Python numpy 模块，inexact() 实例源码
我们从Python开源项目中，提取了以下19个代码示例，用于说明如何使用numpy.inexact()。
def _replace_nan(a, val):
"""
If `a` is of inexact type, make a copy of `a`, replace NaNs with
the `val` value, and return the copy together with a boolean mask
marking the locations where NaNs were present. If `a` is not of
inexact type, do nothing and return `a` together with a mask of None.
Note that scalars will end up as array scalars, which is important
for using the result as the value of the out argument in some
operations.
Parameters
----------
a : array-like
Input array.
val : float
NaN values are set to val before doing the operation.
Returns
-------
y : ndarray
If `a` is of inexact type, return a copy of `a` with the NaNs
replaced by the fill value, otherwise return `a`.
mask: {bool, None}
If `a` is of inexact type, return a boolean mask marking locations of
NaNs, otherwise return None.
"""
is_new = not isinstance(a, np.ndarray)
if is_new:
a = np.array(a)
if not issubclass(a.dtype.type, np.inexact):
return a, None
if not is_new:
# need copy
a = np.array(a, subok=True)
mask = np.isnan(a)
np.copyto(a, val, where=mask)
return a, mask
def _divide_by_count(a, b, out=None):
"""
Compute a/b ignoring invalid results. If `a` is an array the division
is done in place. If `a` is a scalar, then its type is preserved in the
output. If out is None, then then a is used instead so that the
division is in place. Note that this is only called with `a` an inexact
type.
Parameters
----------
a : {ndarray, numpy scalar}
Numerator. Expected to be of inexact type but not checked.
b : {ndarray, numpy scalar}
Denominator.
out : ndarray, optional
Alternate output array in which to place the result. The default
is ``None``; if provided, it must have the same shape as the
expected output, but the type will be cast if necessary.
Returns
-------
ret : {ndarray, numpy scalar}
The return value is a/b. If `a` was an ndarray the division is done
in place. If `a` is a numpy scalar, the division preserves its type.
"""
with np.errstate(invalid='ignore'):
if isinstance(a, np.ndarray):
if out is None:
return np.divide(a, b, out=a, casting='unsafe')
else:
return np.divide(a, b, out=out, casting='unsafe')
else:
if out is None:
return a.dtype.type(a / b)
else:
# This is questionable, but currently a numpy scalar can
# be output to a zero dimensional array.
return np.divide(a, b, out=out, casting='unsafe')
def _replace_nan(a, val):
"""
If `a` is of inexact type, make a copy of `a`, replace NaNs with
the `val` value, and return the copy together with a boolean mask
marking the locations where NaNs were present. If `a` is not of
inexact type, do nothing and return `a` together with a mask of None.
Note that scalars will end up as array scalars, which is important
for using the result as the value of the out argument in some
operations.
Parameters
----------
a : array-like
Input array.
val : float
NaN values are set to val before doing the operation.
Returns
-------
y : ndarray
If `a` is of inexact type, return a copy of `a` with the NaNs
replaced by the fill value, otherwise return `a`.
mask: {bool, None}
If `a` is of inexact type, return a boolean mask marking locations of
NaNs, otherwise return None.
"""
is_new = not isinstance(a, np.ndarray)
if is_new:
a = np.array(a)
if not issubclass(a.dtype.type, np.inexact):
return a, None
if not is_new:
# need copy
a = np.array(a, subok=True)
mask = np.isnan(a)
np.copyto(a, val, where=mask)
return a, mask
def _divide_by_count(a, b, out=None):
"""
Compute a/b ignoring invalid results. If `a` is an array the division
is done in place. If `a` is a scalar, then its type is preserved in the
output. If out is None, then then a is used instead so that the
division is in place. Note that this is only called with `a` an inexact
type.
Parameters
----------
a : {ndarray, numpy scalar}
Numerator. Expected to be of inexact type but not checked.
b : {ndarray, numpy scalar}
Denominator.
out : ndarray, optional
Alternate output array in which to place the result. The default
is ``None``; if provided, it must have the same shape as the
expected output, but the type will be cast if necessary.
