def doprint(self, funcname, args, expr):
"""Returns the function definition code as a string."""
from sympy import Dummy
funcbody = []
if not iterable(args):
args = [args]
argstrs, expr = self._preprocess(args, expr)
# Generate argument unpacking and final argument list
funcargs = []
unpackings = []
for argstr in argstrs:
if iterable(argstr):
funcargs.append(self._argrepr(Dummy()))
unpackings.extend(self._print_unpacking(argstr, funcargs[-1]))
else:
funcargs.append(argstr)
funcsig = 'def {}({}):'.format(funcname, ', '.join(funcargs))
# Wrap input arguments before unpacking
funcbody.extend(self._print_funcargwrapping(funcargs))
funcbody.extend(unpackings)
funcbody.append('return ({})'.format(self._exprrepr(expr)))
funclines = [funcsig]
funclines.extend(' ' + line for line in funcbody)
return '\n'.join(funclines) + '\n'
def reduce_inequalities(inequalities, symbols=[]):
"""Reduce a system of inequalities with rational coefficients.
Examples
========
>>> from sympy import sympify as S, Symbol
>>> from sympy.abc import x, y
>>> from sympy.solvers.inequalities import reduce_inequalities
>>> reduce_inequalities(0 <= x + 3, [])
And(-3 <= x, x < oo)
>>> reduce_inequalities(0 <= x + y*2 - 1, [x])
x >= -2*y + 1
"""
if not iterable(inequalities):
inequalities = [inequalities]
inequalities = [sympify(i) for i in inequalities]
gens = set().union(*[i.free_symbols for i in inequalities])
if not iterable(symbols):
symbols = [symbols]
symbols = (set(symbols) or gens) & gens
if any(i.is_real is False for i in symbols):
raise TypeError(
filldedent(
"""
inequalities cannot contain symbols that are not real."""
)
)
# make vanilla symbol real
recast = dict([(i, Dummy(i.name, real=True)) for i in gens if i.is_real is None])
inequalities = [i.xreplace(recast) for i in inequalities]
symbols = set([i.xreplace(recast) for i in symbols])
# prefilter
keep = []
for i in inequalities:
if isinstance(i, Relational):
i = i.func(i.lhs.as_expr() - i.rhs.as_expr(), 0)
elif i not in (True, False):
i = Eq(i, 0)
if i == True:
continue
elif i == False:
return S.false
if i.lhs.is_number:
raise NotImplementedError("could not determine truth value of %s" % i)
keep.append(i)
inequalities = keep
del keep
# solve system
rv = _reduce_inequalities(inequalities, symbols)
# restore original symbols and return
return rv.xreplace(dict([(v, k) for k, v in recast.items()]))
def test_iterable_is_sequence():
ordered = [list(), tuple(), Tuple(), Matrix([[]])]
unordered = [set()]
not_sympy_iterable = [{}, '', u('')]
assert all(is_sequence(i) for i in ordered)
assert all(not is_sequence(i) for i in unordered)
assert all(iterable(i) for i in ordered + unordered)
assert all(not iterable(i) for i in not_sympy_iterable)
assert all(iterable(i, exclude=None) for i in not_sympy_iterable)
def test_iterable_ordered_iter():
ordered = [list(), tuple(), Tuple(), Matrix([[]])]
unordered = [set()]
not_sympy_iterable = [{}, '']
assert all(ordered_iter(i) for i in ordered)
assert all(not ordered_iter(i) for i in unordered)
assert all(iterable(i) for i in ordered + unordered)
assert all(not iterable(i) for i in not_sympy_iterable)
assert all(iterable(i, exclude=None) for i in not_sympy_iterable)
def test_iterable():
assert iterable(0) is False
assert iterable(1) is False
assert iterable(None) is False
class Test1(NotIterable):
pass
assert iterable(Test1()) is False
class Test2(NotIterable):
_iterable = True
assert iterable(Test2()) is True
class Test3(object):
pass
assert iterable(Test3()) is False
class Test4(object):
_iterable = True
assert iterable(Test4()) is True
class Test5(object):
def __iter__(self):
yield 1
assert iterable(Test5()) is True
class Test6(Test5):
_iterable = False
assert iterable(Test6()) is False
def _find_opts(expr):
if expr.is_Atom or expr.is_Order:
return
if iterable(expr):
list(map(_find_opts, expr))
return
if expr in seen_subexp:
return expr
seen_subexp.add(expr)
list(map(_find_opts, expr.args))
if _coeff_isneg(expr):
neg_expr = -expr
if not neg_expr.is_Atom:
opt_subs[expr] = Mul(S.NegativeOne, neg_expr, evaluate=False)
seen_subexp.add(neg_expr)
expr = neg_expr
if expr.is_Mul:
muls.add(expr)
elif expr.is_Add:
adds.add(expr)
elif expr.is_Pow:
if _coeff_isneg(expr.exp):
opt_subs[expr] = Pow(Pow(expr.base, -expr.exp), S.NegativeOne,
evaluate=False)
def _find_repeated(expr):
if not isinstance(expr, Basic):
return
if expr.is_Atom or expr.is_Order:
return
if iterable(expr):
args = expr
else:
if expr in seen_subexp:
for ign in ignore:
if ign in expr.free_symbols:
break
else:
to_eliminate.add(expr)
return
seen_subexp.add(expr)
if expr in opt_subs:
expr = opt_subs[expr]
args = expr.args
list(map(_find_repeated, args))
def _find_opts(expr):
if not isinstance(expr, Basic):
return
if expr.is_Atom or expr.is_Order:
return
if iterable(expr):
list(map(_find_opts, expr))
return
if expr in seen_subexp:
return expr
seen_subexp.add(expr)
list(map(_find_opts, expr.args))
if _coeff_isneg(expr):
neg_expr = -expr
if not neg_expr.is_Atom:
opt_subs[expr] = Mul(S.NegativeOne, neg_expr, evaluate=False)
seen_subexp.add(neg_expr)
expr = neg_expr
if isinstance(expr, (Mul, MatMul)):
muls.add(expr)
elif isinstance(expr, (Add, MatAdd)):
adds.add(expr)
elif isinstance(expr, (Pow, MatPow)):
if _coeff_isneg(expr.exp):
opt_subs[expr] = Pow(Pow(expr.base, -expr.exp), S.NegativeOne,
evaluate=False)
def fuzzy_and(*args):
"""Return True (all True), False (any False) or None.
If `a` is an iterable it must have more than one element."""
if (len(args) == 1 and iterable(args[0]) or
len(args) > 2):
if len(args) == 1:
args = args[0]
rv = True
i = 0
for ai in args:
ai = fuzzy_bool(ai)
if ai is False:
return False
if rv: # this will stop updating if a None is ever trapped
rv = ai
i += 1
if i < 2:
raise ValueError('iterables must have 2 or more elements')
return rv
a, b = [fuzzy_bool(i) for i in args]
if a is True and b is True:
return True
elif a is False or b is False:
return False
def __new__(cls, *args, **kwargs):
evaluate = kwargs.get('evaluate', global_evaluate[0])
if iterable(args[0]):
if isinstance(args[0], Point) and not evaluate:
return args[0]
args = args[0]
# unpack the arguments into a friendly Tuple
# if we were already a Point, we're doing an excess
# iteration, but we'll worry about efficiency later
coords = Tuple(*args)
if any(a.is_number and im(a) for a in coords):
raise ValueError('Imaginary coordinates not permitted.')
# Turn any Floats into rationals and simplify
# any expressions before we instantiate
if evaluate:
coords = coords.xreplace(dict(
[(f, simplify(nsimplify(f, rational=True)))
for f in coords.atoms(Float)]))
if len(coords) == 2:
return Point2D(coords, **kwargs)
if len(coords) == 3:
return Point3D(coords, **kwargs)
return GeometryEntity.__new__(cls, *coords)
def lambdastr(args, expr, printer=None):
"""
Returns a string that can be evaluated to a lambda function.
>>> from sympy.abc import x, y, z
>>> from sympy.utilities.lambdify import lambdastr
>>> lambdastr(x, x**2)
'lambda x: (x**2)'
>>> lambdastr((x,y,z), [z,y,x])
'lambda x,y,z: ([z, y, x])'
"""
if printer is not None:
if inspect.isfunction(printer):
lambdarepr = printer
else:
if inspect.isclass(printer):
lambdarepr = lambda expr: printer().doprint(expr)
else:
lambdarepr = lambda expr: printer.doprint(expr)
else:
#XXX: This has to be done here because of circular imports
from sympy.printing.lambdarepr import lambdarepr
# Transform everything to strings.
from sympy.matrices import DeferredVector
expr = lambdarepr(expr)
if isinstance(args, str):
pass
elif iterable(args, exclude=DeferredVector):
args = ",".join(str(a) for a in args)
else:
args = str(args)
return "lambda %s: (%s)" % (args, expr)
def _walsh_hadamard_transform(seq, inverse=False):
"""Utility function for the Walsh Hadamard Transform"""
if not iterable(seq):
raise TypeError("Expected a sequence of coefficients "
"for Walsh Hadamard Transform")
a = [sympify(arg) for arg in seq]
n = len(a)
if n < 2:
return a
if n&(n - 1):
n = 2**n.bit_length()
a += [S.Zero]*(n - len(a))
h = 2
while h <= n:
hf, ut = h // 2, n // h
for i in range(0, n, h):
for j in range(hf):
u, v = a[i + j], a[i + j + hf]
a[i + j], a[i + j + hf] = u + v, u - v
h *= 2
if inverse:
a = [x/n for x in a]
return a
def _preprocess(self, args, expr):
"""Preprocess args, expr to replace arguments that do not map
to valid Python identifiers.
Returns string form of args, and updated expr.
"""
from sympy import Dummy, Function, flatten, Derivative, ordered, Basic
from sympy.matrices import DeferredVector
# Args of type Dummy can cause name collisions with args
# of type Symbol. Force dummify of everything in this
# situation.
dummify = self._dummify or any(
isinstance(arg, Dummy) for arg in flatten(args))
argstrs = [None]*len(args)
for arg, i in reversed(list(ordered(zip(args, range(len(args)))))):
if iterable(arg):
s, expr = self._preprocess(arg, expr)
elif isinstance(arg, DeferredVector):
s = str(arg)
elif isinstance(arg, Basic) and arg.is_symbol:
s = self._argrepr(arg)
if dummify or not self._is_safe_ident(s):
dummy = Dummy()
s = self._argrepr(dummy)
expr = self._subexpr(expr, {arg: dummy})
elif dummify or isinstance(arg, (Function, Derivative)):
dummy = Dummy()
s = self._argrepr(dummy)
expr = self._subexpr(expr, {arg: dummy})
else:
s = str(arg)
argstrs[i] = s
return argstrs, expr
def convert(self, elem, M=None):
"""
Convert ``elem`` into the internal representation.
This method is called implicitly whenever computations involve elements
not in the internal representation.
>>> from sympy.abc import x
>>> from sympy import QQ
>>> F = QQ.old_poly_ring(x).free_module(2)
>>> F.convert([1, 0])
[1, 0]
"""
if isinstance(elem, FreeModuleElement):
if elem.module is self:
return elem
if elem.module.rank != self.rank:
raise CoercionFailed
return FreeModuleElement(self,
tuple(self.ring.convert(x, elem.module.ring) for x in elem.data))
elif iterable(elem):
tpl = tuple(self.ring.convert(x) for x in elem)
if len(tpl) != self.rank:
raise CoercionFailed
return FreeModuleElement(self, tpl)
elif elem is 0:
return FreeModuleElement(self, (self.ring.convert(0),)*self.rank)
else:
raise CoercionFailed
def is_sequence(i, include=None):
"""
Return a boolean indicating whether ``i`` is a sequence in the SymPy
sense. If anything that fails the test below should be included as
being a sequence for your application, set 'include' to that object's
type; multiple types should be passed as a tuple of types.
Note: although generators can generate a sequence, they often need special
handling to make sure their elements are captured before the generator is
exhausted, so these are not included by default in the definition of a
sequence.
See also: iterable
Examples
========
>>> from sympy.utilities.iterables import is_sequence
>>> from types import GeneratorType
>>> is_sequence([])
True
>>> is_sequence(set())
False
>>> is_sequence('abc')
False
>>> is_sequence('abc', include=str)
True
>>> generator = (c for c in 'abc')
>>> is_sequence(generator)
False
>>> is_sequence(generator, include=(str, GeneratorType))
True
"""
return hasattr(i, "__getitem__") and iterable(i) or bool(include) and isinstance(i, include)
def sub_args(args, dummies_dict):
if isinstance(args, str):
return args
elif isinstance(args, DeferredVector):
return str(args)
elif iterable(args):
flatten = lambda *n: (e for a in n for e in
(flatten(*a) if iterable(a) else (a,)))
dummies = flatten([sub_args(a, dummies_dict) for a in args])
return ",".join(str(a) for a in dummies)
else:
if isinstance(args, Function):
dummies = Dummy()
dummies_dict.update({args : dummies})
return str(dummies)
else:
return str(args)
def flatten(arg):
if isinstance(arg, Set):
if arg.is_ProductSet:
return sum(map(flatten, arg.args), [])
else:
return [arg]
elif iterable(arg):
return sum(map(flatten, arg), [])
raise TypeError("Input must be Sets or iterables of Sets")
def flatten(arg):
if isinstance(arg, Set):
if arg.is_Intersection:
return sum(map(flatten, arg.args), [])
else:
return [arg]
if iterable(arg): # and not isinstance(arg, Set) (implicit)
return sum(map(flatten, arg), [])
raise TypeError("Input must be Sets or iterables of Sets")
def group_to_finiteset(group):
"""
Converts ``group`` to a :class:``FiniteSet`` if it is an
iterable.
"""
if iterable(group):
return FiniteSet(group)
else:
return group
def _flatten(arg):
if isinstance(arg, SeqBase):
if isinstance(arg, SeqMul):
return sum(map(_flatten, arg.args), [])
else:
return [arg]
elif iterable(arg):
return sum(map(_flatten, arg), [])
raise TypeError("Input must be Sequences or " " iterables of Sequences")
def _rebuild(expr):
if not isinstance(expr, Basic):
return expr
if not expr.args:
return expr
if iterable(expr):
new_args = [_rebuild(arg) for arg in expr]
return expr.func(*new_args)
if expr in subs:
return subs[expr]
orig_expr = expr
if expr in opt_subs:
expr = opt_subs[expr]
# If enabled, parse Muls and Adds arguments by order to ensure
# replacement order independent from hashes
if order != 'none':
if isinstance(expr, (Mul, MatMul)):
c, nc = expr.args_cnc()
if c == [1]:
args = nc
else:
args = list(ordered(c)) + nc
elif isinstance(expr, (Add, MatAdd)):
args = list(ordered(expr.args))
else:
args = expr.args
else:
args = expr.args
new_args = list(map(_rebuild, args))
if new_args != args:
new_expr = expr.func(*new_args)
else:
new_expr = expr
if orig_expr in to_eliminate:
try:
sym = next(symbols)
except StopIteration:
raise ValueError("Symbols iterator ran out of symbols.")
if isinstance(orig_expr, MatrixExpr):
sym = MatrixSymbol(sym.name, orig_expr.rows,
orig_expr.cols)
subs[orig_expr] = sym
replacements.append((sym, new_expr))
return sym
else:
return new_expr
def _number_theoretic_transform(seq, prime, inverse=False):
"""Utility function for the Number Theoretic Transform"""
if not iterable(seq):
raise TypeError("Expected a sequence of integer coefficients "
"for Number Theoretic Transform")
p = as_int(prime)
if isprime(p) == False:
raise ValueError("Expected prime modulus for "
"Number Theoretic Transform")
a = [as_int(x) % p for x in seq]
n = len(a)
if n < 1:
return a
b = n.bit_length() - 1
if n&(n - 1):
b += 1
n = 2**b
if (p - 1) % n:
raise ValueError("Expected prime modulus of the form (m*2**k + 1)")
a += [0]*(n - len(a))
for i in range(1, n):
j = int(ibin(i, b, str=True)[::-1], 2)
if i < j:
a[i], a[j] = a[j], a[i]
pr = primitive_root(p)
rt = pow(pr, (p - 1) // n, p)
if inverse:
rt = pow(rt, p - 2, p)
w = [1]*(n // 2)
for i in range(1, n // 2):
w[i] = w[i - 1]*rt % p
h = 2
while h <= n:
hf, ut = h // 2, n // h
for i in range(0, n, h):
for j in range(hf):
u, v = a[i + j], a[i + j + hf]*w[ut * j]
a[i + j], a[i + j + hf] = (u + v) % p, (u - v) % p
h *= 2
if inverse:
rv = pow(n, p - 2, p)
a = [x*rv % p for x in a]
return a
def posify(eq):
"""Return eq (with generic symbols made positive) and a
dictionary containing the mapping between the old and new
symbols.
