quippy.farray.FortranArray – Fortran one-based array indexing

quippy contains a FortranArray class and various utility functions for handling one-based array indexing.

class quippy.farray.FortranArray

Subclass of numpy.ndarray which uses Fortran-style one-based indexing.

The first element is numbered one rather than zero and trying to access element zero will raise an IndexError exception. Negative indices are unchanged; -1 still refers to the highest index along a dimension. Slices are also Fortran-style, i.e. inclusive on the second element so 1:2 includes the one-based elements 1 and 2, equivalent to a C-style slice of 0:2. The self.flat iterator is still indexed from zero.

all(axis=None, out=None)

One-based analogue of numpy.ndarray.all()

any(axis=None, out=None)

One-based analogue of numpy.ndarray.any()

argmax(axis=None, out=None)

Return one-based indices of the maximum values along the given axis of a.

Refer to numpy.ndarray.argmax for detailed documentation.

argmin(axis=None, out=None)

Return one-based indices of the minimum values along the given axis of a.

Refer to numpy.ndarray.argmax for detailed documentation.

argsort(axis=None, kind='quicksort', order=None)

Returns the indices that would sort this array.

Refer to numpy.argsort for full documentation.

col_iter()

Iterator for MxN arrays to return cols [...,i] for i=1,N one by one as Mx1 arrays.

cols

Iterator for MxN arrays to return cols [...,i] for i=1,N one by one as Mx1 arrays.

mapindices(indx)

Transform from Fortran-style one-based to C-style zero based indices.

indx can be any object that can be used to subscript an ndarray - scalar integer or slice, sequence of integers and slices or an ndarray of integer or boolean kind.

mean(axis=None, dtype=None, out=None)

One-based analogue of numpy.ndarray.mean()

nonzero()

Return the one-based indices of the elements of a which are not zero.

norm()

Return sqrt(norm2(self))

norm2()

Squared norm of a 1D or 2D array.

For a 1D array, returns dot(self,self) For a 2D array, must have shape (3,n). Returns array a where a[i] = dot(self[:,i],self[:,i])

put(indices, values, mode='raise')

Set a.flat[n] = values[n] for all n in indices.

Refer to numpy.put for full documentation.

row_iter()

Iterate over this FortranArray treating first dimension as fastest varying

rows

Iterate over this FortranArray treating first dimension as fastest varying

stripstrings()

Return contents as string (0- and 1-dimensional arrays) or array of strings with trailing spaces removed (2-dimensional arrays).

Raises ValueError if this FortranArray does not have a string datatype.

sum(axis=None, dtype=None, out=None)

One-based analogue of numpy.ndarray.sum()

take(indices, axis=None, out=None, mode='raise')

Return an array formed from the elements of a at the given one-based indices.

Refer to numpy.take for full documentation.

There are also some functions which operate on FortranArray instances:

quippy.farray.farray(seq, dtype=None)

Convert seq to a FortranArray with Fortran ordering.

>>> fa = farray([1,2,3])

A copy of the data in seq will be made if necessary.

quippy.farray.fenumerate(seq)

One-based equivalent of enumerate

quippy.farray.fidentity(n)

Return the n dimensional identity matrix as a FortranArray.

quippy.farray.frange(min, max=None, step=1)

Fortran equivalent of range() builtin.

Returns an iterator for integers from min to max inclusive, increasing by step each time.

>>> list(frange(3))
[1, 2, 3]
>>> list(frange(3,6,2))
[3, 5]

quippy.farray.fvar(seq)

Create one or more rank-0 instances of FortranArray and inject them into the global namespace.

This is a convenience function useful for making arrays to use as intent(in,out) arguments to a Fortran function. A single string argument causes variables with one-letter names to be created. The new arrays are also returned.

The following examples are equivalent:

>>> fvar("abc")
(FortranArray(0.0), FortranArray(0.0), FortranArray(0.0))
>>> a, b, c = fvar(['a','b','c'])

quippy.farray.fzeros(shape, dtype=<type 'float'>)

Create an empty FortranArray with Fortran ordering.

quippy.farray.padded_str_array(d, length)

Return FortranArray with shape (length, len(d)), filled with rows from d padded with spaces

quippy.farray.convert_farray_to_ndarray(func)

Decorator to convert all occurences of farray in arguments to ndarray. If result contains ndarray, they will be converted back to farray.

quippy.farray.convert_ndarray_to_farray(func)

Decorator to convert all occurences of ndarray in arguments to farray. If results contains farray, they will be converted back to ndarray.

quippy.farray.n2f(x)

Return x.view(FortranArray)

quippy.farray.f2n(x)

Return x.view(numpy.ndarray)