Returns
-------
ret : {ndarray, numpy scalar}
The return value is a/b. If `a` was an ndarray the division is done
in place. If `a` is a numpy scalar, the division preserves its type.
"""
with np.errstate(invalid='ignore'):
if isinstance(a, np.ndarray):
if out is None:
return np.divide(a, b, out=a, casting='unsafe')
else:
return np.divide(a, b, out=out, casting='unsafe')
else:
if out is None:
return a.dtype.type(a / b)
else:
# This is questionable, but currently a numpy scalar can
# be output to a zero dimensional array.
return np.divide(a, b, out=out, casting='unsafe')
def _replace_nan(a, val):
"""
If `a` is of inexact type, make a copy of `a`, replace NaNs with
the `val` value, and return the copy together with a boolean mask
marking the locations where NaNs were present. If `a` is not of
inexact type, do nothing and return `a` together with a mask of None.
Note that scalars will end up as array scalars, which is important
for using the result as the value of the out argument in some
operations.
Parameters
----------
a : array-like
Input array.
val : float
NaN values are set to val before doing the operation.
Returns
-------
y : ndarray
If `a` is of inexact type, return a copy of `a` with the NaNs
replaced by the fill value, otherwise return `a`.
mask: {bool, None}
If `a` is of inexact type, return a boolean mask marking locations of
NaNs, otherwise return None.
"""
is_new = not isinstance(a, np.ndarray)
if is_new:
a = np.array(a)
if not issubclass(a.dtype.type, np.inexact):
return a, None
if not is_new:
# need copy
a = np.array(a, subok=True)
mask = np.isnan(a)
np.copyto(a, val, where=mask)
return a, mask
def _divide_by_count(a, b, out=None):
"""
Compute a/b ignoring invalid results. If `a` is an array the division
is done in place. If `a` is a scalar, then its type is preserved in the
output. If out is None, then then a is used instead so that the
division is in place. Note that this is only called with `a` an inexact
type.
Parameters
----------
a : {ndarray, numpy scalar}
Numerator. Expected to be of inexact type but not checked.
b : {ndarray, numpy scalar}
Denominator.
out : ndarray, optional
Alternate output array in which to place the result. The default
is ``None``; if provided, it must have the same shape as the
expected output, but the type will be cast if necessary.
Returns
-------
ret : {ndarray, numpy scalar}
The return value is a/b. If `a` was an ndarray the division is done
in place. If `a` is a numpy scalar, the division preserves its type.
"""
with np.errstate(invalid='ignore'):
if isinstance(a, np.ndarray):
if out is None:
return np.divide(a, b, out=a, casting='unsafe')
else:
return np.divide(a, b, out=out, casting='unsafe')
else:
if out is None:
return a.dtype.type(a / b)
else:
# This is questionable, but currently a numpy scalar can
# be output to a zero dimensional array.
return np.divide(a, b, out=out, casting='unsafe')
def minimum_dtype(x, dtype=np.bool_):
"""returns the "most basic" dtype which represents `x` properly, which
provides at least the same value range as the specified dtype."""
def check_type(x, dtype):
try:
converted = dtype.type(x)
except (ValueError, OverflowError):
return False
# False if some overflow has happened
return converted == x or math.isnan(x)
def type_loop(x, dtype, dtype_dict, default=None):
while True:
try:
dtype = np.dtype(dtype_dict[dtype.name])
if check_type(x, dtype):
return np.dtype(dtype)
except KeyError:
if default is not None:
return np.dtype(default)
raise ValueError("Can not determine dtype of %r" % x)
dtype = np.dtype(dtype)
if check_type(x, dtype):
return dtype
if np.issubdtype(dtype, np.inexact):
return type_loop(x, dtype, _next_float_dtype)
else:
return type_loop(x, dtype, _next_int_dtype, default=np.float32)
def _replace_nan(a, val):
"""
If `a` is of inexact type, make a copy of `a`, replace NaNs with
the `val` value, and return the copy together with a boolean mask
marking the locations where NaNs were present. If `a` is not of
inexact type, do nothing and return `a` together with a mask of None.
