mirror of
https://github.com/Cantera/cantera.git
synced 2025-02-25 18:55:29 -06:00
[Python] streamline implementation
Simplify code using: * explicit type conversions * leverage of set properties Co-authored-by: Ray Speth <speth@mit.edu>
This commit is contained in:
Ingmar Schoegl
committed by
Bryan W. Weber
Bryan W. Weber
parent
ae7b54fed0
commit
924c916ba9
@@ -98,7 +98,7 @@ class TestPyToAnyValue(utilities.CanteraTest):
|
||||
|
||||
def test_raises_named(self):
|
||||
self.check_raises({'abcd': [[['spam', 'eggs'], ['foo', 'bar']]]},
|
||||
NotImplementedError, "value with key 'abcd'")
|
||||
NotImplementedError, "with key 'abcd'")
|
||||
|
||||
def test_raises_int(self):
|
||||
self.check_raises([[[1, 2, 3], [4, 5, 6]]],
|
||||
|
||||
@@ -130,181 +130,161 @@ cdef anymap_to_dict(CxxAnyMap& m):
|
||||
for item in m.ordered()}
|
||||
|
||||
|
||||
cdef CxxAnyValue python_to_anyvalue(v, name=None) except *:
|
||||
cdef CxxAnyValue python_to_anyvalue(item, name=None) except *:
|
||||
# convert Python values to C++ AnyValue
|
||||
cdef CxxAnyMap a
|
||||
cdef CxxAnyValue b
|
||||
|
||||
if name is None:
|
||||
msg = "Unable to convert value to AnyValue:\n"
|
||||
msg = "Unable to convert item of type '{}' to AnyValue:\n".format(type(item))
|
||||
else:
|
||||
msg = "Unable to convert value with key '{}' to AnyValue:\n".format(name)
|
||||
msg = ("Unable to convert item of type '{}' with key '{}' to AnyValue:\n"
|
||||
"".format(type(item), name))
|
||||
|
||||
if v is None:
|
||||
if item is None:
|
||||
return b
|
||||
elif isinstance(v, set):
|
||||
elif isinstance(item, set):
|
||||
raise NotImplementedError("{}cannot process Python set.".format(msg))
|
||||
elif isinstance(v, dict):
|
||||
a = dict_to_anymap(v)
|
||||
b = a
|
||||
return b
|
||||
elif isinstance(item, dict):
|
||||
return <CxxAnyValue>dict_to_anymap(item)
|
||||
|
||||
def is_scalar(v):
|
||||
return not hasattr(v, '__len__') or isinstance(v, str)
|
||||
def is_scalar(item):
|
||||
return not hasattr(item, '__len__') or isinstance(item, str)
|
||||
|
||||
all_same = True
|
||||
multi_dict = False
|
||||
if not is_scalar(v):
|
||||
itype = set()
|
||||
if not (is_scalar(item) or isinstance(item, np.ndarray)):
|
||||
itype = set(type(i) for i in item)
|
||||
# ensure that data are homogeneous
|
||||
# (np.array converts inhomogeneous sequences to string arrays)
|
||||
if not len(v):
|
||||
if not len(item):
|
||||
# np.array converts empty arrays to float
|
||||
pass
|
||||
elif isinstance(v, np.ndarray):
|
||||
elif is_scalar(item[0]):
|
||||
# the content may be inhomogeneous or 1-D homogeneous, but it will
|
||||
# not be a 2-D array and thus does not have to be checked
|
||||
pass
|
||||
elif itype == {dict}:
|
||||
# list of dictionaries
|
||||
pass
|
||||
elif is_scalar(v[0]):
|
||||
all_same = all([type(val) == type(v[0]) for val in v])
|
||||
elif isinstance(v[0], dict):
|
||||
multi_dict = all([isinstance(val, dict) for val in v])
|
||||
else:
|
||||
all_same = []
|
||||
itype = set()
|
||||
sizes = []
|
||||
for row in range(len(v)):
|
||||
sizes.append(len(v[row]))
|
||||
all_same.append(all([type(val) == type(v[0][0]) for val in v[row]]))
|
||||
if not all(all_same):
|
||||
for row in item:
|
||||
sizes.append(len(row))
|
||||
itype = itype.union(set(type(val) for val in row))
|
||||
if len(itype) != 1:
|
||||
raise NotImplementedError(
|
||||
"{}cannot process nested sequences with inhomogeneous data "
|
||||
"types.".format(msg)
|
||||
)
|
||||
"{}cannot process nested sequences with inhomogeneous data "
|
||||
"types.".format(msg))
|
||||
elif np.unique(sizes).size != 1:
|
||||
raise NotImplementedError(
|
||||
"{}cannot process arrays represented by ragged nested "
|
||||
"sequences.".format(msg)
|
||||
)
|
||||
all_same = all(all_same)
|
||||
"{}cannot process arrays represented by ragged nested "
|
||||
"sequences.".format(msg))
|
||||
|
||||
cdef vector[CxxAnyValue] vv_any
|
||||
if not all_same:
|
||||
if len(itype) > 1:
|
||||
# inhomogeneous sequence
|
||||
for val in v:
|
||||
for val in item:
|
||||
vv_any.push_back(python_to_anyvalue(val, name))
|
||||
b = vv_any
|
||||
return b
|
||||
|
||||
cdef vector[CxxAnyMap] vv_map
|
||||
if multi_dict:
|
||||
for val in v:
|
||||
if itype == {dict}:
|
||||
for val in item:
|
||||
vv_map.push_back(dict_to_anymap(val))
|
||||
b = vv_map
|
||||
return b
|
