fipy.tools¶

Utility modules, functions, and values

fipy.tools.serialComm: CommWrapper¶: Serial MPI communicator when running in parallel.

fipy.tools.parallelComm: CommWrapper¶: Parallel MPI communicator when running in parallel.

class fipy.tools.PhysicalField(value, unit=None, array=None)¶

Bases: object

Field or quantity with units.

Physical Fields can be constructed in one of two ways:

PhysicalField(*value*, *unit*), where *value* is a number of arbitrary type and *unit* is a string containing the unit name
>>> print(PhysicalField(value = 10., unit = 'm'))
10.0 m
PhysicalField(*string*), where *string* contains both the value and the unit. This form is provided to make interactive use more convenient
>>> print(PhysicalField(value = "10. m"))
10.0 m

Dimensionless quantities, with a unit of 1, can be specified in several ways

>>> print(PhysicalField(value = "1"))
1.0 1
>>> print(PhysicalField(value = 2., unit = " "))
2.0 1
>>> print(PhysicalField(value = 2.))
2.0 1

Physical arrays are also possible (and are the reason this code was adapted from Konrad Hinsen’s original PhysicalQuantity). The value can be a Numeric array:

>>> a = numerix.array(((3., 4.), (5., 6.)))
>>> print(PhysicalField(value = a, unit = "m"))
[[ 3.  4.]
 [ 5.  6.]] m

or a tuple:

>>> print(PhysicalField(value = ((3., 4.), (5., 6.)), unit = "m"))
[[ 3.  4.]
 [ 5.  6.]] m

or as a single value to be applied to every element of a supplied array:

>>> print(PhysicalField(value = 2., unit = "m", array = a))
[[ 2.  2.]
 [ 2.  2.]] m

Every element in an array has the same unit, which is stored only once for the whole array.

__abs__()¶

Return the absolute value of the quantity. The unit is unchanged.

>>> print(abs(PhysicalField(((3., -2.), (-1., 4.)), 'm')))
[[ 3.  2.]
 [ 1.  4.]] m

__add__(other)¶

Add two physical quantities, so long as their units are compatible. The unit of the result is the unit of the first operand.

>>> print(PhysicalField(10., 'km') + PhysicalField(10., 'm'))
10.01 km
>>> print(PhysicalField(10., 'km') + PhysicalField(10., 'J'))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

__array__(dtype=None, copy=None)¶

Return a dimensionless PhysicalField as a Numeric array.

>>> print(numerix.array(PhysicalField(((2., 3.), (4., 5.)), "m/m")))
[[ 2.  3.]
 [ 4.  5.]]

As a special case, fields with angular units are converted to base units (radians) and then assumed dimensionless.

>>> print(numerix.array(PhysicalField(((2., 3.), (4., 5.)), "deg")))
[[ 0.03490659  0.05235988]
 [ 0.06981317  0.08726646]]

If the array is not dimensionless, the numerical value in its base units is returned.

>>> numerix.array(PhysicalField(((2., 3.), (4., 5.)), "mm"))
array([[ 0.002,  0.003],
       [ 0.004,  0.005]])

__array_wrap__(arr, context=None, return_scalar=False)¶

Required to prevent numpy not calling the reverse binary operations. Both the following tests are examples ufuncs.

>>> from fipy.tools.dimensions.physicalField import PhysicalField
>>> print(type(numerix.array([1.0, 2.0]) * PhysicalField([1.0, 2.0], unit="m")))
<class 'fipy.tools.dimensions.physicalField.PhysicalField'>

>>> print(type(numerix.array([1.0, 2.0]) * PhysicalField([1.0, 2.0])))
<class 'fipy.tools.dimensions.physicalField.PhysicalField'>

>>> from scipy.special import gamma as Gamma 
>>> print(isinstance(Gamma(PhysicalField([1.0, 2.0])), type(numerix.array(1)))) 
1

__bool__()¶

Test if the quantity is zero.

Should this only pass if the unit offset is zero?

__div__(other)¶

Divide two physical quantities. The unit of the result is the unit of the first operand divided by the unit of the second.

>>> print(PhysicalField(10., 'm') / PhysicalField(2., 's'))
5.0 m/s

As a special case, if the result is dimensionless, the value is returned without units, rather than with a dimensionless unit of 1. This facilitates passing physical quantities to packages such as Numeric that cannot use units, while ensuring the quantities have the desired units

>>> print((PhysicalField(1., 'inch')
...        / PhysicalField(1., 'mm')))
25.4

__eq__(other)¶: Return self==value.

__float__()¶

Return a dimensionless PhysicalField quantity as a float.