Note that scalars will end up as array scalars, which is important
for using the result as the value of the out argument in some
operations.
Parameters
----------
a : array-like
Input array.
val : float
NaN values are set to val before doing the operation.
Returns
-------
y : ndarray
If `a` is of inexact type, return a copy of `a` with the NaNs
replaced by the fill value, otherwise return `a`.
mask: {bool, None}
If `a` is of inexact type, return a boolean mask marking locations of
NaNs, otherwise return None.
"""
is_new = not isinstance(a, np.ndarray)
if is_new:
a = np.array(a)
if not issubclass(a.dtype.type, np.inexact):
return a, None
if not is_new:
# need copy
a = np.array(a, subok=True)
mask = np.isnan(a)
np.copyto(a, val, where=mask)
return a, mask
def _divide_by_count(a, b, out=None):
"""
Compute a/b ignoring invalid results. If `a` is an array the division
is done in place. If `a` is a scalar, then its type is preserved in the
output. If out is None, then then a is used instead so that the
division is in place. Note that this is only called with `a` an inexact
type.
Parameters
----------
a : {ndarray, numpy scalar}
Numerator. Expected to be of inexact type but not checked.
b : {ndarray, numpy scalar}
Denominator.
out : ndarray, optional
Alternate output array in which to place the result. The default
is ``None``; if provided, it must have the same shape as the
expected output, but the type will be cast if necessary.
Returns
-------
ret : {ndarray, numpy scalar}
The return value is a/b. If `a` was an ndarray the division is done
in place. If `a` is a numpy scalar, the division preserves its type.
"""
with np.errstate(invalid='ignore'):
if isinstance(a, np.ndarray):
if out is None:
return np.divide(a, b, out=a, casting='unsafe')
else:
return np.divide(a, b, out=out, casting='unsafe')
else:
if out is None:
return a.dtype.type(a / b)
else:
# This is questionable, but currently a numpy scalar can
# be output to a zero dimensional array.
return np.divide(a, b, out=out, casting='unsafe')
def _replace_nan(a, val):
"""
If `a` is of inexact type, make a copy of `a`, replace NaNs with
the `val` value, and return the copy together with a boolean mask
marking the locations where NaNs were present. If `a` is not of
inexact type, do nothing and return `a` together with a mask of None.
Note that scalars will end up as array scalars, which is important
for using the result as the value of the out argument in some
operations.
Parameters
----------
a : array-like
Input array.
val : float
NaN values are set to val before doing the operation.
Returns
-------
y : ndarray
If `a` is of inexact type, return a copy of `a` with the NaNs
replaced by the fill value, otherwise return `a`.
mask: {bool, None}
If `a` is of inexact type, return a boolean mask marking locations of
NaNs, otherwise return None.
"""
is_new = not isinstance(a, np.ndarray)
if is_new:
a = np.array(a)
if not issubclass(a.dtype.type, np.inexact):
return a, None
if not is_new:
# need copy
a = np.array(a, subok=True)
mask = np.isnan(a)
np.copyto(a, val, where=mask)
return a, mask
def _divide_by_count(a, b, out=None):
"""
Compute a/b ignoring invalid results. If `a` is an array the division
is done in place. If `a` is a scalar, then its type is preserved in the
output. If out is None, then then a is used instead so that the
division is in place. Note that this is only called with `a` an inexact
type.
Parameters
----------
a : {ndarray, numpy scalar}
Numerator. Expected to be of inexact type but not checked.
b : {ndarray, numpy scalar}
Denominator.
out : ndarray, optional
Alternate output array in which to place the result. The default
is ``None``; if provided, it must have the same shape as the
expected output, but the type will be cast if necessary.