Any symbol that has positive=None will be replaced with a positive dummy
symbol having the same name. This replacement will allow more symbolic
processing of expressions, especially those involving powers and
logarithms.
A dictionary that can be sent to subs to restore eq to its original
symbols is also returned.
>>> from sympy import posify, Symbol, log, solve
>>> from sympy.abc import x
>>> posify(x + Symbol('p', positive=True) + Symbol('n', negative=True))
(_x + n + p, {_x: x})
>>> eq = 1/x
>>> log(eq).expand()
log(1/x)
>>> log(posify(eq)[0]).expand()
-log(_x)
>>> p, rep = posify(eq)
>>> log(p).expand().subs(rep)
-log(x)
It is possible to apply the same transformations to an iterable
of expressions:
>>> eq = x**2 - 4
>>> solve(eq, x)
[-2, 2]
>>> eq_x, reps = posify([eq, x]); eq_x
[_x**2 - 4, _x]
>>> solve(*eq_x)
[2]
"""
eq = sympify(eq)
if iterable(eq):
f = type(eq)
eq = list(eq)
syms = set()
for e in eq:
syms = syms.union(e.atoms(Symbol))
reps = {}
for s in syms:
reps.update(dict((v, k) for k, v in posify(s)[1].items()))
for i, e in enumerate(eq):
eq[i] = e.subs(reps)
return f(eq), dict([(r, s) for s, r in reps.items()])
reps = dict([(s, Dummy(s.name, positive=True))
for s in eq.free_symbols if s.is_positive is None])
eq = eq.subs(reps)
return eq, dict([(r, s) for s, r in reps.items()])
def __init__(self, monom, gens=None):
if not iterable(monom):
rep, gens = dict_from_expr(sympify(monom), gens=gens)
if len(rep) == 1 and list(rep.values())[0] == 1:
monom = list(rep.keys())[0]
else:
raise ValueError("Expected a monomial got %s" % monom)
self.exponents = tuple(map(int, monom))
self.gens = gens
def __new__(cls, *args, **kwargs):
if iterable(args[0]):
coords = tuple([sympify(x) for x in args[0]])
else:
coords = tuple([sympify(x) for x in args])
if len(coords) != 2:
raise NotImplementedError("Only two dimensional points currently supported")
return GeometryEntity.__new__(cls, *coords)
def rewrite(self, *args, **hints):
""" Rewrite functions in terms of other functions.
Rewrites expression containing applications of functions
of one kind in terms of functions of different kind. For
example you can rewrite trigonometric functions as complex
exponentials or combinatorial functions as gamma function.
As a pattern this function accepts a list of functions to
to rewrite (instances of DefinedFunction class). As rule
you can use string or a destination function instance (in
this case rewrite() will use the str() function).
There is also possibility to pass hints on how to rewrite
the given expressions. For now there is only one such hint
defined called 'deep'. When 'deep' is set to False it will
forbid functions to rewrite their contents.
>>> from sympy import sin, exp, I
>>> from sympy.abc import x, y
Unspecified pattern:
>>> sin(x).rewrite(exp)
-I*(exp(I*x) - exp(-I*x))/2
Pattern as a single function:
>>> sin(x).rewrite(sin, exp)
-I*(exp(I*x) - exp(-I*x))/2
Pattern as a list of functions:
>>> sin(x).rewrite([sin, ], exp)
-I*(exp(I*x) - exp(-I*x))/2
"""
if self.is_Atom or not args:
return self
else:
pattern = args[:-1]
if isinstance(args[-1], basestring):
rule = '_eval_rewrite_as_' + args[-1]
else:
rule = '_eval_rewrite_as_' + args[-1].__name__
if not pattern:
return self._eval_rewrite(None, rule, **hints)
else:
if iterable(pattern[0]):
pattern = pattern[0]
pattern = [ p.__class__ for p in pattern if self.has(p) ]
if pattern:
return self._eval_rewrite(tuple(pattern), rule, **hints)
else:
return self
def flat_indexes(elems):
n = 0
for el in elems:
if iterable(el):
for ndeep in flat_indexes(el):
yield (n,) + ndeep
else:
yield (n,)
n += 1
def __new__(cls, *args, **kwargs):
if iterable(args[0]):
coords = Tuple(*args[0])
elif isinstance(args[0], Point):
coords = args[0].args
else:
coords = Tuple(*args)
if len(coords) != 2:
raise NotImplementedError("Only two dimensional points currently supported")
return GeometryEntity.__new__(cls, *coords)
def __new__(cls, *args, **kw_args):
"""The constructor for the Prufer object.
Examples
========
>>> from sympy.combinatorics.prufer import Prufer
A Prufer object can be constructed from a list of edges:
>>> a = Prufer([[0, 1], [0, 2], [0, 3]])
>>> a.prufer_repr
[0, 0]
If the number of nodes is given, no checking of the nodes will
be performed; it will be assumed that nodes 0 through n - 1 are
present:
>>> Prufer([[0, 1], [0, 2], [0, 3]], 4)
Prufer([[0, 1], [0, 2], [0, 3]], 4)
A Prufer object can be constructed from a Prufer sequence:
>>> b = Prufer([1, 3])
>>> b.tree_repr
[[0, 1], [1, 3], [2, 3]]
"""
ret_obj = Basic.__new__(cls, *args, **kw_args)
args = [list(args[0])]
if args[0] and iterable(args[0][0]):
if not args[0][0]:
raise ValueError(
'Prufer expects at least one edge in the tree.')
if len(args) > 1:
nnodes = args[1]
else:
nodes = set(flatten(args[0]))
nnodes = max(nodes) + 1
if nnodes != len(nodes):
missing = set(range(nnodes)) - nodes
if len(missing) == 1:
msg = 'Node %s is missing.' % missing.pop()
else:
msg = 'Nodes %s are missing.' % list(sorted(missing))
raise ValueError(msg)
ret_obj._tree_repr = [list(i) for i in args[0]]
ret_obj._nodes = nnodes
else:
ret_obj._prufer_repr = args[0]
ret_obj._nodes = len(ret_obj._prufer_repr) + 2
return ret_obj
def __new__(cls, *args):
if len(args) == 1 and iterable(args[0]):
args = args[0]
args = map(sympify, args)
if len(args) == 0:
return EmptySet()
args = frozenset(args) # remove duplicates
obj = Basic.__new__(cls, *args)
obj._elements = args
return obj
def __new__(cls, *args, **kwargs):
eval = kwargs.get('evaluate', global_evaluate[0])
check = True
if isinstance(args[0], Point2D):
if not eval:
return args[0]
args = args[0].args
check = False
else:
if iterable(args[0]):
args = args[0]
if len(args) != 2:
raise ValueError(
"Only two dimensional points currently supported")
coords = Tuple(*args)
if check:
if any(a.is_number and im(a) for a in coords):
raise ValueError('Imaginary args not permitted.')
if eval:
coords = coords.xreplace(
dict([(f, simplify(nsimplify(f, rational=True)))
for f in coords.atoms(Float)]))
return GeometryEntity.__new__(cls, *coords)
def _find_repeated(expr):
if not isinstance(expr, Basic):
return
if expr.is_Atom or expr.is_Order:
return
if iterable(expr):
args = expr
else:
if expr in seen_subexp:
to_eliminate.add(expr)
return
seen_subexp.add(expr)
if expr in opt_subs:
expr = opt_subs[expr]
args = expr.args
list(map(_find_repeated, args))
def _find_opts(expr):
if not isinstance(expr, Basic):
return
if expr.is_Atom or expr.is_Order:
return
if iterable(expr):
list(map(_find_opts, expr))
return
if expr in seen_subexp:
return expr
seen_subexp.add(expr)
list(map(_find_opts, expr.args))
if _coeff_isneg(expr):
neg_expr = -expr
if not neg_expr.is_Atom:
opt_subs[expr] = Mul(S.NegativeOne, neg_expr, evaluate=False)
seen_subexp.add(neg_expr)
expr = neg_expr
if isinstance(expr, (Mul, MatMul)):
muls.add(expr)
elif isinstance(expr, (Add, MatAdd)):
adds.add(expr)
elif isinstance(expr, (Pow, MatPow)):
if _coeff_isneg(expr.exp):
opt_subs[expr] = Pow(Pow(expr.base, -expr.exp),
S.NegativeOne,
evaluate=False)
def postorder_traversal(node):
"""
Do a postorder traversal of a tree.
This generator recursively yields nodes that it has visited in a postorder
fashion. That is, it descends through the tree depth-first to yield all of
a node's children's postorder traversal before yielding the node itself.
Parameters
----------
node : sympy expression
The expression to traverse.
Yields
------
subtree : sympy expression
All of the subtrees in the tree.
Examples
--------
>>> from sympy import symbols
>>> from sympy.utilities.iterables import postorder_traversal
>>> from sympy.abc import x, y, z
>>> set(postorder_traversal((x+y)*z)) == set([z, y, x, x + y, z*(x + y)])
True
"""
if isinstance(node, Basic):
for arg in node.args:
for subtree in postorder_traversal(arg):
yield subtree
elif iterable(node):
for item in node:
for subtree in postorder_traversal(item):
yield subtree
yield node
def _mobius_transform(seq, sgn, subset):
r"""Utility function for performing Möbius Transform using
Yate's Dynamic Programming (DP)"""
if not iterable(seq):
raise TypeError("Expected a sequence of coefficients")
a = [sympify(arg) for arg in seq]
n = len(a)
if n < 2:
return a
if n & (n - 1):
n = 2**n.bit_length()
a += [S.Zero] * (n - len(a))
if subset:
i = 1
while i < n:
for j in range(n):
if j & i:
a[j] += sgn * a[j ^ i]
i *= 2
else:
i = 1
while i < n:
for j in range(n):
if j & i:
continue
a[j] += sgn * a[j ^ i]
i *= 2
return a
def is_sequence(i, include=None):
"""
Return a boolean indicating whether ``i`` is a sequence in the SymPy
sense. If anything that fails the test below should be included as
being a sequence for your application, set 'include' to that object's
type; multiple types should be passed as a tuple of types.
Note: although generators can generate a sequence, they often need special
handling to make sure their elements are captured before the generator is
exhausted, so these are not included by default in the definition of a
sequence.
See also: iterable
Examples
========
>>> from sympy.utilities.iterables import is_sequence
>>> from types import GeneratorType
>>> is_sequence([])
True
>>> is_sequence(set())
False
>>> is_sequence('abc')
False
>>> is_sequence('abc', include=str)
True
>>> generator = (c for c in 'abc')
>>> is_sequence(generator)
False
>>> is_sequence(generator, include=(str, GeneratorType))
True
"""
return (hasattr(i, '__getitem__') and iterable(i)
or bool(include) and isinstance(i, include))
def PolynomialRing(dom, *gens, **opts):
r"""
Create a generalized multivariate polynomial ring.
A generalized polynomial ring is defined by a ground field `K`, a set
of generators (typically `x_1, \dots, x_n`) and a monomial order `<`.
The monomial order can be global, local or mixed. In any case it induces
a total ordering on the monomials, and there exists for every (non-zero)
polynomial `f \in K[x_1, \dots, x_n]` a well-defined "leading monomial"
`LM(f) = LM(f, >)`. One can then define a multiplicative subset
`S = S_> = \{f \in K[x_1, \dots, x_n] | LM(f) = 1\}`. The generalized
polynomial ring corresponding to the monomial order is
`R = S^{-1}K[x_1, \dots, x_n]`.
If `>` is a so-called global order, that is `1` is the smallest monomial,
then we just have `S = K` and `R = K[x_1, \dots, x_n]`.
Examples
========
A few examples may make this clearer.
>>> from sympy.abc import x, y
>>> from sympy import QQ
Our first ring uses global lexicographic order.
>>> R1 = QQ.old_poly_ring(x, y, order=(("lex", x, y),))
The second ring uses local lexicographic order. Note that when using a
single (non-product) order, you can just specify the name and omit the
variables:
>>> R2 = QQ.old_poly_ring(x, y, order="ilex")
The third and fourth rings use a mixed orders:
>>> o1 = (("ilex", x), ("lex", y))
>>> o2 = (("lex", x), ("ilex", y))
>>> R3 = QQ.old_poly_ring(x, y, order=o1)
>>> R4 = QQ.old_poly_ring(x, y, order=o2)
We will investigate what elements of `K(x, y)` are contained in the various
rings.
>>> L = [x, 1/x, y/(1 + x), 1/(1 + y), 1/(1 + x*y)]
>>> test = lambda R: [f in R for f in L]
The first ring is just `K[x, y]`:
>>> test(R1)
[True, False, False, False, False]
The second ring is R1 localised at the maximal ideal (x, y):
>>> test(R2)
[True, False, True, True, True]
The third ring is R1 localised at the prime ideal (x):
>>> test(R3)
[True, False, True, False, True]
Finally the fourth ring is R1 localised at `S = K[x, y] \setminus yK[y]`:
>>> test(R4)
[True, False, False, True, False]
"""
order = opts.get("order", GeneralizedPolynomialRing.default_order)
if iterable(order):
order = build_product_order(order, gens)
order = monomial_key(order)
opts['order'] = order
if order.is_global:
return GlobalPolynomialRing(dom, *gens, **opts)
else:
return GeneralizedPolynomialRing(dom, *gens, **opts)
def isiter(l):
return iterable(l, exclude=(str, DeferredVector))
def unpack_lhs(lvalues):
return '[{}]'.format(', '.join(
unpack_lhs(val) if iterable(val) else val for val in lvalues))
def __new__(cls, *args):
"""
Construct a new instance of Diagram.
If no arguments are supplied, an empty diagram is created.
If at least an argument is supplied, ``args[0]`` is
interpreted as the premises of the diagram. If ``args[0]`` is
a list, it is interpreted as a list of :class:`Morphism`'s, in
which each :class:`Morphism` has an empty set of properties.
If ``args[0]`` is a Python dictionary or a :class:`Dict`, it
is interpreted as a dictionary associating to some
:class:`Morphism`'s some properties.
If at least two arguments are supplied ``args[1]`` is
interpreted as the conclusions of the diagram. The type of
``args[1]`` is interpreted in exactly the same way as the type
of ``args[0]``. If only one argument is supplied, the diagram
has no conclusions.