>>> float(PhysicalField("2. m/m"))
2.0

As a special case, quantities with angular units are converted to base units (radians) and then assumed dimensionless.

>>> print(numerix.round(float(PhysicalField("2. deg")), 6))
0.034907

If the quantity is not dimensionless, the conversion fails.

>>> float(PhysicalField("2. m"))
Traceback (most recent call last):
    ...
TypeError: Not possible to convert a PhysicalField with dimensions to float

Just as a Numeric array cannot be cast to float, neither can PhysicalField arrays

>>> float(PhysicalField(((2., 3.), (4., 5.)), "m/m")) 
Traceback (most recent call last):
    ...
TypeError: only ...-1 arrays can be converted to Python scalars

__ge__(other)¶: Return self>=value.

__getitem__(index)¶

Return the specified element of the array. The unit of the result will be the unit of the array.

>>> a = PhysicalField(((3., 4.), (5., 6.)), "m")
>>> print(a[1, 1])
6.0 m

__gt__(other)¶

Compare self to other, returning an array of Boolean values corresponding to the test against each element.

>>> a = PhysicalField(((3., 4.), (5., 6.)), "m")
>>> print(numerix.allclose(a > PhysicalField("13 ft"),
...                        [[False, True], [ True, True]]))
True

Appropriately formatted dimensional quantity strings can also be compared.

>>> print(numerix.allclose(a > "13 ft",
...                        [[False, True], [ True, True]]))
True

Arrays are compared element to element

>>> print(numerix.allclose(a > PhysicalField(((3., 13.), (17., 6.)), "ft"),
...                        [[ True, True], [False, True]]))
True

Units must be compatible

>>> print(a > PhysicalField("1 lb"))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

And so must array dimensions

>>> print(a > PhysicalField(((3., 13., 4.), (17., 6., 2.)), "ft")) 
Traceback (most recent call last):
    ...
ValueError: shape mismatch: objects cannot be broadcast to a single shape

__hash__()¶: Return hash(self).

__le__(other)¶: Return self<=value.

__lt__(other)¶: Return self<value.

__mod__(other)¶

Return the remainder of dividing two physical quantities. The unit of the result is the unit of the first operand divided by the unit of the second.

>>> print(PhysicalField(11., 'm') % PhysicalField(2., 's'))
1.0 m/s

__mul__(other)¶

Multiply two physical quantities. The unit of the result is the product of the units of the operands.

>>> print(PhysicalField(10., 'N') * PhysicalField(10., 'm') == PhysicalField(100., 'N*m'))
True

As a special case, if the result is dimensionless, the value is returned without units, rather than with a dimensionless unit of 1. This facilitates passing physical quantities to packages such as Numeric that cannot use units, while ensuring the quantities have the desired units.

>>> print((PhysicalField(10., 's') * PhysicalField(2., 'Hz')))
20.0

__ne__(other)¶: Return self!=value.

__neg__()¶

Return the negative of the quantity. The unit is unchanged.

>>> print(-PhysicalField(((3., -2.), (-1., 4.)), 'm'))
[[-3.  2.]
 [ 1. -4.]] m

__nonzero__()¶

Test if the quantity is zero.

Should this only pass if the unit offset is zero?

__pow__(other)¶

Raise a PhysicalField to a power. The unit is raised to the same power.

>>> print(PhysicalField(10., 'm')**2)
100.0 m**2

__radd__(other)¶

Add two physical quantities, so long as their units are compatible. The unit of the result is the unit of the first operand.

>>> print(PhysicalField(10., 'km') + PhysicalField(10., 'm'))
10.01 km
>>> print(PhysicalField(10., 'km') + PhysicalField(10., 'J'))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

__repr__()¶

Return representation of a physical quantity suitable for re-use

>>> PhysicalField(value = 3., unit = "eV")
PhysicalField(3.0,'eV')

__rmul__(other)¶

Multiply two physical quantities. The unit of the result is the product of the units of the operands.

>>> print(PhysicalField(10., 'N') * PhysicalField(10., 'm') == PhysicalField(100., 'N*m'))
True

As a special case, if the result is dimensionless, the value is returned without units, rather than with a dimensionless unit of 1. This facilitates passing physical quantities to packages such as Numeric that cannot use units, while ensuring the quantities have the desired units.