Returns
-------
ret : {ndarray, numpy scalar}
The return value is a/b. If `a` was an ndarray the division is done
in place. If `a` is a numpy scalar, the division preserves its type.
"""
with np.errstate(invalid='ignore', divide='ignore'):
if isinstance(a, np.ndarray):
if out is None:
return np.divide(a, b, out=a, casting='unsafe')
else:
return np.divide(a, b, out=out, casting='unsafe')
else:
if out is None:
return a.dtype.type(a / b)
else:
# This is questionable, but currently a numpy scalar can
# be output to a zero dimensional array.
return np.divide(a, b, out=out, casting='unsafe')
def _median(a, axis=None, out=None, overwrite_input=False):
# can't be reasonably be implemented in terms of percentile as we have to
# call mean to not break astropy
a = np.asanyarray(a)
# Set the partition indexes
if axis is None:
sz = a.size
else:
sz = a.shape[axis]
if sz % 2 == 0:
szh = sz // 2
kth = [szh - 1, szh]
else:
kth = [(sz - 1) // 2]
# Check if the array contains any nan's
if np.issubdtype(a.dtype, np.inexact):
kth.append(-1)
if overwrite_input:
if axis is None:
part = a.ravel()
part.partition(kth)
else:
a.partition(kth, axis=axis)
part = a
else:
part = partition(a, kth, axis=axis)
if part.shape == ():
# make 0-D arrays work
return part.item()
if axis is None:
axis = 0
indexer = [slice(None)] * part.ndim
index = part.shape[axis] // 2
if part.shape[axis] % 2 == 1:
# index with slice to allow mean (below) to work
indexer[axis] = slice(index, index+1)
else:
indexer[axis] = slice(index-1, index+1)
# Check if the array contains any nan's
if np.issubdtype(a.dtype, np.inexact) and sz > 0:
# warn and return nans like mean would
rout = mean(part[indexer], axis=axis, out=out)
return np.lib.utils._median_nancheck(part, rout, axis, out)
else:
# if there are no nans
# Use mean in odd and even case to coerce data type
# and check, use out array.
return mean(part[indexer], axis=axis, out=out)
def object_to_json(o, level=0, indent=4, space=" ", newline="\n"):
ret = ""
if isinstance(o, dict):
ret += "{" + newline
comma = ""
for k,v in o.iteritems():
ret += comma
comma = ",\n"
ret += space * indent * level
ret += '"' + str(k) + '":' + space
ret += object_to_json(v, level + 1)
ret += newline + space * indent * (level - 1) + "}"
elif isinstance(o, basestring):
ret += '"' + o + '"'
elif isinstance(o, list):
ret += "[" + ", ".join([object_to_json(e, level+1) for e in o]) + "]"
elif isinstance(o, bool):
ret += "true" if o else "false"
elif isinstance(o, (int, long)):
ret += str(o)
elif isinstance(o, datetime):
ret += o.strftime('%Y-%m-%d %H:%M:%S.%f')
elif isinstance(o, bson.ObjectId):
ret += 'ObjectId(%s)' % o
elif isinstance(o, float):
ret += '%.7g' % o
elif isinstance(o, numpy.ndarray) and numpy.issubdtype(o.dtype, numpy.integer):
ret += "[" + ','.join(map(str, o.flatten().tolist())) + "]"
elif isinstance(o, numpy.ndarray) and numpy.issubdtype(o.dtype, numpy.inexact):
ret += "[" + ','.join(map(lambda x: '%.7g' % x, o.flatten().tolist())) + "]"
elif isinstance(o, bson.timestamp.Timestamp):
ret += o.as_datetime().strftime('%Y-%m-%d %H:%M:%S.%f')
elif o is None:
ret += 'null'
else:
raise TypeError("Unknown type '%s' for json serialization" % str(type(o)))
return ret
def _replace_nan(a, val):
"""
If `a` is of inexact type, make a copy of `a`, replace NaNs with
the `val` value, and return the copy together with a boolean mask
marking the locations where NaNs were present. If `a` is not of
inexact type, do nothing and return `a` together with a mask of None.