Examples
========
>>> from sympy.categories import Object, NamedMorphism
>>> from sympy.categories import IdentityMorphism, Diagram
>>> A = Object("A")
>>> B = Object("B")
>>> C = Object("C")
>>> f = NamedMorphism(A, B, "f")
>>> g = NamedMorphism(B, C, "g")
>>> d = Diagram([f, g])
>>> IdentityMorphism(A) in d.premises.keys()
True
>>> g * f in d.premises.keys()
True
>>> d = Diagram([f, g], {g * f: "unique"})
>>> d.conclusions[g * f]
{unique}
"""
premises = {}
conclusions = {}
# Here we will keep track of the objects which appear in the
# premises.
objects = EmptySet()
if len(args) >= 1:
# We've got some premises in the arguments.
premises_arg = args[0]
if isinstance(premises_arg, list):
# The user has supplied a list of morphisms, none of
# which have any attributes.
empty = EmptySet()
for morphism in premises_arg:
objects |= FiniteSet(morphism.domain, morphism.codomain)
Diagram._add_morphism_closure(premises, morphism, empty)
elif isinstance(premises_arg, dict) or isinstance(
premises_arg, Dict):
# The user has supplied a dictionary of morphisms and
# their properties.
for morphism, props in premises_arg.items():
objects |= FiniteSet(morphism.domain, morphism.codomain)
Diagram._add_morphism_closure(
premises, morphism,
FiniteSet(
*props) if iterable(props) else FiniteSet(props))
if len(args) >= 2:
# We also have some conclusions.
conclusions_arg = args[1]
if isinstance(conclusions_arg, list):
# The user has supplied a list of morphisms, none of
# which have any attributes.
empty = EmptySet()
for morphism in conclusions_arg:
# Check that no new objects appear in conclusions.
if (morphism.domain in objects) and \
(morphism.codomain in objects):
# No need to add identities and recurse
# composites this time.
Diagram._add_morphism_closure(conclusions,
morphism,
empty,
add_identities=False,
recurse_composites=False)
elif isinstance(conclusions_arg, dict) or \
isinstance(conclusions_arg, Dict):
# The user has supplied a dictionary of morphisms and
# their properties.
for morphism, props in conclusions_arg.items():
# Check that no new objects appear in conclusions.
if (morphism.domain in objects) and \
(morphism.codomain in objects):
# No need to add identities and recurse
# composites this time.
Diagram._add_morphism_closure(
conclusions,
morphism,
FiniteSet(*props)
if iterable(props) else FiniteSet(props),
add_identities=False,
recurse_composites=False)
return Basic.__new__(cls, Dict(premises), Dict(conclusions), objects)
def lambdastr(args, expr, printer=None, dummify=False):
"""
Returns a string that can be evaluated to a lambda function.
Examples
========
>>> from sympy.abc import x, y, z
>>> from sympy.utilities.lambdify import lambdastr
>>> lambdastr(x, x**2)
'lambda x: (x**2)'
>>> lambdastr((x,y,z), [z,y,x])
'lambda x,y,z: ([z, y, x])'
Although tuples may not appear as arguments to lambda in Python 3,
lambdastr will create a lambda function that will unpack the original
arguments so that nested arguments can be handled:
>>> lambdastr((x, (y, z)), x + y)
'lambda _0,_1: (lambda x,y,z: (x + y))(*list(__flatten_args__([_0,_1])))'
"""
# Transforming everything to strings.
from sympy.matrices import DeferredVector
from sympy import Dummy, sympify, Symbol, Function, flatten
if printer is not None:
if inspect.isfunction(printer):
lambdarepr = printer
else:
if inspect.isclass(printer):
lambdarepr = lambda expr: printer().doprint(expr)
else:
lambdarepr = lambda expr: printer.doprint(expr)
else:
#XXX: This has to be done here because of circular imports
from sympy.printing.lambdarepr import lambdarepr
def sub_args(args, dummies_dict):
if isinstance(args, str):
return args
elif isinstance(args, DeferredVector):
return str(args)
elif iterable(args):
dummies = flatten([sub_args(a, dummies_dict) for a in args])
return ",".join(str(a) for a in dummies)
else:
#Sub in dummy variables for functions or symbols
if isinstance(args, (Function, Symbol)):
dummies = Dummy()
dummies_dict.update({args: dummies})
return str(dummies)
else:
return str(args)
def sub_expr(expr, dummies_dict):
try:
expr = sympify(expr).xreplace(dummies_dict)
except Exception:
if isinstance(expr, DeferredVector):
pass
elif isinstance(expr, dict):
k = [sub_expr(sympify(a), dummies_dict) for a in expr.keys()]
v = [sub_expr(sympify(a), dummies_dict) for a in expr.values()]
expr = dict(zip(k, v))
elif isinstance(expr, tuple):
expr = tuple(sub_expr(sympify(a), dummies_dict) for a in expr)
elif isinstance(expr, list):
expr = [sub_expr(sympify(a), dummies_dict) for a in expr]
return expr
# Transform args
def isiter(l):
return iterable(l, exclude=(str, DeferredVector))
if isiter(args) and any(isiter(i) for i in args):
from sympy.utilities.iterables import flatten
import re
dum_args = [str(Dummy(str(i))) for i in range(len(args))]
iter_args = ','.join(
[i if isiter(a) else i for i, a in zip(dum_args, args)])
lstr = lambdastr(flatten(args), expr, printer=printer, dummify=dummify)
flat = '__flatten_args__'
rv = 'lambda %s: (%s)(*list(%s([%s])))' % (','.join(dum_args), lstr,
flat, iter_args)
if len(re.findall(r'\b%s\b' % flat, rv)) > 1:
raise ValueError('the name %s is reserved by lambdastr' % flat)
return rv
dummies_dict = {}
if dummify:
args = sub_args(args, dummies_dict)
else:
if isinstance(args, str):
pass
elif iterable(args, exclude=DeferredVector):
args = ",".join(str(a) for a in args)
# Transform expr
if dummify:
if isinstance(expr, str):
pass
else:
expr = sub_expr(expr, dummies_dict)
expr = lambdarepr(expr)
return "lambda %s: (%s)" % (args, expr)
def inv(l):
if iterable(l):
return tuple(inv(x) for x in l)
return -l
def autowrap(expr,
language=None,
backend='Cython',
tempdir=None,
args=None,
flags=None,
verbose=False,
helpers=None):
"""Generates python callable binaries based on the math expression.
Parameters
----------
expr
The SymPy expression that should be wrapped as a binary routine.
language : string, optional
If supplied, (options: 'C' or 'F95'), specifies the language of the
generated code. If ``None`` [default], the language is inferred based
upon the specified backend.
backend : string, optional
Backend used to wrap the generated code. Either 'f2py',
or 'cython' [default].
tempdir : string, optional
Path to directory for temporary files. If this argument is supplied,
the generated code and the wrapper input files are left intact in the
specified path.
args : iterable, optional
An ordered iterable of symbols. Specifies the argument sequence for the
function.
flags : iterable, optional
Additional option flags that will be passed to the backend.
verbose : bool, optional
If True, autowrap will not mute the command line backends. This can be
helpful for debugging.
helpers : iterable, optional
Used to define auxillary expressions needed for the main expr. If the
main expression needs to call a specialized function it should be put
in the ``helpers`` iterable. Autowrap will then make sure that the
compiled main expression can link to the helper routine. Items should
be tuples with (<funtion_name>, <sympy_expression>, <arguments>). It
is mandatory to supply an argument sequence to helper routines.
>>> from sympy.abc import x, y, z
>>> from sympy.utilities.autowrap import autowrap
>>> expr = ((x - y + z)**(13)).expand()
>>> binary_func = autowrap(expr)
>>> binary_func(1, 4, 2)
-1.0
"""
if language:
_validate_backend_language(backend, language)
else:
language = _infer_language(backend)
helpers = [helpers] if helpers else ()
flags = flags if flags else ()
args = list(args) if iterable(args, exclude=set) else args
code_generator = get_code_generator(language, "autowrap")
CodeWrapperClass = _get_code_wrapper_class(backend)
code_wrapper = CodeWrapperClass(code_generator, tempdir, flags, verbose)
helps = []
for name_h, expr_h, args_h in helpers:
helps.append(make_routine(name_h, expr_h, args_h))
for name_h, expr_h, args_h in helpers:
if expr.has(expr_h):
name_h = binary_function(name_h, expr_h, backend='dummy')
expr = expr.subs(expr_h, name_h(*args_h))
try:
routine = make_routine('autofunc', expr, args)
except CodeGenArgumentListError as e:
# if all missing arguments are for pure output, we simply attach them
# at the end and try again, because the wrappers will silently convert
# them to return values anyway.
new_args = []
for missing in e.missing_args:
if not isinstance(missing, OutputArgument):
raise
new_args.append(missing.name)
routine = make_routine('autofunc', expr, args + new_args)
return code_wrapper.wrap_code(routine, helpers=helps)
def convolution_subset(a, b):
"""
Performs Subset Convolution of given sequences.
The indices of each argument, considered as bit strings, correspond to
subsets of a finite set.
The sequence is automatically padded to the right with zeros, as the
definition of subset based on bitmasks (indices) requires the size of
sequence to be a power of 2.
Parameters
==========
a, b : iterables
The sequences for which convolution is performed.
Examples
========
>>> from sympy import symbols, S, I
>>> from sympy.discrete.convolutions import convolution_subset
>>> u, v, x, y, z = symbols('u v x y z')
>>> convolution_subset([u, v], [x, y])
[u*x, u*y + v*x]
>>> convolution_subset([u, v, x], [y, z])
[u*y, u*z + v*y, x*y, x*z]
>>> convolution_subset([1, S(2)/3], [3, 4])
[3, 6]
>>> convolution_subset([1, 3, S(5)/7], [7])
[7, 21, 5, 0]
References
==========
.. [1] https://people.csail.mit.edu/rrw/presentations/subset-conv.pdf
"""
if not a or not b:
return []
if not iterable(a) or not iterable(b):
raise TypeError("Expected a sequence of coefficients for convolution")
a = [sympify(arg) for arg in a]
b = [sympify(arg) for arg in b]
n = max(len(a), len(b))
if n & (n - 1): # not a power of 2
n = 2**n.bit_length()
# padding with zeros
a += [S.Zero] * (n - len(a))
b += [S.Zero] * (n - len(b))
c = [S.Zero] * n
for mask in range(n):
smask = mask
while smask > 0:
c[mask] += expand_mul(a[smask] * b[mask ^ smask])
smask = (smask - 1) & mask
c[mask] += expand_mul(a[smask] * b[mask ^ smask])
return c
def default_sort_key(item, order=None):
"""Return a key that can be used for sorting.
The key has the structure:
(class_key, (len(args), args), exponent.sort_key(), coefficient)
This key is supplied by the sort_key routine of Basic objects when
``item`` is a Basic object or an object (other than a string) that
sympifies to a Basic object. Otherwise, this function produces the
key.
The ``order`` argument is passed along to the sort_key routine and is
used to determine how the terms *within* an expression are ordered.
(See examples below) ``order`` options are: 'lex', 'grlex', 'grevlex',
and reversed values of the same (e.g. 'rev-lex'). The default order
value is None (which translates to 'lex').
Examples
========
>>> from sympy import S, I, default_sort_key, sin, cos, sqrt
>>> from sympy.core.function import UndefinedFunction
>>> from sympy.abc import x
The following are equivalent ways of getting the key for an object:
>>> x.sort_key() == default_sort_key(x)
True
Here are some examples of the key that is produced:
>>> default_sort_key(UndefinedFunction('f'))
((0, 0, 'UndefinedFunction'), (1, ('f',)), ((1, 0, 'Number'),
(0, ()), (), 1), 1)
>>> default_sort_key('1')
((0, 0, 'str'), (1, ('1',)), ((1, 0, 'Number'), (0, ()), (), 1), 1)
>>> default_sort_key(S.One)
((1, 0, 'Number'), (0, ()), (), 1)
>>> default_sort_key(2)
((1, 0, 'Number'), (0, ()), (), 2)
While sort_key is a method only defined for SymPy objects,
default_sort_key will accept anything as an argument so it is
more robust as a sorting key. For the following, using key=
lambda i: i.sort_key() would fail because 2 doesn't have a sort_key
method; that's why default_sort_key is used. Note, that it also
handles sympification of non-string items likes ints:
>>> a = [2, I, -I]
>>> sorted(a, key=default_sort_key)
[2, -I, I]
The returned key can be used anywhere that a key can be specified for
a function, e.g. sort, min, max, etc...:
>>> a.sort(key=default_sort_key); a[0]
2
>>> min(a, key=default_sort_key)
2
Note
----
The key returned is useful for getting items into a canonical order
that will be the same across platforms. It is not directly useful for
sorting lists of expressions:
>>> a, b = x, 1/x
Since ``a`` has only 1 term, its value of sort_key is unaffected by
``order``:
>>> a.sort_key() == a.sort_key('rev-lex')
True
If ``a`` and ``b`` are combined then the key will differ because there
are terms that can be ordered:
>>> eq = a + b
>>> eq.sort_key() == eq.sort_key('rev-lex')
False
>>> eq.as_ordered_terms()
[x, 1/x]
>>> eq.as_ordered_terms('rev-lex')
[1/x, x]
But since the keys for each of these terms are independent of ``order``'s
value, they don't sort differently when they appear separately in a list:
>>> sorted(eq.args, key=default_sort_key)
[1/x, x]
>>> sorted(eq.args, key=lambda i: default_sort_key(i, order='rev-lex'))
[1/x, x]
The order of terms obtained when using these keys is the order that would
be obtained if those terms were *factors* in a product.
Although it is useful for quickly putting expressions in canonical order,
it does not sort expressions based on their complexity defined by the
number of operations, power of variables and others:
>>> sorted([sin(x)*cos(x), sin(x)], key=default_sort_key)
[sin(x)*cos(x), sin(x)]
>>> sorted([x, x**2, sqrt(x), x**3], key=default_sort_key)
[sqrt(x), x, x**2, x**3]
See Also
========
ordered, sympy.core.expr.as_ordered_factors, sympy.core.expr.as_ordered_terms
"""
from sympy.core import S, Basic
from sympy.core.sympify import sympify, SympifyError
from sympy.core.compatibility import iterable
if isinstance(item, Basic):
return item.sort_key(order=order)
if iterable(item, exclude=string_types):
if isinstance(item, dict):
args = item.items()
unordered = True
elif isinstance(item, set):
args = item
unordered = True
else:
# e.g. tuple, list
args = list(item)
unordered = False
args = [default_sort_key(arg, order=order) for arg in args]
if unordered:
# e.g. dict, set
args = sorted(args)
cls_index, args = 10, (len(args), tuple(args))
else:
if not isinstance(item, string_types):
try:
item = sympify(item)
except SympifyError:
# e.g. lambda x: x
pass
else:
if isinstance(item, Basic):
# e.g int -> Integer
return default_sort_key(item)
# e.g. UndefinedFunction
# e.g. str
cls_index, args = 0, (1, (str(item), ))
return (cls_index, 0,
item.__class__.__name__), args, S.One.sort_key(), S.One
def lambdify(args,
expr,
modules=None,
printer=None,
use_imps=True,
dummify=False):
"""
Translates a SymPy expression into an equivalent numeric function
For example, to convert the SymPy expression ``sin(x) + cos(x)`` to an
equivalent NumPy function that numerically evaluates it:
>>> from sympy import sin, cos, symbols, lambdify
>>> import numpy as np
>>> x = symbols('x')
>>> expr = sin(x) + cos(x)
>>> expr
sin(x) + cos(x)
>>> f = lambdify(x, expr, 'numpy')
>>> a = np.array([1, 2])
>>> f(a)
[1.38177329 0.49315059]
The primary purpose of this function is to provide a bridge from SymPy
expressions to numerical libraries such as NumPy, SciPy, NumExpr, mpmath,
and tensorflow. In general, SymPy functions do not work with objects from
other libraries, such as NumPy arrays, and functions from numeric
libraries like NumPy or mpmath do not work on SymPy expressions.
``lambdify`` bridges the two by converting a SymPy expression to an
equivalent numeric function.
The basic workflow with ``lambdify`` is to first create a SymPy expression
representing whatever mathematical function you wish to evaluate. This
should be done using only SymPy functions and expressions. Then, use
``lambdify`` to convert this to an equivalent function for numerical
evaluation. For instance, above we created ``expr`` using the SymPy symbol
``x`` and SymPy functions ``sin`` and ``cos``, then converted it to an
equivalent NumPy function ``f``, and called it on a NumPy array ``a``.
.. warning::
This function uses ``exec``, and thus shouldn't be used on unsanitized
input.
Arguments
=========
The first argument of ``lambdify`` is a variable or list of variables in
the expression. Variable lists may be nested. Variables can be Symbols,
undefined functions, or matrix symbols. The order and nesting of the
variables corresponds to the order and nesting of the parameters passed to
the lambdified function. For instance,
>>> from sympy.abc import x, y, z
>>> f = lambdify([x, (y, z)], x + y + z)
>>> f(1, (2, 3))
6
The second argument of ``lambdify`` is the expression, list of
expressions, or matrix to be evaluated. Lists may be nested. If the
expression is a list, the output will also be a list.