>>> print((PhysicalField(10., 's') * PhysicalField(2., 'Hz')))
20.0

__setitem__(index, value)¶

Assign the specified element of the array, performing appropriate conversions.

>>> a = PhysicalField(((3., 4.), (5., 6.)), "m")
>>> a[0, 1] = PhysicalField("6 ft")
>>> print(a)
[[ 3.      1.8288]
 [ 5.      6.    ]] m
>>> a[1, 0] = PhysicalField("2 min")
Traceback (most recent call last):
    ...
TypeError: Incompatible units

__str__()¶

Return human-readable form of a physical quantity

>>> print(PhysicalField(value = 3., unit = "eV"))
3.0 eV

__sub__(other)¶

Subtract two physical quantities, so long as their units are compatible. The unit of the result is the unit of the first operand.

>>> print(PhysicalField(10., 'km') - PhysicalField(10., 'm'))
9.99 km
>>> print(PhysicalField(10., 'km') - PhysicalField(10., 'J'))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

__truediv__(other)¶

Divide two physical quantities. The unit of the result is the unit of the first operand divided by the unit of the second.

>>> print(PhysicalField(10., 'm') / PhysicalField(2., 's'))
5.0 m/s

As a special case, if the result is dimensionless, the value is returned without units, rather than with a dimensionless unit of 1. This facilitates passing physical quantities to packages such as Numeric that cannot use units, while ensuring the quantities have the desired units

>>> print((PhysicalField(1., 'inch')
...        / PhysicalField(1., 'mm')))
25.4

add(other)¶

Add two physical quantities, so long as their units are compatible. The unit of the result is the unit of the first operand.

>>> print(PhysicalField(10., 'km') + PhysicalField(10., 'm'))
10.01 km
>>> print(PhysicalField(10., 'km') + PhysicalField(10., 'J'))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

allclose(other, atol=None, rtol=1e-08)¶

This function tests whether or not self and other are equal subject to the given relative and absolute tolerances. The formula used is:

| self - other | < atol + rtol * | other |

This means essentially that both elements are small compared to atol or their difference divided by other’s value is small compared to rtol.

allequal(other)¶: This function tests whether or not self and other are exactly equal.

arccos()¶

Return the inverse cosine of the PhysicalField in radians

>>> print(PhysicalField(0).arccos().allclose("1.57079632679 rad"))
1

The input PhysicalField must be dimensionless

>>> print(numerix.round(PhysicalField("1 m").arccos(), 6))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

arccosh()¶

Return the inverse hyperbolic cosine of the PhysicalField

>>> print(numerix.allclose(PhysicalField(2).arccosh(),
...                        1.31695789692))
1

The input PhysicalField must be dimensionless

>>> print(numerix.round(PhysicalField("1. m").arccosh(), 6))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

arcsin()¶

Return the inverse sine of the PhysicalField in radians

>>> print(PhysicalField(1).arcsin().allclose("1.57079632679 rad"))
1

The input PhysicalField must be dimensionless

>>> print(numerix.round(PhysicalField("1 m").arcsin(), 6))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

arctan()¶

Return the arctangent of the PhysicalField in radians

>>> print(numerix.round(PhysicalField(1).arctan(), 6))
0.785398

The input PhysicalField must be dimensionless

>>> print(numerix.round(PhysicalField("1 m").arctan(), 6))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

arctan2(other)¶

Return the arctangent of self divided by other in radians

>>> print(numerix.round(PhysicalField(2.).arctan2(PhysicalField(5.)), 6))
0.380506

The input PhysicalField objects must be in the same dimensions

>>> print(numerix.round(PhysicalField(2.54, "cm").arctan2(PhysicalField(1., "inch")), 6))
0.785398

>>> print(numerix.round(PhysicalField(2.).arctan2(PhysicalField("5. m")), 6))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

arctanh()¶

Return the inverse hyperbolic tangent of the PhysicalField

>>> print(PhysicalField(0.5).arctanh())
0.549306144334

The input PhysicalField must be dimensionless

>>> print(numerix.round(PhysicalField("1 m").arctanh(), 6))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

ceil()¶

Return the smallest integer greater than or equal to the PhysicalField.

>>> print(PhysicalField(2.2, "m").ceil())
3.0 m

conjugate()¶

Return the complex conjugate of the PhysicalField.

>>> print(PhysicalField(2.2 - 3j, "ohm").conjugate() == PhysicalField(2.2 + 3j, "ohm"))
True

convertToUnit(unit)¶

Changes the unit to unit and adjusts the value such that the combination is equivalent. The new unit is by a string containing its name. The new unit must be compatible with the previous unit of the object.