||||
|
||||
# data are homogeneous: convert to np.array for convenience
|
||||
vv = np.array(v)
|
||||
vv = np.array(item)
|
||||
ndim = vv.ndim
|
||||
|
||||
cdef double v_double
|
||||
cdef vector[double] vv_double
|
||||
cdef vector[vector[double]] vvv_double
|
||||
if vv.dtype == float:
|
||||
if vv.size == 0:
|
||||
# empty list
|
||||
return b
|
||||
if ndim == 0:
|
||||
v_double = v
|
||||
b = v_double
|
||||
return b
|
||||
if ndim == 1:
|
||||
pass
|
||||
elif ndim == 0:
|
||||
b = <double>item
|
||||
elif ndim == 1:
|
||||
for val in vv:
|
||||
vv_double.push_back(val)
|
||||
vv_double.push_back(<double>val)
|
||||
b = vv_double
|
||||
return b
|
||||
if ndim == 2:
|
||||
elif ndim == 2:
|
||||
vvv_double.resize(vv.shape[0])
|
||||
for row in range(vv.shape[0]):
|
||||
for val in vv[row, :]:
|
||||
vvv_double[row].push_back(val)
|
||||
vvv_double[row].push_back(<double>val)
|
||||
b = vvv_double
|
||||
return b
|
||||
raise NotImplementedError(
|
||||
"{}cannot process float array with {} dimensions".format(msg, ndim)
|
||||
)
|
||||
else:
|
||||
raise NotImplementedError(
|
||||
"{}cannot process float array with {} dimensions".format(msg, ndim))
|
||||
return b
|
||||
|
||||
cdef long v_int
|
||||
cdef vector[long] vv_int
|
||||
cdef vector[vector[long]] vvv_int
|
||||
if vv.dtype == int:
|
||||
if ndim == 0:
|
||||
v_int = v
|
||||
b = v_int
|
||||
return b
|
||||
if ndim == 1:
|
||||
b = <long>item
|
||||
elif ndim == 1:
|
||||
for val in vv:
|
||||
vv_int.push_back(val)
|
||||
vv_int.push_back(<long>val)
|
||||
b = vv_int
|
||||
return b
|
||||
if ndim == 2:
|
||||
elif ndim == 2:
|
||||
vvv_int.resize(vv.shape[0])
|
||||
for row in range(vv.shape[0]):
|
||||
for val in vv[row, :]:
|
||||
vvv_int[row].push_back(val)
|
||||
vvv_int[row].push_back(<long>val)
|
||||
b = vvv_int
|
||||
return b
|
||||
raise NotImplementedError(
|
||||
"{}cannot process integer array with {} dimensions".format(msg, ndim)
|
||||
)
|
||||
else:
|
||||
raise NotImplementedError(
|
||||
"{}cannot process integer array with {} dimensions".format(msg, ndim))
|
||||
return b
|
||||
|
||||
cdef cbool v_bool
|
||||
cdef vector[cbool] vv_bool
|
||||
cdef vector[vector[cbool]] vvv_bool
|
||||
if vv.dtype == bool:
|
||||
if ndim == 0:
|
||||
v_bool = v
|
||||
b = v_bool
|
||||
return b
|
||||
if ndim == 1:
|
||||
b = <cbool>item
|
||||
elif ndim == 1:
|
||||
for val in vv:
|
||||
vv_bool.push_back(val)
|
||||
vv_bool.push_back(<cbool>val)
|
||||
b = vv_bool
|
||||
return b
|
||||
if ndim == 2:
|
||||
elif ndim == 2:
|
||||
vvv_bool.resize(vv.shape[0])
|
||||
for row in range(vv.shape[0]):
|
||||
for val in vv[row, :]:
|
||||
vvv_bool[row].push_back(val)
|
||||
vvv_bool[row].push_back(<cbool>val)
|
||||
b = vvv_bool
|
||||
return b
|
||||
raise NotImplementedError(
|
||||
"{}cannot process boolean array with {} dimensions".format(msg, ndim)
|
||||
)
|
||||
else:
|
||||
raise NotImplementedError(
|
||||
"{}cannot process boolean array with {} dimensions".format(msg, ndim))
|
||||
return b
|
||||
|
||||
cdef string v_string
|
||||
cdef vector[string] vv_string
|
||||
cdef vector[vector[string]] vvv_string
|
||||
if isinstance(v, str) or (ndim and isinstance(vv.ravel()[0], str)):
|
||||
if isinstance(item, str) or (ndim and isinstance(vv.ravel()[0], str)):
|
||||
if ndim == 0:
|
||||
v_string = stringify(v)
|
||||
b = v_string
|
||||
return b
|
||||
if ndim == 1:
|
||||
b = stringify(item)
|
||||
elif ndim == 1:
|
||||
for val in vv:
|
||||
vv_string.push_back(stringify(val))
|
||||
b = vv_string
|
||||
return b
|
||||
if ndim == 2:
|
||||
elif ndim == 2:
|
||||
vvv_string.resize(vv.shape[0])
|
||||
for row in range(vv.shape[0]):
|
||||
for val in vv[row, :]:
|
||||
vvv_string[row].push_back(stringify(val))
|
||||
b = vvv_string
|
||||
return b
|
||||
raise NotImplementedError(
|
||||
"{}cannot process string array with {} dimensions".format(msg, ndim)
|
||||
)
|
||||
else:
|
||||
raise NotImplementedError(
|
||||
"{}cannot process string array with {} dimensions".format(msg, ndim))
|
||||
return b
|
||||
|
||||
raise NotImplementedError(
|
||||
"{}unknown conversion for variable with value\n".format(msg,v)
|
||||
)
|
||||
"{}unknown conversion for variable with value\n".format(msg, item))
|
||||
|
||||
|
||||
cdef CxxAnyMap dict_to_anymap(dict dd) except *:
|
||||
|
||||
Reference in New Issue
Block a user