Note that scalars will end up as array scalars, which is important
for using the result as the value of the out argument in some
operations.
Parameters
----------
a : array-like
Input array.
val : float
NaN values are set to val before doing the operation.
Returns
-------
y : ndarray
If `a` is of inexact type, return a copy of `a` with the NaNs
replaced by the fill value, otherwise return `a`.
mask: {bool, None}
If `a` is of inexact type, return a boolean mask marking locations of
NaNs, otherwise return None.
"""
is_new = not isinstance(a, np.ndarray)
if is_new:
a = np.array(a)
if not issubclass(a.dtype.type, np.inexact):
return a, None
if not is_new:
# need copy
a = np.array(a, subok=True)
mask = np.isnan(a)
np.copyto(a, val, where=mask)
return a, mask
def _divide_by_count(a, b, out=None):
"""
Compute a/b ignoring invalid results. If `a` is an array the division
is done in place. If `a` is a scalar, then its type is preserved in the
output. If out is None, then then a is used instead so that the
division is in place. Note that this is only called with `a` an inexact
type.
Parameters
----------
a : {ndarray, numpy scalar}
Numerator. Expected to be of inexact type but not checked.
b : {ndarray, numpy scalar}
Denominator.
out : ndarray, optional
Alternate output array in which to place the result. The default
is ``None``; if provided, it must have the same shape as the
expected output, but the type will be cast if necessary.
Returns
-------
ret : {ndarray, numpy scalar}
The return value is a/b. If `a` was an ndarray the division is done
in place. If `a` is a numpy scalar, the division preserves its type.
"""
with np.errstate(invalid='ignore'):
if isinstance(a, np.ndarray):
if out is None:
return np.divide(a, b, out=a, casting='unsafe')
else:
return np.divide(a, b, out=out, casting='unsafe')
else:
if out is None:
return a.dtype.type(a / b)
else:
# This is questionable, but currently a numpy scalar can
# be output to a zero dimensional array.
return np.divide(a, b, out=out, casting='unsafe')
def nan_to_num(array, fill_value = 0.0, copy = True):
"""
Replace NaNs with another fill value.
Parameters
----------
array : array_like
Input data.
fill_value : float, optional
NaNs will be replaced by ``fill_value``. Default is 0.0, in keeping
with ``numpy.nan_to_num``.
copy : bool, optional
Whether to create a copy of `array` (True) or to replace values
in-place (False). The in-place operation only occurs if
casting to an array does not require a copy.
Returns
-------
out : ndarray
Array without NaNs. If ``array`` was not of floating or complearray type,
``array`` is returned unchanged.
Notes
-----
Contrary to ``numpy.nan_to_num``, this functions does not handle
infinite values.
See Also
--------
numpy.nan_to_num : replace NaNs and Infs with zeroes.