>>> f = lambdify(x, [x, [x + 1, x + 2]])
>>> f(1)
[1, [2, 3]]
If it is a matrix, an array will be returned (for the NumPy module).
>>> from sympy import Matrix
>>> f = lambdify(x, Matrix([x, x + 1]))
>>> f(1)
[[1]
[2]]
Note that the argument order here, variables then expression, is used to
emulate the Python ``lambda`` keyword. ``lambdify(x, expr)`` works
(roughly) like ``lambda x: expr`` (see :ref:`lambdify-how-it-works` below).
The third argument, ``modules`` is optional. If not specified, ``modules``
defaults to ``["scipy", "numpy"]`` if SciPy is installed, ``["numpy"]`` if
only NumPy is installed, and ``["math", "mpmath", "sympy"]`` if neither is
installed. That is, SymPy functions are replaced as far as possible by
either ``scipy`` or ``numpy`` functions if available, and Python's
standard library ``math``, or ``mpmath`` functions otherwise.
``modules`` can be one of the following types
- the strings ``"math"``, ``"mpmath"``, ``"numpy"``, ``"numexpr"``,
``"scipy"``, ``"sympy"``, or ``"tensorflow"``. This uses the
corresponding printer and namespace mapping for that module.
- a module (e.g., ``math``). This uses the global namespace of the
module. If the module is one of the above known modules, it will also
use the corresponding printer and namespace mapping (i.e.,
``modules=numpy`` is equivalent to ``modules="numpy"``).
- a dictionary that maps names of SymPy functions to arbitrary functions
(e.g., ``{'sin': custom_sin}``).
- a list that contains a mix of the arguments above, with higher priority
given to entries appearing first (e.g., to use the NumPy module but
override the ``sin`` function with a custom version, you can use
``[{'sin': custom_sin}, 'numpy']``).
The ``dummify`` keyword argument controls whether or not the variables in
the provided expression that are not valid Python identifiers are
substituted with dummy symbols. This allows for undefined functions like
``Function('f')(t)`` to be supplied as arguments. By default, the
variables are only dummified if they are not valid Python identifiers. Set
``dummify=True`` to replace all arguments with dummy symbols (if ``args``
is not a string) - for example, to ensure that the arguments do not
redefine any built-in names.
.. _lambdify-how-it-works:
How it works
============
When using this function, it helps a great deal to have an idea of what it
is doing. At its core, lambdify is nothing more than a namespace
translation, on top of a special printer that makes some corner cases work
properly.
To understand lambdify, first we must properly understand how Python
namespaces work. Say we had two files. One called ``sin_cos_sympy.py``,
with
.. code:: python
# sin_cos_sympy.py
from sympy import sin, cos
def sin_cos(x):
return sin(x) + cos(x)
and one called ``sin_cos_numpy.py`` with
.. code:: python
# sin_cos_numpy.py
from numpy import sin, cos
def sin_cos(x):
return sin(x) + cos(x)
The two files define an identical function ``sin_cos``. However, in the
first file, ``sin`` and ``cos`` are defined as the SymPy ``sin`` and
``cos``. In the second, they are defined as the NumPy versions.
If we were to import the first file and use the ``sin_cos`` function, we
would get something like
>>> from sin_cos_sympy import sin_cos # doctest: +SKIP
>>> sin_cos(1) # doctest: +SKIP
cos(1) + sin(1)
On the other hand, if we imported ``sin_cos`` from the second file, we
would get
>>> from sin_cos_numpy import sin_cos # doctest: +SKIP
>>> sin_cos(1) # doctest: +SKIP
1.38177329068
In the first case we got a symbolic output, because it used the symbolic
``sin`` and ``cos`` functions from SymPy. In the second, we got a numeric
result, because ``sin_cos`` used the numeric ``sin`` and ``cos`` functions
from NumPy. But notice that the versions of ``sin`` and ``cos`` that were
used was not inherent to the ``sin_cos`` function definition. Both
``sin_cos`` definitions are exactly the same. Rather, it was based on the
names defined at the module where the ``sin_cos`` function was defined.
The key point here is that when function in Python references a name that
is not defined in the function, that name is looked up in the "global"
namespace of the module where that function is defined.
Now, in Python, we can emulate this behavior without actually writing a
file to disk using the ``exec`` function. ``exec`` takes a string
containing a block of Python code, and a dictionary that should contain
the global variables of the module. It then executes the code "in" that
dictionary, as if it were the module globals. The following is equivalent
to the ``sin_cos`` defined in ``sin_cos_sympy.py``:
>>> import sympy
>>> module_dictionary = {'sin': sympy.sin, 'cos': sympy.cos}
>>> exec('''
... def sin_cos(x):
... return sin(x) + cos(x)
... ''', module_dictionary)
>>> sin_cos = module_dictionary['sin_cos']
>>> sin_cos(1)
cos(1) + sin(1)
and similarly with ``sin_cos_numpy``:
>>> import numpy
>>> module_dictionary = {'sin': numpy.sin, 'cos': numpy.cos}
>>> exec('''
... def sin_cos(x):
... return sin(x) + cos(x)
... ''', module_dictionary)
>>> sin_cos = module_dictionary['sin_cos']
>>> sin_cos(1)
1.38177329068
So now we can get an idea of how ``lambdify`` works. The name "lambdify"
comes from the fact that we can think of something like ``lambdify(x,
sin(x) + cos(x), 'numpy')`` as ``lambda x: sin(x) + cos(x)``, where
``sin`` and ``cos`` come from the ``numpy`` namespace. This is also why
the symbols argument is first in ``lambdify``, as opposed to most SymPy
functions where it comes after the expression: to better mimic the
``lambda`` keyword.
``lambdify`` takes the input expression (like ``sin(x) + cos(x)``) and
1. Converts it to a string
2. Creates a module globals dictionary based on the modules that are
passed in (by default, it uses the NumPy module)
3. Creates the string ``"def func({vars}): return {expr}"``, where ``{vars}`` is the
list of variables separated by commas, and ``{expr}`` is the string
created in step 1., then ``exec``s that string with the module globals
namespace and returns ``func``.
In fact, functions returned by ``lambdify`` support inspection. So you can
see exactly how they are defined by using ``inspect.getsource``, or ``??`` if you
are using IPython or the Jupyter notebook.
>>> f = lambdify(x, sin(x) + cos(x))
>>> import inspect
>>> print(inspect.getsource(f))
def _lambdifygenerated(x):
return (sin(x) + cos(x))
This shows us the source code of the function, but not the namespace it
was defined in. We can inspect that by looking at the ``__globals__``
attribute of ``f``:
>>> f.__globals__['sin']
<ufunc 'sin'>
>>> f.__globals__['cos']
<ufunc 'cos'>
>>> f.__globals__['sin'] is numpy.sin
True
This shows us that ``sin`` and ``cos`` in the namespace of ``f`` will be
``numpy.sin`` and ``numpy.cos``.
Note that there are some convenience layers in each of these steps, but at
the core, this is how ``lambdify`` works. Step 1 is done using the
``LambdaPrinter`` printers defined in the printing module (see
:mod:`sympy.printing.lambdarepr`). This allows different SymPy expressions
to define how they should be converted to a string for different modules.
You can change which printer ``lambdify`` uses by passing a custom printer
in to the ``printer`` argument.
Step 2 is augmented by certain translations. There are default
translations for each module, but you can provide your own by passing a
list to the ``modules`` argument. For instance,
>>> def mysin(x):
... print('taking the sin of', x)
... return numpy.sin(x)
...
>>> f = lambdify(x, sin(x), [{'sin': mysin}, 'numpy'])
>>> f(1)
taking the sin of 1
0.8414709848078965
The globals dictionary is generated from the list by merging the
dictionary ``{'sin': mysin}`` and the module dictionary for NumPy. The
merging is done so that earlier items take precedence, which is why
``mysin`` is used above instead of ``numpy.sin``.
If you want to modify the way ``lambdify`` works for a given function, it
is usually easiest to do so by modifying the globals dictionary as such.
In more complicated cases, it may be necessary to create and pass in a
custom printer.
Finally, step 3 is augmented with certain convenience operations, such as
the addition of a docstring.
Understanding how ``lambdify`` works can make it easier to avoid certain
gotchas when using it. For instance, a common mistake is to create a
lambdified function for one module (say, NumPy), and pass it objects from
another (say, a SymPy expression).
For instance, say we create
>>> from sympy.abc import x
>>> f = lambdify(x, x + 1, 'numpy')
Now if we pass in a NumPy array, we get that array plus 1
>>> import numpy
>>> a = numpy.array([1, 2])
>>> f(a)
[2 3]
But what happens if you make the mistake of passing in a SymPy expression
instead of a NumPy array:
>>> f(x + 1)
x + 2
This worked, but it was only by accident. Now take a different lambdified
function:
>>> from sympy import sin
>>> g = lambdify(x, x + sin(x), 'numpy')
This works as expected on NumPy arrays:
>>> g(a)
[1.84147098 2.90929743]
But if we try to pass in a SymPy expression, it fails
>>> try:
... g(x + 1)
... # NumPy release after 1.17 raises TypeError instead of
... # AttributeError
... except (AttributeError, TypeError):
... raise AttributeError() # doctest: +IGNORE_EXCEPTION_DETAIL
Traceback (most recent call last):
...
AttributeError:
Now, let's look at what happened. The reason this fails is that ``g``
calls ``numpy.sin`` on the input expression, and ``numpy.sin`` does not
know how to operate on a SymPy object. **As a general rule, NumPy
functions do not know how to operate on SymPy expressions, and SymPy
functions do not know how to operate on NumPy arrays. This is why lambdify
exists: to provide a bridge between SymPy and NumPy.**
However, why is it that ``f`` did work? That's because ``f`` doesn't call
any functions, it only adds 1. So the resulting function that is created,
``def _lambdifygenerated(x): return x + 1`` does not depend on the globals
namespace it is defined in. Thus it works, but only by accident. A future
version of ``lambdify`` may remove this behavior.
Be aware that certain implementation details described here may change in
future versions of SymPy. The API of passing in custom modules and
printers will not change, but the details of how a lambda function is
created may change. However, the basic idea will remain the same, and
understanding it will be helpful to understanding the behavior of
lambdify.
**In general: you should create lambdified functions for one module (say,
NumPy), and only pass it input types that are compatible with that module
(say, NumPy arrays).** Remember that by default, if the ``module``
argument is not provided, ``lambdify`` creates functions using the NumPy
and SciPy namespaces.
Examples
========
>>> from sympy.utilities.lambdify import implemented_function
>>> from sympy import sqrt, sin, Matrix
>>> from sympy import Function
>>> from sympy.abc import w, x, y, z
>>> f = lambdify(x, x**2)
>>> f(2)
4
>>> f = lambdify((x, y, z), [z, y, x])
>>> f(1,2,3)
[3, 2, 1]
>>> f = lambdify(x, sqrt(x))
>>> f(4)
2.0
>>> f = lambdify((x, y), sin(x*y)**2)
>>> f(0, 5)
0.0
>>> row = lambdify((x, y), Matrix((x, x + y)).T, modules='sympy')
>>> row(1, 2)
Matrix([[1, 3]])
``lambdify`` can be used to translate SymPy expressions into mpmath
functions. This may be preferable to using ``evalf`` (which uses mpmath on
the backend) in some cases.
>>> import mpmath
>>> f = lambdify(x, sin(x), 'mpmath')
>>> f(1)
0.8414709848078965
Tuple arguments are handled and the lambdified function should
be called with the same type of arguments as were used to create
the function:
>>> f = lambdify((x, (y, z)), x + y)
>>> f(1, (2, 4))
3
The ``flatten`` function can be used to always work with flattened
arguments:
>>> from sympy.utilities.iterables import flatten
>>> args = w, (x, (y, z))
>>> vals = 1, (2, (3, 4))
>>> f = lambdify(flatten(args), w + x + y + z)
>>> f(*flatten(vals))
10
Functions present in ``expr`` can also carry their own numerical
implementations, in a callable attached to the ``_imp_`` attribute. This
can be used with undefined functions using the ``implemented_function``
factory:
>>> f = implemented_function(Function('f'), lambda x: x+1)
>>> func = lambdify(x, f(x))
>>> func(4)
5
``lambdify`` always prefers ``_imp_`` implementations to implementations
in other namespaces, unless the ``use_imps`` input parameter is False.
Usage with Tensorflow:
>>> import tensorflow as tf
>>> from sympy import Max, sin, lambdify
>>> from sympy.abc import x
>>> f = Max(x, sin(x))
>>> func = lambdify(x, f, 'tensorflow')
After tensorflow v2, eager execution is enabled by default.
If you want to get the compatible result across tensorflow v1 and v2
as same as this tutorial, run this line.
>>> tf.compat.v1.enable_eager_execution()
If you have eager execution enabled, you can get the result out
immediately as you can use numpy.
If you pass tensorflow objects, you may get an ``EagerTensor``
object instead of value.
>>> result = func(tf.constant(1.0))
>>> print(result)
tf.Tensor(1.0, shape=(), dtype=float32)
>>> print(result.__class__)
<class 'tensorflow.python.framework.ops.EagerTensor'>
You can use ``.numpy()`` to get the numpy value of the tensor.
>>> result.numpy()
1.0
>>> var = tf.Variable(2.0)
>>> result = func(var) # also works for tf.Variable and tf.Placeholder
>>> result.numpy()
2.0
And it works with any shape array.
>>> tensor = tf.constant([[1.0, 2.0], [3.0, 4.0]])
>>> result = func(tensor)
>>> result.numpy()
[[1. 2.]
[3. 4.]]
Notes
=====
- For functions involving large array calculations, numexpr can provide a
significant speedup over numpy. Please note that the available functions
for numexpr are more limited than numpy but can be expanded with
``implemented_function`` and user defined subclasses of Function. If
specified, numexpr may be the only option in modules. The official list
of numexpr functions can be found at:
https://numexpr.readthedocs.io/en/latest/user_guide.html#supported-functions
- In previous versions of SymPy, ``lambdify`` replaced ``Matrix`` with
``numpy.matrix`` by default. As of SymPy 1.0 ``numpy.array`` is the
default. To get the old default behavior you must pass in
``[{'ImmutableDenseMatrix': numpy.matrix}, 'numpy']`` to the
``modules`` kwarg.
>>> from sympy import lambdify, Matrix
>>> from sympy.abc import x, y
>>> import numpy
>>> array2mat = [{'ImmutableDenseMatrix': numpy.matrix}, 'numpy']
>>> f = lambdify((x, y), Matrix([x, y]), modules=array2mat)
>>> f(1, 2)
[[1]
[2]]
- In the above examples, the generated functions can accept scalar
values or numpy arrays as arguments. However, in some cases
the generated function relies on the input being a numpy array:
>>> from sympy import Piecewise
>>> from sympy.utilities.pytest import ignore_warnings
>>> f = lambdify(x, Piecewise((x, x <= 1), (1/x, x > 1)), "numpy")
>>> with ignore_warnings(RuntimeWarning):
... f(numpy.array([-1, 0, 1, 2]))
[-1. 0. 1. 0.5]
>>> f(0)
Traceback (most recent call last):
...
ZeroDivisionError: division by zero
In such cases, the input should be wrapped in a numpy array:
>>> with ignore_warnings(RuntimeWarning):
... float(f(numpy.array([0])))
0.0
Or if numpy functionality is not required another module can be used:
>>> f = lambdify(x, Piecewise((x, x <= 1), (1/x, x > 1)), "math")
>>> f(0)
0
"""
from sympy.core.symbol import Symbol
# If the user hasn't specified any modules, use what is available.
if modules is None:
try:
_import("scipy")
except ImportError:
try:
_import("numpy")
except ImportError:
# Use either numpy (if available) or python.math where possible.
# XXX: This leads to different behaviour on different systems and
# might be the reason for irreproducible errors.
modules = ["math", "mpmath", "sympy"]
else:
modules = ["numpy"]
else:
modules = ["numpy", "scipy"]
# Get the needed namespaces.
namespaces = []
# First find any function implementations
if use_imps:
namespaces.append(_imp_namespace(expr))
# Check for dict before iterating
if isinstance(modules,
(dict, string_types)) or not hasattr(modules, '__iter__'):
namespaces.append(modules)
else:
# consistency check
if _module_present('numexpr', modules) and len(modules) > 1:
raise TypeError("numexpr must be the only item in 'modules'")
namespaces += list(modules)
# fill namespace with first having highest priority
namespace = {}
for m in namespaces[::-1]:
buf = _get_namespace(m)
namespace.update(buf)
if hasattr(expr, "atoms"):
#Try if you can extract symbols from the expression.
