pyyeti.nastran.op4.load

pyyeti.nastran.op4.load(filename=None, namelist=None, into='dct', justmatrix=False, sparse=False)[source]

Read all matching matrices from op4 file into dictionary or list; non-member version of OP4.load().

This is a the same as read() except justmatrix default is False.

Parameters:

  • filename (string or None; optional) – Name of op4 file to read. Can also be the name of a directory or None; in these cases, a GUI is opened for file selection.

  • namelist (list, string, or None; optional) – List of variable names to read in, or string with name of the single variable to read in, or None. If None, all matrices are read in.

  • into (string; optional) – Either ‘dct’ or ‘list’. Use ‘list’ if multiple matrices share the same name. See below.

  • justmatrix (bool; optional) – If True, only the matrix is stored in the dictionary. If False, a tuple of (matrix, form, mtype) is stored. This option is ignored if into == 'list'.

  • sparse (bool or None or two-tuple_like; optional) – Specifies whether output matrices will be regular numpy arrays or sparse arrays. If not two-tuple_like:

    sparse

    Action

    None

    Auto setting: each matrix will be sparse if and only if it was written in a sparse format

    True

    Matrices will be returned in sparse format

    False

    Matrices will be returned in regular (dense) numpy arrays

    If sparse is two-tuple_like, the first element is either None, True, or False (see table above) and the second element is a callable, as in: X = callable(X). A common usage of the callable would be to convert from “COO” sparse form (see scipy.sparse.coo_matrix) to a more desirable form. For example, to ensure all matrices are returned in CSC form (see scipy.sparse.csc_matrix) use:

    sparse=(True, scipy.sparse.coo_matrix.tocsc)

    The callable is ignored for non-sparse matrices.

Returns:

  • dct (collections.OrderedDict, if into == 'dct') – Keys are the lower-case matrix names and the values are either just the matrix or a tuple of: (matrix, form, mtype) depending on justmatrix.

  • tup (tuple, if into == 'list') – Tuple of 4 lists: (names, matrices, forms, mtypes)

Notes

The default form for sparse matrices is the “COO” sparse form (see scipy.sparse.coo_matrix). To override, provide a callable in the sparse option (see above).

Examples

This examples translates a sparse format binary op4 file to a simpler ascii format while preserving the matrix forms.

First, create a file in “bigmat” sparse format and set the form on the “m” matrix to be symmetric (form=6):

>>> import numpy as np
>>> from pyyeti.nastran import op4
>>> m = np.array([[1, 2], [2.1, 3]])
>>> k = np.array([3, 5])
>>> b = np.array([4, 6])
>>> names = ['m', 'k', 'b']
>>> values = [eval(n) for n in names]
>>> op4.write('mkb.op4', names, values, forms=[6, 2, 2],
...           sparse='bigmat')

Now, translate it to simple ascii, preserving the forms:

>>> dct = op4.load('mkb.op4')
>>> op4.write('mkb_ascii.op4', dct, binary=False)

Check that the order and forms are the same:

>>> _ = op4.dir('mkb.op4')
m       ,      2 x 2     , form=6, mtype=2
k       ,      1 x 2     , form=2, mtype=2
b       ,      1 x 2     , form=2, mtype=2

>>> _ = op4.dir('mkb_ascii.op4')
m       ,      2 x 2     , form=6, mtype=2
k       ,      1 x 2     , form=2, mtype=2
b       ,      1 x 2     , form=2, mtype=2

Clean up:

>>> import os
>>> os.remove('mkb.op4')
>>> os.remove('mkb_ascii.op4')

See also

read(), write(), dir()