"""
array = np.array(array, subok = True, copy = copy)
dtype = array.dtype.type
# Non-inexact types do not have NaNs
if not np.issubdtype(dtype, np.inexact):
return array
iscomplex = np.issubdtype(dtype, np.complexfloating)
dest = (array.real, array.imag) if iscomplex else (array,)
for d in dest:
np.copyto(d, fill_value, where = np.isnan(d))
return array
def to_json(o, level=0):
INDENT = 2
SPACE = " "
NEWLINE = "\n"
ret = ""
if isinstance(o, dict):
ret += "{" + NEWLINE
comma = ""
for k in sorted(o.keys()):
v = o[k]
ret += comma
comma = ",\n"
ret += SPACE * INDENT * (level+1)
ret += '"' + str(k) + '":' + SPACE
ret += to_json(v, level + 1)
ret += NEWLINE + SPACE * INDENT * level + "}"
elif isinstance(o, str):
ret += '"' + o + '"'
elif isinstance(o, list) or isinstance(o, tuple):
ret += "[" + ",".join([to_json(e, level+1) for e in o]) + "]"
elif isinstance(o, bool):
ret += "true" if o else "false"
elif isinstance(o, int):
ret += str(o)
elif isinstance(o, float):
ret += '%.7g' % o
elif isinstance(o, np.ndarray) and np.issubdtype(o.dtype, np.integer):
ret += "[" + ','.join(map(str, o.flatten().tolist())) + "]"
elif isinstance(o, np.ndarray) and np.issubdtype(o.dtype, np.inexact):
ret += "[" + \
','.join(map(lambda x: '%.7g' % x, o.flatten().tolist())) + "]"
elif o is None:
ret += 'null'
elif hasattr(o, '__class__'):
ret += '"' + o.__class__.__name__ + '"'
else:
raise TypeError(
"Unknown type '%s' for json serialization" % str(type(o)))
return ret
# format object and its properties into printable dict
def check_dtype(dtype, func_str, a, n):
if np.isscalar(a) or not a.shape:
if func_str not in ("sum", "prod", "len"):
raise ValueError("scalar inputs are supported only for 'sum', "
"'prod' and 'len'")
a_dtype = np.dtype(type(a))
else:
a_dtype = a.dtype
if dtype is not None:
# dtype set by the user
# Careful here: np.bool != np.bool_ !
if np.issubdtype(dtype, np.bool_) and \
not('all' in func_str or 'any' in func_str):
raise TypeError("function %s requires a more complex datatype "
"than bool" % func_str)
if not np.issubdtype(dtype, np.integer) and func_str in ('len', 'nanlen'):
raise TypeError("function %s requires an integer datatype" % func_str)
# TODO: Maybe have some more checks here
return np.dtype(dtype)
else:
try:
return np.dtype(_forced_types[func_str])
except KeyError:
if func_str in _forced_float_types:
if np.issubdtype(a_dtype, np.floating):
return a_dtype
else:
return np.dtype(np.float64)
else:
if func_str == 'sum':
# Try to guess the minimally required int size
if np.issubdtype(a_dtype, np.int64):
# It's not getting bigger anymore
# TODO: strictly speaking it might need float
return np.dtype(np.int64)
elif np.issubdtype(a_dtype, np.integer):
maxval = np.iinfo(a_dtype).max * n
return minimum_dtype(maxval, a_dtype)
elif np.issubdtype(a_dtype, np.bool_):
return minimum_dtype(n, a_dtype)
else:
# floating, inexact, whatever
return a_dtype
elif func_str in _forced_same_type:
return a_dtype
else:
if isinstance(a_dtype, np.integer):
return np.dtype(np.int64)
else:
return a_dtype
def _median(a, axis=None, out=None, overwrite_input=False):
if overwrite_input:
if axis is None:
asorted = a.ravel()
asorted.sort()
else:
a.sort(axis=axis)
asorted = a
else:
asorted = sort(a, axis=axis)
if axis is None:
axis = 0
elif axis < 0:
axis += asorted.ndim
if asorted.ndim == 1:
idx, odd = divmod(count(asorted), 2)
return asorted[idx + odd - 1 : idx + 1].mean(out=out)
counts = count(asorted, axis=axis)
h = counts // 2
# create indexing mesh grid for all but reduced axis
axes_grid = [np.arange(x) for i, x in enumerate(asorted.shape)
if i != axis]
ind = np.meshgrid(*axes_grid, sparse=True, indexing='ij')
# insert indices of low and high median
ind.insert(axis, np.maximum(0, h - 1))
low = asorted[tuple(ind)]
ind[axis] = h
high = asorted[tuple(ind)]
# duplicate high if odd number of elements so mean does nothing
odd = counts % 2 == 1
if asorted.ndim > 1:
np.copyto(low, high, where=odd)
elif odd:
low = high
if np.issubdtype(asorted.dtype, np.inexact):
# avoid inf / x = masked
s = np.ma.sum([low, high], axis=0, out=out)
np.true_divide(s.data, 2., casting='unsafe', out=s.data)
else:
s = np.ma.mean([low, high], axis=0, out=out)
return s