#Move on if expr.atoms in not implemented.
syms = expr.atoms(Symbol)
for term in syms:
namespace.update({str(term): term})
if printer is None:
if _module_present('mpmath', namespaces):
from sympy.printing.pycode import MpmathPrinter as Printer
elif _module_present('scipy', namespaces):
from sympy.printing.pycode import SciPyPrinter as Printer
elif _module_present('numpy', namespaces):
from sympy.printing.pycode import NumPyPrinter as Printer
elif _module_present('numexpr', namespaces):
from sympy.printing.lambdarepr import NumExprPrinter as Printer
elif _module_present('tensorflow', namespaces):
from sympy.printing.tensorflow import TensorflowPrinter as Printer
elif _module_present('sympy', namespaces):
from sympy.printing.pycode import SymPyPrinter as Printer
else:
from sympy.printing.pycode import PythonCodePrinter as Printer
user_functions = {}
for m in namespaces[::-1]:
if isinstance(m, dict):
for k in m:
user_functions[k] = k
printer = Printer({
'fully_qualified_modules': False,
'inline': True,
'allow_unknown_functions': True,
'user_functions': user_functions
})
# Get the names of the args, for creating a docstring
if not iterable(args):
args = (args, )
names = []
# Grab the callers frame, for getting the names by inspection (if needed)
callers_local_vars = inspect.currentframe().f_back.f_locals.items()
for n, var in enumerate(args):
if hasattr(var, 'name'):
names.append(var.name)
else:
# It's an iterable. Try to get name by inspection of calling frame.
name_list = [
var_name for var_name, var_val in callers_local_vars
if var_val is var
]
if len(name_list) == 1:
names.append(name_list[0])
else:
# Cannot infer name with certainty. arg_# will have to do.
names.append('arg_' + str(n))
# Create the function definition code and execute it
funcname = '_lambdifygenerated'
if _module_present('tensorflow', namespaces):
funcprinter = _TensorflowEvaluatorPrinter(printer, dummify)
else:
funcprinter = _EvaluatorPrinter(printer, dummify)
funcstr = funcprinter.doprint(funcname, args, expr)
# Collect the module imports from the code printers.
imp_mod_lines = []
for mod, keys in (getattr(printer, 'module_imports', None) or {}).items():
for k in keys:
if k not in namespace:
ln = "from %s import %s" % (mod, k)
try:
exec_(ln, {}, namespace)
except ImportError:
# Tensorflow 2.0 has issues with importing a specific
# function from its submodule.
# https://github.com/tensorflow/tensorflow/issues/33022
ln = "%s = %s.%s" % (k, mod, k)
exec_(ln, {}, namespace)
imp_mod_lines.append(ln)
# Provide lambda expression with builtins, and compatible implementation of range
namespace.update({'builtins': builtins, 'range': range})
funclocals = {}
global _lambdify_generated_counter
filename = '<lambdifygenerated-%s>' % _lambdify_generated_counter
_lambdify_generated_counter += 1
c = compile(funcstr, filename, 'exec')
exec_(c, namespace, funclocals)
# mtime has to be None or else linecache.checkcache will remove it
linecache.cache[filename] = (len(funcstr), None, funcstr.splitlines(True),
filename)
func = funclocals[funcname]
# Apply the docstring
sig = "func({0})".format(", ".join(str(i) for i in names))
sig = textwrap.fill(sig, subsequent_indent=' ' * 8)
expr_str = str(expr)
if len(expr_str) > 78:
expr_str = textwrap.wrap(expr_str, 75)[0] + '...'
func.__doc__ = ("Created with lambdify. Signature:\n\n"
"{sig}\n\n"
"Expression:\n\n"
"{expr}\n\n"
"Source code:\n\n"
"{src}\n\n"
"Imported modules:\n\n"
"{imp_mods}").format(sig=sig,
expr=expr_str,
src=funcstr,
imp_mods='\n'.join(imp_mod_lines))
return func
def factor_terms(expr, radical=False, clear=False, fraction=False):
"""Remove common factors from terms in all arguments without
changing the underlying structure of the expr. No expansion or
simplification (and no processing of non-commutatives) is performed.
If radical=True then a radical common to all terms will be factored
out of any Add sub-expressions of the expr.
If clear=False (default) then coefficients will not be separated
from a single Add if they can be distributed to leave one or more
terms with integer coefficients.
If fraction=True (default is False) then a common denominator will be
constructed for the expression.
Examples
========
>>> from sympy import factor_terms, Symbol, Mul, primitive
>>> from sympy.abc import x, y
>>> factor_terms(x + x*(2 + 4*y)**3)
x*(8*(2*y + 1)**3 + 1)
>>> A = Symbol('A', commutative=False)
>>> factor_terms(x*A + x*A + x*y*A)
x*(y*A + 2*A)
When ``clear`` is False, a rational will only be factored out of an
Add expression if all terms of the Add have coefficients that are
fractions:
>>> factor_terms(x/2 + 1, clear=False)
x/2 + 1
>>> factor_terms(x/2 + 1, clear=True)
(x + 2)/2
This only applies when there is a single Add that the coefficient
multiplies:
>>> factor_terms(x*y/2 + y, clear=True)
y*(x + 2)/2
>>> factor_terms(x*y/2 + y, clear=False) == _
True
See Also
========
gcd_terms, sympy.polys.polytools.terms_gcd
"""
expr = sympify(expr)
is_iterable = iterable(expr)
if not isinstance(expr, Basic) or expr.is_Atom:
if is_iterable:
return type(expr)([factor_terms(i,
radical=radical,
clear=clear,
fraction=fraction) for i in expr])
return expr
if expr.is_Pow or expr.is_Function or is_iterable or not hasattr(expr, 'args_cnc'):
args = expr.args
newargs = tuple([factor_terms(i,
radical=radical,
clear=clear,
fraction=fraction) for i in args])
if newargs == args:
return expr
return expr.func(*newargs)
cont, p = expr.as_content_primitive(radical=radical)
if p.is_Add:
list_args = [gcd_terms(a,
isprimitive=True,
clear=clear,
fraction=fraction) for a in Add.make_args(p)]
p = Add._from_args(list_args) # gcd_terms will fix up ordering
elif p.args:
p = p.func(
*[factor_terms(a, radical, clear, fraction) for a in p.args])
p = gcd_terms(p,
isprimitive=True,
clear=clear,
fraction=fraction)
return _keep_coeff(cont, p, clear=clear)
def lambdify(args,
expr,
modules=None,
printer=None,
use_imps=True,
dummify=True):
"""
Returns a lambda function for fast calculation of numerical values.
If not specified differently by the user, ``modules`` defaults to
``["numpy"]`` if NumPy is installed, and ``["math", "mpmath", "sympy"]``
if it isn't, that is, SymPy functions are replaced as far as possible by
either ``numpy`` functions if available, and Python's standard library
``math``, or ``mpmath`` functions otherwise. To change this behavior, the
"modules" argument can be used. It accepts:
- the strings "math", "mpmath", "numpy", "numexpr", "sympy", "tensorflow"
- any modules (e.g. math)
- dictionaries that map names of sympy functions to arbitrary functions
- lists that contain a mix of the arguments above, with higher priority
given to entries appearing first.
.. warning::
Note that this function uses ``eval``, and thus shouldn't be used on
unsanitized input.
The default behavior is to substitute all arguments in the provided
expression with dummy symbols. This allows for applied functions (e.g.
f(t)) to be supplied as arguments. Call the function with dummify=False if
dummy substitution is unwanted (and `args` is not a string). If you want
to view the lambdified function or provide "sympy" as the module, you
should probably set dummify=False.
For functions involving large array calculations, numexpr can provide a
significant speedup over numpy. Please note that the available functions
for numexpr are more limited than numpy but can be expanded with
implemented_function and user defined subclasses of Function. If specified,
numexpr may be the only option in modules. The official list of numexpr
functions can be found at:
https://github.com/pydata/numexpr#supported-functions
In previous releases ``lambdify`` replaced ``Matrix`` with ``numpy.matrix``
by default. As of release 1.0 ``numpy.array`` is the default.
To get the old default behavior you must pass in ``[{'ImmutableDenseMatrix':
numpy.matrix}, 'numpy']`` to the ``modules`` kwarg.
>>> from sympy import lambdify, Matrix
>>> from sympy.abc import x, y
>>> import numpy
>>> array2mat = [{'ImmutableDenseMatrix': numpy.matrix}, 'numpy']
>>> f = lambdify((x, y), Matrix([x, y]), modules=array2mat)
>>> f(1, 2)
matrix([[1],
[2]])
Usage
=====
(1) Use one of the provided modules:
>>> from sympy import sin, tan, gamma
>>> from sympy.utilities.lambdify import lambdastr
>>> from sympy.abc import x, y
>>> f = lambdify(x, sin(x), "math")
Attention: Functions that are not in the math module will throw a name
error when the lambda function is evaluated! So this would
be better:
>>> f = lambdify(x, sin(x)*gamma(x), ("math", "mpmath", "sympy"))
(2) Use some other module:
>>> import numpy
>>> f = lambdify((x,y), tan(x*y), numpy)
Attention: There are naming differences between numpy and sympy. So if
you simply take the numpy module, e.g. sympy.atan will not be
translated to numpy.arctan. Use the modified module instead
by passing the string "numpy":
>>> f = lambdify((x,y), tan(x*y), "numpy")
>>> f(1, 2)
-2.18503986326
>>> from numpy import array
>>> f(array([1, 2, 3]), array([2, 3, 5]))
[-2.18503986 -0.29100619 -0.8559934 ]
(3) Use a dictionary defining custom functions:
>>> def my_cool_function(x): return 'sin(%s) is cool' % x
>>> myfuncs = {"sin" : my_cool_function}
>>> f = lambdify(x, sin(x), myfuncs); f(1)
'sin(1) is cool'
Examples
========
>>> from sympy.utilities.lambdify import implemented_function
>>> from sympy import sqrt, sin, Matrix
>>> from sympy import Function
>>> from sympy.abc import w, x, y, z
>>> f = lambdify(x, x**2)
>>> f(2)
4
>>> f = lambdify((x, y, z), [z, y, x])
>>> f(1,2,3)
[3, 2, 1]
>>> f = lambdify(x, sqrt(x))
>>> f(4)
2.0
>>> f = lambdify((x, y), sin(x*y)**2)
>>> f(0, 5)
0.0
>>> row = lambdify((x, y), Matrix((x, x + y)).T, modules='sympy')
>>> row(1, 2)
Matrix([[1, 3]])
Tuple arguments are handled and the lambdified function should
be called with the same type of arguments as were used to create
the function.:
>>> f = lambdify((x, (y, z)), x + y)
>>> f(1, (2, 4))
3
A more robust way of handling this is to always work with flattened
arguments:
>>> from sympy.utilities.iterables import flatten
>>> args = w, (x, (y, z))
>>> vals = 1, (2, (3, 4))
>>> f = lambdify(flatten(args), w + x + y + z)
>>> f(*flatten(vals))
10
Functions present in `expr` can also carry their own numerical
implementations, in a callable attached to the ``_imp_``
attribute. Usually you attach this using the
``implemented_function`` factory:
>>> f = implemented_function(Function('f'), lambda x: x+1)
>>> func = lambdify(x, f(x))
>>> func(4)
5
``lambdify`` always prefers ``_imp_`` implementations to implementations
in other namespaces, unless the ``use_imps`` input parameter is False.
Usage with Tensorflow module:
>>> import tensorflow as tf
>>> f = Max(x, sin(x))
>>> func = lambdify(x, f, 'tensorflow')
>>> result = func(tf.constant(1.0))
>>> result # a tf.Tensor representing the result of the calculation
<tf.Tensor 'Maximum:0' shape=() dtype=float32>
>>> sess = tf.Session()
>>> sess.run(result) # compute result
1.0
>>> var = tf.Variable(1.0)
>>> sess.run(tf.global_variables_initializer())
>>> sess.run(func(var)) # also works for tf.Variable and tf.Placeholder
1.0
>>> tensor = tf.constant([[1.0, 2.0], [3.0, 4.0]]) # works with any shape tensor
>>> sess.run(func(tensor))
array([[ 1., 2.],
[ 3., 4.]], dtype=float32)
"""
from sympy.core.symbol import Symbol
from sympy.utilities.iterables import flatten
# If the user hasn't specified any modules, use what is available.
module_provided = True
if modules is None:
module_provided = False
try:
_import("numpy")
except ImportError:
# Use either numpy (if available) or python.math where possible.
# XXX: This leads to different behaviour on different systems and
# might be the reason for irreproducible errors.
modules = ["math", "mpmath", "sympy"]
else:
modules = ["numpy"]
# Get the needed namespaces.
namespaces = []
# First find any function implementations
if use_imps:
namespaces.append(_imp_namespace(expr))
# Check for dict before iterating
if isinstance(modules, (dict, str)) or not hasattr(modules, '__iter__'):
namespaces.append(modules)
else:
# consistency check
if _module_present('numexpr', modules) and len(modules) > 1:
raise TypeError("numexpr must be the only item in 'modules'")
namespaces += list(modules)
# fill namespace with first having highest priority
namespace = {}
for m in namespaces[::-1]:
buf = _get_namespace(m)
namespace.update(buf)
if hasattr(expr, "atoms"):
#Try if you can extract symbols from the expression.
#Move on if expr.atoms in not implemented.
syms = expr.atoms(Symbol)
for term in syms:
namespace.update({str(term): term})
if printer is None:
if _module_present('mpmath', namespaces):
from sympy.printing.pycode import MpmathPrinter as Printer
elif _module_present('numpy', namespaces):
from sympy.printing.pycode import NumPyPrinter as Printer
elif _module_present('numexpr', namespaces):
from sympy.printing.lambdarepr import NumExprPrinter as Printer
elif _module_present('tensorflow', namespaces):
from sympy.printing.lambdarepr import TensorflowPrinter as Printer
elif _module_present('sympy', namespaces):
from sympy.printing.pycode import SymPyPrinter as Printer
else:
from sympy.printing.pycode import PythonCodePrinter as Printer
user_functions = {}
for m in namespaces[::-1]:
if isinstance(m, dict):
for k in m:
user_functions[k] = k
printer = Printer({
'fully_qualified_modules': False,
'inline': True,
'user_functions': user_functions
})
# Get the names of the args, for creating a docstring
if not iterable(args):
args = (args, )
names = []
# Grab the callers frame, for getting the names by inspection (if needed)
callers_local_vars = inspect.currentframe().f_back.f_locals.items()
for n, var in enumerate(args):
if hasattr(var, 'name'):
names.append(var.name)
else:
# It's an iterable. Try to get name by inspection of calling frame.
name_list = [
var_name for var_name, var_val in callers_local_vars
if var_val is var
]
if len(name_list) == 1:
names.append(name_list[0])
else:
# Cannot infer name with certainty. arg_# will have to do.
names.append('arg_' + str(n))
# Create lambda function.
lstr = lambdastr(args, expr, printer=printer, dummify=dummify)
flat = '__flatten_args__'
imp_mod_lines = []
for mod, keys in (getattr(printer, 'module_imports', None) or {}).items():
for k in keys:
if k not in namespace:
imp_mod_lines.append("from %s import %s" % (mod, k))
for ln in imp_mod_lines:
exec_(ln, {}, namespace)
if flat in lstr:
namespace.update({flat: flatten})
# Provide lambda expression with builtins, and compatible implementation of range
namespace.update({'builtins': builtins, 'range': range})
func = eval(lstr, namespace)
# For numpy lambdify, wrap all input arguments in arrays.
# This is a fix for gh-11306.
if module_provided and _module_present('numpy', namespaces):
def array_wrap(funcarg):
@wraps(funcarg)
def wrapper(*argsx, **kwargsx):
asarray = namespace['asarray']
newargs = [
asarray(i) if isinstance(i, integer_types +
(float, complex)) else i
for i in argsx
]
return funcarg(*newargs, **kwargsx)
return wrapper
func = array_wrap(func)
# Apply the docstring
sig = "func({0})".format(", ".join(str(i) for i in names))
sig = textwrap.fill(sig, subsequent_indent=' ' * 8)
expr_str = str(expr)
if len(expr_str) > 78:
expr_str = textwrap.wrap(expr_str, 75)[0] + '...'
func.__doc__ = ("Created with lambdify. Signature:\n\n"
"{sig}\n\n"
"Expression:\n\n"
"{expr}\n\n"
"Source code:\n\n"
"{src}\n\n"
"Imported modules:\n\n"
"{imp_mods}").format(sig=sig,
expr=expr_str,
src=lstr,
imp_mods='\n'.join(imp_mod_lines))
return func
def _as_finite_diff(derivative, points=1, x0=None, wrt=None):
"""
Returns an approximation of a derivative of a function in
the form of a finite difference formula. The expression is a
weighted sum of the function at a number of discrete values of
(one of) the independent variable(s).
Parameters
==========
derivative: a Derivative instance
points: sequence or coefficient, optional
If sequence: discrete values (length >= order+1) of the
independent variable used for generating the finite
difference weights.
If it is a coefficient, it will be used as the step-size
for generating an equidistant sequence of length order+1
centered around ``x0``. default: 1 (step-size 1)
x0: number or Symbol, optional
the value of the independent variable (``wrt``) at which the
derivative is to be approximated. Default: same as ``wrt``.
wrt: Symbol, optional
"with respect to" the variable for which the (partial)
derivative is to be approximated for. If not provided it
is required that the Derivative is ordinary. Default: ``None``.
Examples
========
>>> from sympy import symbols, Function, exp, sqrt, Symbol, as_finite_diff
>>> from sympy.utilities.exceptions import SymPyDeprecationWarning
>>> import warnings
>>> warnings.simplefilter("ignore", SymPyDeprecationWarning)
>>> x, h = symbols('x h')
>>> f = Function('f')
>>> as_finite_diff(f(x).diff(x))
-f(x - 1/2) + f(x + 1/2)
The default step size and number of points are 1 and ``order + 1``
respectively. We can change the step size by passing a symbol
as a parameter:
>>> as_finite_diff(f(x).diff(x), h)
-f(-h/2 + x)/h + f(h/2 + x)/h
We can also specify the discretized values to be used in a sequence:
>>> as_finite_diff(f(x).diff(x), [x, x+h, x+2*h])
-3*f(x)/(2*h) + 2*f(h + x)/h - f(2*h + x)/(2*h)
The algorithm is not restricted to use equidistant spacing, nor
do we need to make the approximation around ``x0``, but we can get
an expression estimating the derivative at an offset:
>>> e, sq2 = exp(1), sqrt(2)
>>> xl = [x-h, x+h, x+e*h]
>>> as_finite_diff(f(x).diff(x, 1), xl, x+h*sq2)
2*h*((h + sqrt(2)*h)/(2*h) - (-sqrt(2)*h + h)/(2*h))*f(E*h + x)/((-h + E*h)*(h + E*h)) +
(-(-sqrt(2)*h + h)/(2*h) - (-sqrt(2)*h + E*h)/(2*h))*f(-h + x)/(h + E*h) +
(-(h + sqrt(2)*h)/(2*h) + (-sqrt(2)*h + E*h)/(2*h))*f(h + x)/(-h + E*h)
Partial derivatives are also supported:
>>> y = Symbol('y')
>>> d2fdxdy=f(x,y).diff(x,y)
>>> as_finite_diff(d2fdxdy, wrt=x)
-Derivative(f(x - 1/2, y), y) + Derivative(f(x + 1/2, y), y)
See also
========
sympy.calculus.finite_diff.apply_finite_diff
sympy.calculus.finite_diff.finite_diff_weights
"""
if derivative.is_Derivative:
pass
elif derivative.is_Atom:
return derivative
else:
return derivative.fromiter(
[_as_finite_diff(ar, points, x0, wrt) for ar
in derivative.args], **derivative.assumptions0)
if wrt is None:
old = None
for v in derivative.variables:
if old is v:
continue
derivative = _as_finite_diff(derivative, points, x0, v)
old = v
return derivative
order = derivative.variables.count(wrt)
if x0 is None:
x0 = wrt
if not iterable(points):
if getattr(points, 'is_Function', False) and wrt in points.args:
points = points.subs(wrt, x0)
# points is simply the step-size, let's make it a
# equidistant sequence centered around x0
if order % 2 == 0:
# even order => odd number of points, grid point included
points = [x0 + points*i for i
in range(-order//2, order//2 + 1)]
else:
# odd order => even number of points, half-way wrt grid point
points = [x0 + points*S(i)/2 for i
in range(-order, order + 1, 2)]
others = [wrt, 0]
for v in set(derivative.variables):
if v == wrt:
continue
others += [v, derivative.variables.count(v)]
if len(points) < order+1:
raise ValueError("Too few points for order %d" % order)
return apply_finite_diff(order, points, [
Derivative(derivative.expr.subs({wrt: x}), *others) for
x in points], x0)
def solve(f, *symbols, **flags):
"""
Algebraically solves equations and systems of equations.
Currently supported are:
- univariate polynomial,
- transcendental
- piecewise combinations of the above
- systems of linear and polynomial equations
- sytems containing relational expressions.
Input is formed as:
f
- a single Expr or Poly that must be zero,
- an Equality
- a Relational expression or boolean
- iterable of one or more of the above
symbols (Symbol, Function or Derivative) specified as
- none given (all free symbols will be used)
- single symbol
- denested list of symbols
e.g. solve(f, x, y)
- ordered iterable of symbols
e.g. solve(f, [x, y])
flags
- ``simplified``, when False, will not simplify solutions
(default=True except for polynomials of
order 3 or greater)
The output varies according to the input and can be seen by example:
>>> from sympy import solve, Poly, Eq, Function, exp
>>> from sympy.abc import x, y, z, a, b
o boolean or univariate Relational
>>> solve(x < 3)
And(im(x) == 0, re(x) < 3)
o single expression and single symbol that is in the expression
>>> solve(x - y, x)
[y]
>>> solve(x - 3, x)
[3]
>>> solve(Eq(x, 3), x)
[3]
>>> solve(Poly(x - 3), x)
[3]
>>> solve(x**2 - y**2, x)
[y, -y]
>>> solve(x**4 - 1, x)
[1, -1, -I, I]
o single expression with no symbol that is in the expression
>>> solve(3, x)
[]
>>> solve(x - 3, y)
[]
o when no symbol is given then all free symbols will be used
and sorted with default_sort_key and the result will be the
same as above as if those symbols had been supplied
>>> solve(x - 3)
[3]
>>> solve(x**2 - y**2)
[y, -y]
o when a Function or Derivative is given as a symbol, it is isolated
algebraically and an implicit solution may be obtained
>>> f = Function('f')
>>> solve(f(x) - x, f(x))
[x]
>>> solve(f(x).diff(x) - f(x) - x, f(x).diff(x))
[x + f(x)]
o single expression and more than 1 symbol
when there is a linear solution
>>> solve(x - y**2, x, y)
{x: y**2}
>>> solve(x**2 - y, x, y)
{y: x**2}
when undetermined coefficients are identified
that are linear
>>> solve((a + b)*x - b + 2, a, b)
{a: -2, b: 2}
that are nonlinear
>>> solve((a + b)*x - b**2 + 2, a, b)
[(-2**(1/2), 2**(1/2)), (2**(1/2), -2**(1/2))]
if there is no linear solution then the first successful
attempt for a nonlinear solution will be returned
>>> solve(x**2 - y**2, x, y)
[y, -y]
>>> solve(x**2 - y**2/exp(x), x, y)
[x*exp(x/2), -x*exp(x/2)]
o iterable of one or more of the above
involving relationals or bools
>>> solve([x < 3, x - 2])
And(im(x) == 0, re(x) == 2)
>>> solve([x > 3, x - 2])
False
when the system is linear
with a solution
>>> solve([x - 3], x)
{x: 3}
>>> solve((x + 5*y - 2, -3*x + 6*y - 15), x, y)
{x: -3, y: 1}
>>> solve((x + 5*y - 2, -3*x + 6*y - 15), x, y, z)
{x: -3, y: 1}
>>> solve((x + 5*y - 2, -3*x + 6*y - z), z, x, y)
{x: -5*y + 2, z: 21*y - 6}
without a solution
>>> solve([x + 3, x - 3])
when the system is not linear
>>> solve([x**2 + y -2, y**2 - 4], x, y)
[(-2, -2), (0, 2), (0, 2), (2, -2)]
Warning: there is a possibility of obtaining ambiguous results
if no symbols are given for a nonlinear system of equations or
are given as a set since the symbols are not presently reported
with the solution. A warning will be issued in this situation.
>>> solve([x - 2, x**2 + y])
<BLANKLINE>
For nonlinear systems of equations, symbols should be
given as a list so as to avoid ambiguity in the results.
solve sorted the symbols as [x, y]
[(2, -4)]
>>> solve([x - 2, x**2 + f(x)], set([f(x), x]))
<BLANKLINE>
For nonlinear systems of equations, symbols should be
given as a list so as to avoid ambiguity in the results.
solve sorted the symbols as [x, f(x)]
[(2, -4)]
See also:
rsolve() for solving recurrence relationships
dsolve() for solving differential equations
"""
# make f and symbols into lists of sympified quantities
# keeping track of how f was passed since if it is a list
# a dictionary of results will be returned.
###########################################################################
def sympified_list(w):
return map(sympify, iff(iterable(w), w, [w]))
bare_f = not iterable(f)
ordered_symbols = (symbols and symbols[0] and
(isinstance(symbols[0], Symbol)
or ordered_iter(symbols[0], include=GeneratorType)))
f, symbols = (sympified_list(w) for w in [f, symbols])
# preprocess equation(s)
###########################################################################
for i, fi in enumerate(f):
if isinstance(fi, Equality):
f[i] = fi.lhs - fi.rhs
elif isinstance(fi, Poly):
f[i] = fi.as_expr()
elif isinstance(fi, bool) or fi.is_Relational:
return reduce_inequalities(f, assume=flags.get('assume'))
# Any embedded piecewise functions need to be brought out to the
# top level so that the appropriate strategy gets selected.
f[i] = piecewise_fold(f[i])
# preprocess symbol(s)
###########################################################################
if not symbols:
# get symbols from equations or supply dummy symbols so solve(3) behaves
# like solve(3, x).
symbols = set([])
for fi in f:
symbols |= fi.free_symbols or set([Dummy()])
ordered_symbols = False
elif len(symbols) == 1 and iterable(symbols[0]):
symbols = symbols[0]
if not ordered_symbols:
# we do this to make the results returned canonical in case f
# contains a system of nonlinear equations; all other cases should
# be unambiguous
symbols = sorted(symbols, key=lambda i: i.sort_key())
# we can solve for Function and Derivative instances by replacing them
# with Dummy symbols
symbols_new = []
symbol_swapped = False
symbols_passed = list(symbols)
for i, s in enumerate(symbols):
if s.is_Symbol:
s_new = s
elif s.is_Function:
symbol_swapped = True
s_new = Dummy('F%d' % i)
elif s.is_Derivative:
symbol_swapped = True
s_new = Dummy('D%d' % i)
else:
msg = 'expected Symbol, Function or Derivative but got %s'
raise TypeError(msg % type(s))
symbols_new.append(s_new)
if symbol_swapped:
swap_back_dict = dict(zip(symbols_new, symbols))
swap_dict = zip(symbols, symbols_new)
f = [fi.subs(swap_dict) for fi in f]
symbols = symbols_new
#
# try to get a solution
###########################################################################
if bare_f:
# pass f the way it was passed to solve; if it wasn't a list then
# a list of solutions will be returned, otherwise a dictionary is
# going to be returned
f = f[0]
solution = _solve(f, *symbols, **flags)
#
# postprocessing
###########################################################################
# Restore original Functions and Derivatives if a dictionary is returned.
# This is not necessary for
# - the single equation, single unknown case
# since the symbol will have been removed from the solution;
# - the nonlinear poly_system since that only support zero-dimensional
# systems and those results come back as a list
if symbol_swapped and type(solution) is dict:
solution = dict([(swap_back_dict[k], v.subs(swap_back_dict))
for k, v in solution.iteritems()])
# warn if ambiguous results are being obtained
# XXX agree on how to make this unambiguous
# see issue 2405 for logic in how Polys chooses ordering and
# for discussion of what to return see http://groups.google.com/group/sympy
# Apr 18, 2011 posting 'using results from solve'
elif (not ordered_symbols and len(symbols) > 1 and solution
and ordered_iter(solution) and ordered_iter(solution[0])
and any(len(set(s)) > 1 for s in solution)):
msg = ('\n\tFor nonlinear systems of equations, symbols should be' +
'\n\tgiven as a list so as to avoid ambiguity in the results.' +
'\n\tsolve sorted the symbols as %s')
print msg % str(
bool(symbol_swapped) and list(zip(*swap_dict)[0]) or symbols)
#
# done
###########################################################################
return solution
def test_not_interable():
i, j = symbols('i j', integer=True)
A = Indexed('A', i, i + j)
assert not iterable(A)
def euler_equations(L, funcs=(), vars=()):
r"""
Find the Euler-Lagrange equations [1]_ for a given Lagrangian.
Parameters
==========
L : Expr
The Lagrangian that should be a function of the functions listed
in the second argument and their derivatives.
For example, in the case of two functions `f(x,y)`, `g(x,y)` and
two independent variables `x`, `y` the Lagrangian would have the form:
.. math:: L\left(f(x,y),g(x,y),\frac{\partial f(x,y)}{\partial x},
\frac{\partial f(x,y)}{\partial y},
\frac{\partial g(x,y)}{\partial x},
\frac{\partial g(x,y)}{\partial y},x,y\right)
In many cases it is not necessary to provide anything, except the
Lagrangian, it will be auto-detected (and an error raised if this
couldn't be done).
funcs : Function or an iterable of Functions
The functions that the Lagrangian depends on. The Euler equations
are differential equations for each of these functions.
vars : Symbol or an iterable of Symbols
The Symbols that are the independent variables of the functions.
Returns
=======
eqns : list of Eq
The list of differential equations, one for each function.
Examples
========
>>> from sympy import Symbol, Function
>>> from sympy.calculus.euler import euler_equations
>>> x = Function('x')
>>> t = Symbol('t')
>>> L = (x(t).diff(t))**2/2 - x(t)**2/2
>>> euler_equations(L, x(t), t)
[Eq(-x(t) - Derivative(x(t), t, t), 0)]
>>> u = Function('u')
>>> x = Symbol('x')
>>> L = (u(t, x).diff(t))**2/2 - (u(t, x).diff(x))**2/2
>>> euler_equations(L, u(t, x), [t, x])
[Eq(-Derivative(u(t, x), t, t) + Derivative(u(t, x), x, x), 0)]
References
==========
.. [1] http://en.wikipedia.org/wiki/Euler%E2%80%93Lagrange_equation
"""
funcs = tuple(funcs) if iterable(funcs) else (funcs, )
if not funcs:
funcs = tuple(L.atoms(Function))
else:
for f in funcs:
if not isinstance(f, Function):
raise TypeError('Function expected, got: %s' % f)
vars = tuple(vars) if iterable(vars) else (vars, )
if not vars:
vars = funcs[0].args
else:
vars = tuple(sympify(var) for var in vars)
if not all(isinstance(v, Symbol) for v in vars):
raise TypeError('Variables are not symbols, got %s' % vars)
for f in funcs:
if not vars == f.args:
raise ValueError("Variables %s don't match args: %s" % (vars, f))
order = max(
len(d.variables) for d in L.atoms(Derivative) if d.expr in funcs)
eqns = []
for f in funcs:
eq = diff(L, f)
for i in range(1, order + 1):
for p in combinations_with_replacement(vars, i):
eq = eq + S.NegativeOne**i * diff(L, diff(f, *p), *p)
eqns.append(Eq(eq))
return eqns
def sympify(a, locals=None, convert_xor=True, strict=False, rational=False):
"""
Converts an arbitrary expression to a type that can be used inside sympy.
For example, it will convert python ints into instance of sympy.Rational,
floats into instances of sympy.Float, etc. It is also able to coerce symbolic
expressions which inherit from Basic. This can be useful in cooperation
with SAGE.
It currently accepts as arguments:
- any object defined in sympy (except matrices [TODO])
- standard numeric python types: int, long, float, Decimal
- strings (like "0.09" or "2e-19")
- booleans, including ``None`` (will leave them unchanged)
- lists, sets or tuples containing any of the above
If the argument is already a type that sympy understands, it will do
nothing but return that value. This can be used at the beginning of a
function to ensure you are working with the correct type.
>>> from sympy import sympify
>>> sympify(2).is_integer
True
>>> sympify(2).is_real
True
>>> sympify(2.0).is_real
True
>>> sympify("2.0").is_real
True
>>> sympify("2e-45").is_real
True
If the expression could not be converted, a SympifyError is raised.
>>> sympify("x***2")
Traceback (most recent call last):
...
SympifyError: SympifyError: "could not parse u'x***2'"
If the option ``strict`` is set to ``True``, only the types for which an
explicit conversion has been defined are converted. In the other
cases, a SympifyError is raised.
>>> sympify(True)
True
>>> sympify(True, strict=True)
Traceback (most recent call last):
...
SympifyError: SympifyError: True
To extend `sympify` to convert custom objects (not derived from `Basic`),
the static dictionary `convert` is provided. The custom converters are
usually added at import time, and will apply to all objects of the given
class or its derived classes.
For example, all geometry objects derive from `GeometryEntity` class, and
should not be altered by the converter, so we add the following after
defining that class:
>>> from sympy.core.sympify import converter
>>> from sympy.geometry.entity import GeometryEntity
>>> converter[GeometryEntity] = lambda x: x
"""
try:
cls = a.__class__
except AttributeError: #a is probably an old-style class object
cls = type(a)
if cls in sympy_classes:
return a
if cls in (bool, type(None)):
if strict:
raise SympifyError(a)
else:
return a
try:
return converter[cls](a)
except KeyError:
for superclass in getmro(cls):
try:
return converter[superclass](a)
except KeyError:
continue
try:
return a._sympy_()
except AttributeError:
pass
if not isinstance(a, basestring):
for coerce in (float, int):
try:
return sympify(coerce(a))
except (TypeError, ValueError, AttributeError, SympifyError):
continue
if strict:
raise SympifyError(a)
if isinstance(a, tuple):
from containers import Tuple
return Tuple(*[sympify(x, locals=locals, convert_xor=convert_xor,
rational=rational) for x in a])
if iterable(a):
try:
return type(a)([sympify(x, locals=locals, convert_xor=convert_xor,
rational=rational) for x in a])
except TypeError:
# Not all iterables are rebuildable with their type.
pass
if isinstance(a, dict):
try:
return type(a)([sympify(x, locals=locals, convert_xor=convert_xor,
rational=rational) for x in a.iteritems()])
except TypeError:
# Not all iterables are rebuildable with their type.
pass
# At this point we were given an arbitrary expression
# which does not inherit from Basic and doesn't implement
# _sympy_ (which is a canonical and robust way to convert
# anything to SymPy expression).
#
# As a last chance, we try to take "a"'s normal form via unicode()
# and try to parse it. If it fails, then we have no luck and
# return an exception
try:
a = unicode(a)
except Exception, exc:
raise SympifyError(a, exc)
def sympified_list(w):
return map(sympify, iff(iterable(w), w, [w]))
def linrec(coeffs, init, n):
r"""
Evaluation of univariate linear recurrences of homogeneous type
having coefficients independent of the recurrence variable.
Parameters
==========
coeffs : iterable
Coefficients of the recurrence
init : iterable
Initial values of the recurrence
n : Integer
Point of evaluation for the recurrence
Notes
=====
Let `y(n)` be the recurrence of given type, ``c`` be the sequence
of coefficients, ``b`` be the sequence of initial/base values of the
recurrence and ``k`` (equal to ``len(c)``) be the order of recurrence.
Then,
.. math :: y(n) = \begin{cases} b_n & 0 \le n < k \\
c_0 y(n-1) + c_1 y(n-2) + \cdots + c_{k-1} y(n-k) & n \ge k
\end{cases}
Let `x_0, x_1, \ldots, x_n` be a sequence and consider the transformation
that maps each polynomial `f(x)` to `T(f(x))` where each power `x^i` is
replaced by the corresponding value `x_i`. The sequence is then a solution
of the recurrence if and only if `T(x^i p(x)) = 0` for each `i \ge 0` where
`p(x) = x^k - c_0 x^(k-1) - \cdots - c_{k-1}` is the characteristic
polynomial.
Then `T(f(x)p(x)) = 0` for each polynomial `f(x)` (as it is a linear
combination of powers `x^i`). Now, if `x^n` is congruent to
`g(x) = a_0 x^0 + a_1 x^1 + \cdots + a_{k-1} x^{k-1}` modulo `p(x)`, then
`T(x^n) = x_n` is equal to
`T(g(x)) = a_0 x_0 + a_1 x_1 + \cdots + a_{k-1} x_{k-1}`.
Computation of `x^n`,
given `x^k = c_0 x^{k-1} + c_1 x^{k-2} + \cdots + c_{k-1}`
is performed using exponentiation by squaring (refer to [1]_) with
an additional reduction step performed to retain only first `k` powers
of `x` in the representation of `x^n`.
Examples
========
>>> from sympy.discrete.recurrences import linrec
>>> from sympy.abc import x, y, z
>>> linrec(coeffs=[1, 1], init=[0, 1], n=10)
55
>>> linrec(coeffs=[1, 1], init=[x, y], n=10)
34*x + 55*y
>>> linrec(coeffs=[x, y], init=[0, 1], n=5)
x**2*y + x*(x**3 + 2*x*y) + y**2
>>> linrec(coeffs=[1, 2, 3, 0, 0, 4], init=[x, y, z], n=16)
13576*x + 5676*y + 2356*z
References
==========
.. [1] https://en.wikipedia.org/wiki/Exponentiation_by_squaring
.. [2] https://en.wikipedia.org/w/index.php?title=Modular_exponentiation§ion=6#Matrices
See Also
========
sympy.polys.agca.extensions.ExtensionElement.__pow__
"""
if not coeffs:
return S.Zero
if not iterable(coeffs):
raise TypeError("Expected a sequence of coefficients for"
" the recurrence")
if not iterable(init):
raise TypeError("Expected a sequence of values for the initialization"
" of the recurrence")
n = as_int(n)
if n < 0:
raise ValueError("Point of evaluation of recurrence must be a "
"non-negative integer")
c = [sympify(arg) for arg in coeffs]
b = [sympify(arg) for arg in init]
k = len(c)
if len(b) > k:
raise TypeError("Count of initial values should not exceed the "
"order of the recurrence")
else:
b += [S.Zero]*(k - len(b)) # remaining initial values default to zero
def _square_and_reduce(u, offset):
# squares `(u_0 + u_1 x + u_2 x^2 + \cdots + u_{k-1} x^k)` (and
# multiplies by `x` if offset is 1) and reduces the above result of
# length upto `2k` to `k` using the characteristic equation of the
# recurrence given by, `x^k = c_0 x^{k-1} + c_1 x^{k-2} + \cdots + c_{k-1}`
w = [S.Zero]*(2*len(u) - 1 + offset)
for i, p in enumerate(u):
for j, q in enumerate(u):
w[offset + i + j] += p*q
for j in range(len(w) - 1, k - 1, -1):
for i in range(k):
w[j - i - 1] += w[j]*c[i]
return w[:k]
def _final_coeffs(n):
# computes the final coefficient list - `cf` corresponding to the
# point at which recurrence is to be evalauted - `n`, such that,
# `y(n) = cf_0 y(k-1) + cf_1 y(k-2) + \cdots + cf_{k-1} y(0)`
if n < k:
return [S.Zero]*n + [S.One] + [S.Zero]*(k - n - 1)
else:
return _square_and_reduce(_final_coeffs(n//2), n%2)
return b[n] if n < k else sum(u*v for u, v in zip(_final_coeffs(n), b))
def autowrap(expr,
language=None,
backend='f2py',
tempdir=None,
args=None,
flags=None,
verbose=False,
helpers=None,
code_gen=None,
**kwargs):
"""Generates python callable binaries based on the math expression.
Parameters
==========
expr
The SymPy expression that should be wrapped as a binary routine.
language : string, optional
If supplied, (options: 'C' or 'F95'), specifies the language of the
generated code. If ``None`` [default], the language is inferred based
upon the specified backend.
backend : string, optional
Backend used to wrap the generated code. Either 'f2py' [default],
or 'cython'.
tempdir : string, optional
Path to directory for temporary files. If this argument is supplied,
the generated code and the wrapper input files are left intact in the
specified path.
args : iterable, optional
An ordered iterable of symbols. Specifies the argument sequence for the
function.
flags : iterable, optional
Additional option flags that will be passed to the backend.
verbose : bool, optional
If True, autowrap will not mute the command line backends. This can be
helpful for debugging.
helpers : 3-tuple or iterable of 3-tuples, optional
Used to define auxiliary expressions needed for the main expr. If the
main expression needs to call a specialized function it should be
passed in via ``helpers``. Autowrap will then make sure that the
compiled main expression can link to the helper routine. Items should
be 3-tuples with (<function_name>, <sympy_expression>,
<argument_tuple>). It is mandatory to supply an argument sequence to
helper routines.
code_gen : CodeGen instance
An instance of a CodeGen subclass. Overrides ``language``.
include_dirs : [string]
A list of directories to search for C/C++ header files (in Unix form
for portability).
library_dirs : [string]
A list of directories to search for C/C++ libraries at link time.
libraries : [string]
A list of library names (not filenames or paths) to link against.
extra_compile_args : [string]
Any extra platform- and compiler-specific information to use when
compiling the source files in 'sources'. For platforms and compilers
where "command line" makes sense, this is typically a list of
command-line arguments, but for other platforms it could be anything.
extra_link_args : [string]
Any extra platform- and compiler-specific information to use when
linking object files together to create the extension (or to create a
new static Python interpreter). Similar interpretation as for
'extra_compile_args'.
Examples
========
>>> from sympy.abc import x, y, z
>>> from sympy.utilities.autowrap import autowrap
>>> expr = ((x - y + z)**(13)).expand()
>>> binary_func = autowrap(expr)
>>> binary_func(1, 4, 2)
-1.0
"""
if language:
if not isinstance(language, type):
_validate_backend_language(backend, language)
else:
language = _infer_language(backend)
# two cases 1) helpers is an iterable of 3-tuples and 2) helpers is a
# 3-tuple
if iterable(helpers) and len(helpers) != 0 and iterable(helpers[0]):
helpers = helpers if helpers else ()
else:
helpers = [helpers] if helpers else ()
args = list(args) if iterable(args, exclude=set) else args
if code_gen is None:
code_gen = get_code_generator(language, "autowrap")
CodeWrapperClass = {
'F2PY': F2PyCodeWrapper,
'CYTHON': CythonCodeWrapper,
'DUMMY': DummyWrapper
}[backend.upper()]
code_wrapper = CodeWrapperClass(code_gen, tempdir, flags if flags else (),
verbose, **kwargs)
helps = []
for name_h, expr_h, args_h in helpers:
helps.append(code_gen.routine(name_h, expr_h, args_h))
for name_h, expr_h, args_h in helpers:
if expr.has(expr_h):
name_h = binary_function(name_h, expr_h, backend='dummy')
expr = expr.subs(expr_h, name_h(*args_h))
try:
routine = code_gen.routine('autofunc', expr, args)
except CodeGenArgumentListError as e:
# if all missing arguments are for pure output, we simply attach them
# at the end and try again, because the wrappers will silently convert
# them to return values anyway.
new_args = []
for missing in e.missing_args:
if not isinstance(missing, OutputArgument):
raise
new_args.append(missing.name)
routine = code_gen.routine('autofunc', expr, args + new_args)
return code_wrapper.wrap_code(routine, helpers=helps)
def reduce_inequalities(inequalities, symbols=[]):
"""Reduce a system of inequalities with rational coefficients.
Examples
========
>>> from sympy import sympify as S, Symbol
>>> from sympy.abc import x, y
>>> from sympy.solvers.inequalities import reduce_inequalities
>>> reduce_inequalities(0 <= x + 3, [])
(-3 <= x) & (x < oo)
>>> reduce_inequalities(0 <= x + y*2 - 1, [x])
(x < oo) & (x >= 1 - 2*y)
"""
if not iterable(inequalities):
inequalities = [inequalities]
inequalities = [sympify(i) for i in inequalities]
gens = set().union(*[i.free_symbols for i in inequalities])
if not iterable(symbols):
symbols = [symbols]
symbols = (set(symbols) or gens) & gens
if any(i.is_extended_real == False for i in symbols):
raise TypeError(
filldedent('''
inequalities cannot contain symbols that are not real.
'''))
# make vanilla symbol real
recast = {
i: Dummy(i.name, extended_real=True)
for i in gens if i.is_extended_real is None
}
inequalities = [i.xreplace(recast) for i in inequalities]
symbols = {i.xreplace(recast) for i in symbols}
# prefilter
keep = []
for i in inequalities:
if isinstance(i, Relational):
i = i.func(i.lhs.as_expr() - i.rhs.as_expr(), 0)
elif i not in (True, False):
i = Eq(i, 0)
if i == True:
continue
elif i == False:
return S.false
if i.lhs.is_number:
raise NotImplementedError("could not determine truth value of %s" %
i)
keep.append(i)
inequalities = keep
del keep
# solve system
rv = _reduce_inequalities(inequalities, symbols)
# restore original symbols and return
return rv.xreplace({v: k for k, v in recast.items()})
def lambdify(args,
expr,
modules=None,
printer=None,
use_imps=True,
dummify=True):
"""
Returns a lambda function for fast calculation of numerical values.
If not specified differently by the user, SymPy functions are replaced as
far as possible by either python-math, numpy (if available) or mpmath
functions - exactly in this order. To change this behavior, the "modules"
argument can be used. It accepts:
- the strings "math", "mpmath", "numpy", "numexpr", "sympy"
- any modules (e.g. math)
- dictionaries that map names of sympy functions to arbitrary functions
- lists that contain a mix of the arguments above, with higher priority
given to entries appearing first.
The default behavior is to substitute all arguments in the provided
expression with dummy symbols. This allows for applied functions (e.g.
f(t)) to be supplied as arguments. Call the function with dummify=False if
dummy substitution is unwanted (and `args` is not a string). If you want
to view the lambdified function or provide "sympy" as the module, you
should probably set dummify=False.
For functions involving large array calculations, numexpr can provide a
significant speedup over numpy. Please note that the available functions
for numexpr are more limited than numpy but can be expanded with
implemented_function and user defined subclasses of Function. If specified,
numexpr may be the only option in modules. The official list of numexpr
functions can be found at:
https://github.com/pydata/numexpr#supported-functions
Deprecation Warnings
====================
In previous releases ``lambdify`` replaced ``Matrix`` with ``numpy.matrix``
by default. As of release 0.7.6 ``numpy.array`` is being transitioned to
the default. In release 0.7.7 this transition will be complete. For now, to
use the new default behavior you must pass in ``[{'ImmutableMatrix':
numpy.array}, 'numpy']`` to the ``modules`` kwarg.
>>> from sympy import lambdify, Matrix
>>> from sympy.abc import x, y
>>> import numpy
>>> mat2array = [{'ImmutableMatrix': numpy.array}, 'numpy']
>>> f = lambdify((x, y), Matrix([x, y]), modules=mat2array)
>>> f(1, 2)
array([[1],
[2]])
Usage
=====
(1) Use one of the provided modules:
>>> from sympy import sin, tan, gamma
>>> from sympy.utilities.lambdify import lambdastr
>>> from sympy.abc import x, y
>>> f = lambdify(x, sin(x), "math")
Attention: Functions that are not in the math module will throw a name
error when the lambda function is evaluated! So this would
be better:
>>> f = lambdify(x, sin(x)*gamma(x), ("math", "mpmath", "sympy"))
(2) Use some other module:
>>> import numpy
>>> f = lambdify((x,y), tan(x*y), numpy)
Attention: There are naming differences between numpy and sympy. So if
you simply take the numpy module, e.g. sympy.atan will not be
translated to numpy.arctan. Use the modified module instead
by passing the string "numpy":
>>> f = lambdify((x,y), tan(x*y), "numpy")
>>> f(1, 2)
-2.18503986326
>>> from numpy import array
>>> f(array([1, 2, 3]), array([2, 3, 5]))
[-2.18503986 -0.29100619 -0.8559934 ]
(3) Use a dictionary defining custom functions:
>>> def my_cool_function(x): return 'sin(%s) is cool' % x
>>> myfuncs = {"sin" : my_cool_function}
>>> f = lambdify(x, sin(x), myfuncs); f(1)
'sin(1) is cool'
Examples
========
>>> from sympy.utilities.lambdify import implemented_function
>>> from sympy import sqrt, sin, Matrix
>>> from sympy import Function
>>> from sympy.abc import w, x, y, z
>>> f = lambdify(x, x**2)
>>> f(2)
4
>>> f = lambdify((x, y, z), [z, y, x])
>>> f(1,2,3)
[3, 2, 1]
>>> f = lambdify(x, sqrt(x))
>>> f(4)
2.0
>>> f = lambdify((x, y), sin(x*y)**2)
>>> f(0, 5)
0.0
>>> row = lambdify((x, y), Matrix((x, x + y)).T, modules='sympy')
>>> row(1, 2)
Matrix([[1, 3]])
Tuple arguments are handled and the lambdified function should
be called with the same type of arguments as were used to create
the function.:
>>> f = lambdify((x, (y, z)), x + y)
>>> f(1, (2, 4))
3
A more robust way of handling this is to always work with flattened
arguments:
>>> from sympy.utilities.iterables import flatten
>>> args = w, (x, (y, z))
>>> vals = 1, (2, (3, 4))
>>> f = lambdify(flatten(args), w + x + y + z)
>>> f(*flatten(vals))
10
Functions present in `expr` can also carry their own numerical
implementations, in a callable attached to the ``_imp_``
attribute. Usually you attach this using the
``implemented_function`` factory:
>>> f = implemented_function(Function('f'), lambda x: x+1)
>>> func = lambdify(x, f(x))
>>> func(4)
5
``lambdify`` always prefers ``_imp_`` implementations to implementations
in other namespaces, unless the ``use_imps`` input parameter is False.
"""
from sympy.core.symbol import Symbol
from sympy.utilities.iterables import flatten
# If the user hasn't specified any modules, use what is available.
module_provided = True
if modules is None:
module_provided = False
# Use either numpy (if available) or python.math where possible.
# XXX: This leads to different behaviour on different systems and
# might be the reason for irreproducible errors.
modules = ["math", "mpmath", "sympy"]
#Attempt to import numpy
try:
_import("numpy")
except ImportError:
pass
else:
modules.insert(1, "numpy")
# Get the needed namespaces.
namespaces = []
# First find any function implementations
if use_imps:
namespaces.append(_imp_namespace(expr))
# Check for dict before iterating
if isinstance(modules, (dict, str)) or not hasattr(modules, '__iter__'):
namespaces.append(modules)
else:
# consistency check
if 'numexpr' in modules and len(modules) > 1:
raise TypeError("numexpr must be the only item in 'modules'")
namespaces += list(modules)
# fill namespace with first having highest priority
namespace = {}
for m in namespaces[::-1]:
buf = _get_namespace(m)
namespace.update(buf)
_issue_7853_dep_check(namespaces, namespace, expr)
if hasattr(expr, "atoms"):
#Try if you can extract symbols from the expression.
#Move on if expr.atoms in not implemented.
syms = expr.atoms(Symbol)
for term in syms:
namespace.update({str(term): term})
if 'numexpr' in namespaces and printer is None:
#XXX: This has to be done here because of circular imports
from sympy.printing.lambdarepr import NumExprPrinter as printer
# Get the names of the args, for creating a docstring
if not iterable(args):
args = (args, )
names = []
# Grab the callers frame, for getting the names by inspection (if needed)
callers_local_vars = inspect.currentframe().f_back.f_locals.items()
for n, var in enumerate(args):
if hasattr(var, 'name'):
names.append(var.name)
else:
# It's an iterable. Try to get name by inspection of calling frame.
name_list = [
var_name for var_name, var_val in callers_local_vars
if var_val is var
]
if len(name_list) == 1:
names.append(name_list[0])
else:
# Cannot infer name with certainty. arg_# will have to do.
names.append('arg_' + str(n))
# Create lambda function.
lstr = lambdastr(args, expr, printer=printer, dummify=dummify)
flat = '__flatten_args__'
if flat in lstr:
namespace.update({flat: flatten})
func = eval(lstr, namespace)
# Apply the docstring
sig = "func({0})".format(", ".join(str(i) for i in names))
sig = textwrap.fill(sig, subsequent_indent=' ' * 8)
expr_str = str(expr)
if len(expr_str) > 78:
expr_str = textwrap.wrap(expr_str, 75)[0] + '...'
func.__doc__ = ("Created with lambdify. Signature:\n\n{sig}\n\n"
"Expression:\n\n{expr}").format(sig=sig, expr=expr_str)
return func
def lambdastr(args, expr, printer=None, dummify=None):
"""
Returns a string that can be evaluated to a lambda function.
Examples
========
>>> from sympy.abc import x, y, z
>>> from sympy.utilities.lambdify import lambdastr
>>> lambdastr(x, x**2)
'lambda x: (x**2)'
>>> lambdastr((x,y,z), [z,y,x])
'lambda x,y,z: ([z, y, x])'
Although tuples may not appear as arguments to lambda in Python 3,
lambdastr will create a lambda function that will unpack the original
arguments so that nested arguments can be handled:
>>> lambdastr((x, (y, z)), x + y)
'lambda _0,_1: (lambda x,y,z: (x + y))(_0,_1[0],_1[1])'
"""
# Transforming everything to strings.
from sympy.matrices import DeferredVector
from sympy import Dummy, sympify, Symbol, Function, flatten, Derivative, Basic
if printer is not None:
if inspect.isfunction(printer):
lambdarepr = printer
else:
if inspect.isclass(printer):
lambdarepr = lambda expr: printer().doprint(expr)
else:
lambdarepr = lambda expr: printer.doprint(expr)
else:
#XXX: This has to be done here because of circular imports
from sympy.printing.lambdarepr import lambdarepr
def sub_args(args, dummies_dict):
if isinstance(args, string_types):
return args
elif isinstance(args, DeferredVector):
return str(args)
elif iterable(args):
dummies = flatten([sub_args(a, dummies_dict) for a in args])
return ",".join(str(a) for a in dummies)
else:
# replace these with Dummy symbols
if isinstance(args, (Function, Symbol, Derivative)):
dummies = Dummy()
dummies_dict.update({args: dummies})
return str(dummies)
else:
return str(args)
def sub_expr(expr, dummies_dict):
try:
expr = sympify(expr).xreplace(dummies_dict)
except Exception:
if isinstance(expr, DeferredVector):
pass
elif isinstance(expr, dict):
k = [sub_expr(sympify(a), dummies_dict) for a in expr.keys()]
v = [sub_expr(sympify(a), dummies_dict) for a in expr.values()]
expr = dict(zip(k, v))
elif isinstance(expr, tuple):
expr = tuple(sub_expr(sympify(a), dummies_dict) for a in expr)
elif isinstance(expr, list):
expr = [sub_expr(sympify(a), dummies_dict) for a in expr]
return expr
# Transform args
def isiter(l):
return iterable(l, exclude=(str, DeferredVector, NotIterable))
def flat_indexes(iterable):
n = 0
for el in iterable:
if isiter(el):
for ndeep in flat_indexes(el):
yield (n, ) + ndeep
else:
yield (n, )
n += 1
if dummify is None:
dummify = any(
isinstance(a, Basic) and a.atoms(Function, Derivative)
for a in (args if isiter(args) else [args]))
if isiter(args) and any(isiter(i) for i in args):
dum_args = [str(Dummy(str(i))) for i in range(len(args))]
indexed_args = ','.join([
dum_args[ind[0]] + ''.join(["[%s]" % k for k in ind[1:]])
for ind in flat_indexes(args)
])
lstr = lambdastr(flatten(args), expr, printer=printer, dummify=dummify)
return 'lambda %s: (%s)(%s)' % (','.join(dum_args), lstr, indexed_args)
dummies_dict = {}
if dummify:
args = sub_args(args, dummies_dict)
else:
if isinstance(args, string_types):
pass
elif iterable(args, exclude=DeferredVector):
args = ",".join(str(a) for a in args)
# Transform expr
if dummify:
if isinstance(expr, string_types):
pass
else:
expr = sub_expr(expr, dummies_dict)
expr = lambdarepr(expr)
return "lambda %s: (%s)" % (args, expr)
def _solve(f, *symbols, **flags):
""" Return a checked solution for f in terms of one or more of the symbols."""
if not iterable(f):
if len(symbols) != 1:
soln = None
free = f.free_symbols
ex = free - set(symbols)
if len(ex) == 1:
ex = ex.pop()
try:
# may come back as dict or list (if non-linear)
soln = solve_undetermined_coeffs(f, symbols, ex)
except NotImplementedError:
pass
if not soln is None:
return soln
# find first successful solution
failed = []
for s in symbols:
n, d = solve_linear(f, x=[s])
if n.is_Symbol:
soln = {n: cancel(d)}
return soln
failed.append(s)
for s in failed:
try:
soln = _solve(f, s, **flags)
return soln
except NotImplementedError:
pass
else:
msg = "No algorithms are implemented to solve equation %s"
raise NotImplementedError(msg % f)
symbol = symbols[0]
# first see if it really depends on symbol and whether there
# is a linear solution
f_num, sol = solve_linear(f, x=symbols)
if not symbol in f_num.free_symbols:
return []
elif f_num.is_Symbol:
return [cancel(sol)]
strategy = guess_solve_strategy(f, symbol)
result = False # no solution was obtained
if strategy == GS_POLY:
poly = f.as_poly(symbol)
if poly is None:
msg = "Cannot solve equation %s for %s" % (f, symbol)
else:
# for cubics and quartics, if the flag wasn't set, DON'T do it
# by default since the results are quite long. Perhaps one could
# base this decision on a certain critical length of the roots.
if poly.degree() > 2:
flags['simplified'] = flags.get('simplified', False)
result = roots(poly, cubics=True, quartics=True).keys()
elif strategy == GS_RATIONAL:
P, _ = f.as_numer_denom()
dens = denoms(f, x=symbols)
try:
soln = _solve(P, symbol, **flags)
except NotImplementedError:
msg = "Cannot solve equation %s for %s" % (P, symbol)
result = []
else:
if dens:
# reject any result that makes any denom. affirmatively 0;
# if in doubt, keep it
result = [
s for s in soln
if all(not checksol(den, {symbol: s}) for den in dens)
]
else:
result = soln
elif strategy == GS_POLY_CV_1:
args = list(f.args)
if isinstance(f, Pow):
result = _solve(args[0], symbol, **flags)
elif isinstance(f, Add):
# we must search for a suitable change of variables
# collect exponents
exponents_denom = list()
for arg in args:
if isinstance(arg, Pow):
exponents_denom.append(arg.exp.q)
elif isinstance(arg, Mul):
for mul_arg in arg.args:
if isinstance(mul_arg, Pow):
exponents_denom.append(mul_arg.exp.q)
assert len(exponents_denom) > 0
if len(exponents_denom) == 1:
m = exponents_denom[0]
else:
# get the LCM of the denominators
m = reduce(ilcm, exponents_denom)
# x -> y**m.
# we assume positive for simplification purposes
t = Dummy('t', positive=True)
f_ = f.subs(symbol, t**m)
if guess_solve_strategy(f_, t) != GS_POLY:
msg = "Could not convert to a polynomial equation: %s" % f_
result = []
else:
soln = [s**m for s in _solve(f_, t)]
# we might have introduced solutions from another branch
# when changing variables; check and keep solutions
# unless they definitely aren't a solution
result = [
s for s in soln
if checksol(f, {symbol: s}) is not False
]
elif isinstance(f, Mul):
result = []
for m in f.args:
result.extend(_solve(m, symbol, **flags) or [])
elif strategy == GS_POLY_CV_2:
m = 0
args = list(f.args)
if isinstance(f, Add):
for arg in args:
if isinstance(arg, Pow):
m = min(m, arg.exp)
elif isinstance(arg, Mul):
for mul_arg in arg.args:
if isinstance(mul_arg, Pow):
m = min(m, mul_arg.exp)
elif isinstance(f, Mul):
for mul_arg in args:
if isinstance(mul_arg, Pow):
m = min(m, mul_arg.exp)
if m and m != 1:
f_ = simplify(f * symbol**(-m))
try:
sols = _solve(f_, symbol)
except NotImplementedError:
msg = 'Could not solve %s for %s' % (f_, symbol)
else:
# we might have introduced unwanted solutions
# when multiplying by x**-m; check and keep solutions
# unless they definitely aren't a solution
if sols:
result = [
s for s in sols
if checksol(f, {symbol: s}) is not False
]
else:
msg = 'CV_2 calculated %d but it should have been other than 0 or 1' % m
elif strategy == GS_PIECEWISE:
result = set()
for expr, cond in f.args:
candidates = _solve(expr, *symbols)
if isinstance(cond, bool) or cond.is_Number:
if not cond:
continue
# Only include solutions that do not match the condition
# of any of the other pieces.
for candidate in candidates:
matches_other_piece = False
for other_expr, other_cond in f.args:
if isinstance(other_cond, bool) \
or other_cond.is_Number:
continue
if bool(other_cond.subs(symbol, candidate)):
matches_other_piece = True
break
if not matches_other_piece:
result.add(candidate)
else:
for candidate in candidates:
if bool(cond.subs(symbol, candidate)):
result.add(candidate)
result = list(result)
elif strategy == -1:
raise ValueError('Could not parse expression %s' % f)
# this is the fallback for not getting any other solution
if result is False or strategy == GS_TRANSCENDENTAL:
soln = tsolve(f_num, symbol)
dens = denoms(f, x=symbols)
if not dens:
result = soln
else:
# reject any result that makes any denom. affirmatively 0;
# if in doubt, keep it
result = [
s for s in soln
if all(not checksol(den, {symbol: s}) for den in dens)
]
if result is False:
raise NotImplementedError(
msg +
"\nNo algorithms are implemented to solve equation %s" % f)
if flags.get('simplified', True) and strategy != GS_RATIONAL:
result = map(simplify, result)
return result
else:
if not f:
return []
else:
polys = []
for g in f:
poly = g.as_poly(*symbols, **{'extension': True})
if poly is not None:
polys.append(poly)
else:
raise NotImplementedError()
if all(p.is_linear for p in polys):
n, m = len(f), len(symbols)
matrix = zeros((n, m + 1))
for i, poly in enumerate(polys):
for monom, coeff in poly.terms():
try:
j = list(monom).index(1)
matrix[i, j] = coeff
except ValueError:
matrix[i, m] = -coeff
# a dictionary of symbols: values or None
result = solve_linear_system(matrix, *symbols, **flags)
return result
else:
# a list of tuples, T, where T[i] [j] corresponds to the ith solution for symbols[j]
result = solve_poly_system(polys)
return result