>>> e = PhysicalField('2.7 Hartree*Nav')
>>> e.convertToUnit('kcal/mol')
>>> print(e)
1694.27557621 kcal/mol

copy()¶

Make a duplicate.

>>> a = PhysicalField(1, unit = 'inch')
>>> b = a.copy()

The duplicate will not reflect changes made to the original

>>> a.convertToUnit('cm')
>>> print(a)
2.54 cm
>>> print(b)
1 inch

Likewise for arrays

>>> a = PhysicalField(numerix.array((0, 1, 2)), unit  = 'm')
>>> b = a.copy()
>>> a[0] = 3
>>> print(a)
[3 1 2] m
>>> print(b)
[0 1 2] m

cos()¶

Return the cosine of the PhysicalField

>>> print(numerix.round(PhysicalField(2*numerix.pi/6, "rad").cos(), 6))
0.5
>>> print(numerix.round(PhysicalField(60., "deg").cos(), 6))
0.5

The units of the PhysicalField must be an angle

>>> PhysicalField(60., "m").cos()
Traceback (most recent call last):
    ...
TypeError: Incompatible units

cosh()¶

Return the hyperbolic cosine of the PhysicalField

>>> PhysicalField(0.).cosh()
1.0

The units of the PhysicalField must be dimensionless

>>> PhysicalField(60., "m").cosh()
Traceback (most recent call last):
    ...
TypeError: Incompatible units

divide(other)¶

Divide two physical quantities. The unit of the result is the unit of the first operand divided by the unit of the second.

>>> print(PhysicalField(10., 'm') / PhysicalField(2., 's'))
5.0 m/s

As a special case, if the result is dimensionless, the value is returned without units, rather than with a dimensionless unit of 1. This facilitates passing physical quantities to packages such as Numeric that cannot use units, while ensuring the quantities have the desired units

>>> print((PhysicalField(1., 'inch')
...        / PhysicalField(1., 'mm')))
25.4

dot(other)¶

Return the dot product of self with other. The resulting unit is the product of the units of self and other.

>>> v = PhysicalField(((5., 6.), (7., 8.)), "m")
>>> print(PhysicalField(((1., 2.), (3., 4.)), "m").dot(v))
[ 26.  44.] m**2

property dtype¶

Returns the NumPy sctype of the underlying array.

>>> issubclass(PhysicalField(1, 'm').dtype.type, numerix.integer)
True
>>> issubclass(PhysicalField(1., 'm').dtype.type, numerix.floating)
True
>>> issubclass(PhysicalField((1, 1.), 'm').dtype.type, numerix.floating)
True

floor()¶

Return the largest integer less than or equal to the PhysicalField.

>>> print(PhysicalField(2.2, "m").floor())
2.0 m

inBaseUnits()¶

Return the quantity with all units reduced to their base SI elements.

>>> e = PhysicalField('2.7 Hartree*Nav')
>>> print(e.inBaseUnits().allclose("7088849.01085 kg*m**2/s**2/mol"))
1

inDimensionless()¶

Returns the numerical value of a dimensionless quantity.

>>> print(PhysicalField(((2., 3.), (4., 5.))).inDimensionless())
[[ 2.  3.]
 [ 4.  5.]]

It’s an error to convert a quantity with units

>>> print(PhysicalField(((2., 3.), (4., 5.)), "m").inDimensionless())
Traceback (most recent call last):
    ...
TypeError: Incompatible units

inRadians()¶

Converts an angular quantity to radians and returns the numerical value.

>>> print(PhysicalField(((2., 3.), (4., 5.)), "rad").inRadians())
[[ 2.  3.]
 [ 4.  5.]]
>>> print(PhysicalField(((2., 3.), (4., 5.)), "deg").inRadians())
[[ 0.03490659  0.05235988]
 [ 0.06981317  0.08726646]]

As a special case, assumes a dimensionless quantity is already in radians.

>>> print(PhysicalField(((2., 3.), (4., 5.))).inRadians())
[[ 2.  3.]
 [ 4.  5.]]

It’s an error to convert a quantity with non-angular units

>>> print(PhysicalField(((2., 3.), (4., 5.)), "m").inRadians())
Traceback (most recent call last):
    ...
TypeError: Incompatible units

inSIUnits()¶

Return the quantity with all units reduced to SI-compatible elements.

