[0])/1E20, shape=k.shape)

    # Solve
    w, v = linalg.eigsh(k, k=20, M=m, sigma=10000)

    # System natural frequencies
    f = (np.real(w))**0.5/(2*np.pi)    
    
.. code:: 

    print('First four natural frequencies')
    for i in range(4):
        print('{:.3f} Hz'.format(f[i]))

    First four natural frequencies
    1283.200 Hz
    1283.200 Hz

    5781.975 Hz
    6919.399 Hz


Plotting a Mode Shape
---------------------
You can also plot the mode shape of this finite element model.  Since
the constrained degrees of freedom have been removed from the
solution, you have to account for these when displaying the
displacement.

.. code:: python
    
    import pyvista as pv

    # Get the 4th mode shape
    full_mode_shape = v[:, 3] # x, y, z displacement for each node
    
    # reshape and compute the normalized displacement
    disp = full_mode_shape.reshape((-1, 3))
    n = (disp*disp).sum(1)**0.5
    n /= n.max() # normalize
    
    # load an archive file and create a vtk unstructured grid
    archive = pymapdl_reader.Archive(examples.hexarchivefile)
    grid = archive.parse_vtk()
    
    # plot the normalized displacement
    # grid.plot(scalars=n)
    
    # Fancy plot the displacement
    pl = pv.Plotter()
    
    # add the nominal mesh
    pl.add_mesh(grid, style='wireframe')
	  
    # copy the mesh and displace it
    new_grid = grid.copy()
    new_grid.points += disp/80
    pl.add_mesh(new_grid, scalars=n, stitle='Normalized\nDisplacement',
                flipscalars=True)
    
    pl.add_text('Cantliver Beam 4th Mode Shape at {:.4f}'.format(f[3]),
                fontsize=30)
    pl.plot()
    
.. image:: ../images/solved_km.png


This example is built into ``pyansys-mapdl`` and can be run from
``examples.solve_km()``.


FullFile Object Methods
-----------------------
.. autoclass:: ansys.mapdl.reader.full.FullFile
    :members: