Logic expression

[dalek-who]

What a nice tools! However, current codes has trouble dealing with logic expression with more than 2 terms, for example:

x, y, z = sympy.symbols("x,y,z")
expr =x  | y | z 
model = SymPyModule(expressions=[expr])
terms = {
        "x": torch.randint(0, 2, (10,)).bool(),
        "y": torch.randint(0, 2, (10,)).bool(),
        "z": torch.randint(0, 2, (10,)).bool(),
}
result = model(**terms)

Then it will raise an error:

TypeError: logical_or() takes 2 positional arguments but 3 were given

Here's the solution: modify

sympytorch/sympytorch/sympy_module.py

Lines 52 to 53 in ca3e3f4

	sympy.And: torch.logical_and,
	sympy.Or: torch.logical_or,

to

sympy.And: _reduce(torch.logical_and),
sympy.Or: _reduce(torch.logical_or),

, which is the same as mul and add.

Also, besides the modification above, I suggest two more improvements:

1. make this function public
   

   sympytorch/sympytorch/sympy_module.py

   Lines 7 to 10 in ca3e3f4

   	def _reduce(fn):
   	def fn_(*args):
   	return ft.reduce(fn, args)
   	return fn_

2. modify
   

   sympytorch/sympytorch/sympy_module.py

   Line 158 in ca3e3f4

   _func_lookup = co.ChainMap(_global_func_lookup, extra_funcs)

   
   to

_func_lookup = co.ChainMap(extra_funcs, _global_func_lookup)

so that users can overload default lookup tables.

[patrick-kidger]

That all sounds reasonable to me! I'd happily accept a PR on these.

(I probably wouldn't bother making reduce public though, that's a pretty small helper function.)