>>> e = PhysicalField('2.7 Hartree*Nav')
>>> print(e.inSIUnits().allclose("7088849.01085 kg*m**2/s**2/mol"))
1

inUnitsOf(*units)¶

Returns one or more PhysicalField objects that express the same physical quantity in different units. The units are specified by strings containing their names. The units must be compatible with the unit of the object. If one unit is specified, the return value is a single PhysicalField.

>>> freeze = PhysicalField('0 degC')
>>> print(freeze.inUnitsOf('degF').allclose("32.0 degF"))
1

If several units are specified, the return value is a tuple of PhysicalField instances with with one element per unit such that the sum of all quantities in the tuple equals the the original quantity and all the values except for the last one are integers. This is used to convert to irregular unit systems like hour/minute/second. The original object will not be changed.

>>> t = PhysicalField(314159., 's')
>>> from builtins import zip
>>> print(numerix.allclose([e.allclose(v) for (e, v) in zip(t.inUnitsOf('d', 'h', 'min', 's'),
...                                                         ['3.0 d', '15.0 h', '15.0 min', '59.0 s'])],
...                        True))
1

log()¶

Return the natural logarithm of the PhysicalField

>>> print(numerix.round(PhysicalField(10).log(), 6))
2.302585

The input PhysicalField must be dimensionless

>>> print(numerix.round(PhysicalField("1. m").log(), 6))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

log10()¶

Return the base-10 logarithm of the PhysicalField

>>> print(numerix.round(PhysicalField(10.).log10(), 6))
1.0

The input PhysicalField must be dimensionless

>>> print(numerix.round(PhysicalField("1. m").log10(), 6))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

multiply(other)¶

Multiply two physical quantities. The unit of the result is the product of the units of the operands.

>>> print(PhysicalField(10., 'N') * PhysicalField(10., 'm') == PhysicalField(100., 'N*m'))
True

As a special case, if the result is dimensionless, the value is returned without units, rather than with a dimensionless unit of 1. This facilitates passing physical quantities to packages such as Numeric that cannot use units, while ensuring the quantities have the desired units.

>>> print((PhysicalField(10., 's') * PhysicalField(2., 'Hz')))
20.0

property numericValue¶

Return the PhysicalField without units, after conversion to base SI units.

>>> print(numerix.round(PhysicalField("1 inch").numericValue, 6))
0.0254

put(indices, values)¶

put is the opposite of take. The values of self at the locations specified in indices are set to the corresponding value of values.

The indices can be any integer sequence object with values suitable for indexing into the flat form of self. The values must be any sequence of values that can be converted to the typecode of self.

>>> f = PhysicalField((1., 2., 3.), "m")
>>> f.put((2, 0), PhysicalField((2., 3.), "inch"))
>>> print(f)
[ 0.0762  2.      0.0508] m

The units of values must be compatible with self.

>>> f.put(1, PhysicalField(3, "kg"))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

reshape(shape)¶

Changes the shape of self to that specified in shape

>>> print(PhysicalField((1., 2., 3., 4.), "m").reshape((2, 2)))
[[ 1.  2.]
 [ 3.  4.]] m

The new shape must have the same size as the existing one.

>>> print(PhysicalField((1., 2., 3., 4.), "m").reshape((2, 3))) 
Traceback (most recent call last):
    ...
ValueError: total size of new array must be unchanged

property shape¶: Tuple of array dimensions.

sign()¶

Return the sign of the quantity. The unit is unchanged.

>>> from fipy.tools.numerix import sign
>>> print(sign(PhysicalField(((3., -2.), (-1., 4.)), 'm')))
[[ 1. -1.]
 [-1.  1.]]

sin()¶

Return the sine of the PhysicalField

>>> print(PhysicalField(numerix.pi/6, "rad").sin())
0.5
>>> print(PhysicalField(30., "deg").sin())
0.5

The units of the PhysicalField must be an angle

>>> PhysicalField(30., "m").sin()
Traceback (most recent call last):
    ...
TypeError: Incompatible units

sinh()¶

Return the hyperbolic sine of the PhysicalField

>>> PhysicalField(0.).sinh()
0.0

The units of the PhysicalField must be dimensionless

>>> PhysicalField(60., "m").sinh()
Traceback (most recent call last):
    ...
TypeError: Incompatible units

sqrt()¶

Return the square root of the PhysicalField

>>> print(PhysicalField("100. m**2").sqrt())
10.0 m

The resulting unit must be integral

>>> print(PhysicalField("100. m").sqrt())
Traceback (most recent call last):
    ...
TypeError: Illegal exponent

subtract(other)¶

Subtract two physical quantities, so long as their units are compatible. The unit of the result is the unit of the first operand.

>>> print(PhysicalField(10., 'km') - PhysicalField(10., 'm'))
9.99 km
>>> print(PhysicalField(10., 'km') - PhysicalField(10., 'J'))
Traceback (most recent call last):
    ...
TypeError: Incompatible units

sum(index=0)¶

Returns the sum of all of the elements in self along the specified axis (first axis by default).

>>> print(PhysicalField(((1., 2.), (3., 4.)), "m").sum())
[ 4.  6.] m
>>> print(PhysicalField(((1., 2.), (3., 4.)), "m").sum(1))
[ 3.  7.] m

take(indices, axis=0)¶

Return the elements of self specified by the elements of indices. The resulting PhysicalField array has the same units as the original.

>>> print(PhysicalField((1., 2., 3.), "m").take((2, 0)))
[ 3.  1.] m

The optional third argument specifies the axis along which the selection occurs, and the default value (as in the example above) is 0, the first axis.

>>> print(PhysicalField(((1., 2., 3.), (4., 5., 6.)), "m").take((2, 0), axis = 1))
[[ 3.  1.]
 [ 6.  4.]] m

tan()¶

Return the tangent of the PhysicalField

>>> numerix.round(PhysicalField(numerix.pi/4, "rad").tan(), 6)
1.0
>>> numerix.round(PhysicalField(45, "deg").tan(), 6)
1.0

The units of the PhysicalField must be an angle

>>> PhysicalField(45., "m").tan()
Traceback (most recent call last):
    ...
TypeError: Incompatible units

tanh()¶

Return the hyperbolic tangent of the PhysicalField

>>> print(numerix.allclose(PhysicalField(1.).tanh(), 0.761594155956))
True

The units of the PhysicalField must be dimensionless

>>> PhysicalField(60., "m").tanh()
Traceback (most recent call last):
    ...
TypeError: Incompatible units

tostring(max_line_width=75, precision=8, suppress_small=False, separator=' ')¶

Return human-readable form of a physical quantity

>>> p = PhysicalField(value = (3., 3.14159), unit = "eV")
>>> print(p.tostring(precision = 3, separator = '|'))
[ 3.   | 3.142] eV

property unit¶

Return the unit object of self.

>>> PhysicalField("1 m").unit
<PhysicalUnit m>

fipy.tools.SharedTemporaryFile(mode='w+b', buffering=-1, encoding=None, newline=None, suffix='', prefix='tmp', dir=None, delete=True, communicator=DummyComm())¶

Create a temporary file shared by all MPI ranks.

The file is created as NamedTemporaryFile would do it. The name of the returned file-like object is accessible as its name attribute. The file will be automatically deleted when it is closed unless the delete argument is set to False.

>>> from fipy.tools import SharedTemporaryFile, parallelComm
>>> with SharedTemporaryFile(mode='w+', suffix=".tmp") as tmpFile:
...     # write on processor 0
...     if parallelComm.procID == 0:
...         _ = tmpFile.write("shared text")
...
...     parallelComm.Barrier()
...
...     # read on all processors
...     _ = tmpFile.seek(0)
...     txt = tmpFile.read()
>>> print(txt)
shared text

Parameters:

prefix (str) – As for mkstemp
suffix (str) – As for mkstemp
dir (str) – As for mkstemp
mode (str) – The mode argument to io.open (default “w+b”)
buffering (int) – The buffer size argument to io.open (default -1)
encoding (str or None) – The encoding argument to io.open (default None)
newline (str or None) – The newline argument to io.open (default None)
delete (bool) – Whether the file is deleted on close (default True)
communicator (CommWrapper) – MPI communicator describing ranks to share with. A duck-typed object with procID and Nproc attributes is sufficient.

Return type:

file-like object

`fipy.tools.comms`
`fipy.tools.debug`
`fipy.tools.decorators`
`fipy.tools.dimensions`
`fipy.tools.dump`
`fipy.tools.inline`
`fipy.tools.logging`
`fipy.tools.numerix`	Replacement module for NumPy
`fipy.tools.parser`
`fipy.tools.sharedtempfile`	This module provides a generic, high-level interface for creating shared temporary files.
`fipy.tools.test`
`fipy.tools.timer`
`fipy.tools.vector`	Vector utility functions that are inexplicably absent from Numeric
`fipy.tools.version`	Shim for version checking