Hi,
On 29/08/12 20:29, Ronan Lamy wrote:
>>
>>>>>> dir(S)
>>> ['Catalan', 'ComplexInfinity', 'EmptySet', 'EulerGamma', 'Exp1',
>>> 'GoldenRatio', 'Half', 'IdentityFunction', 'ImaginaryUnit',
>>> 'Infinity', 'Integers', 'NaN', 'Naturals', 'NegativeInfinity',
>>> 'NegativeOne', 'NumberSymbol', 'One', 'Pi', 'Reals', 'UniversalSet',
>>> 'Zero', '__call__', '__class__', '__delattr__', '__doc__',
>>> '__format__', '__getattribute__', '__hash__', '__init__',
>>> '__module__', '__new__', '__reduce__', '__reduce_ex__', '__repr__',
>>> '__setattr__', '__sizeof__', '__slots__', '__str__',
>>> '__subclasshook__']
>> Thanks!
>>
>> So, dir(S) is the preferred way of getting a complete list :)
>>
>> It also works in both old and new versions, so that's good. I think the
>> only remaining question is if this is likely to stay working for the
>> foreseeable future?
>
> Actually, if you use sympify(), almost everything in the sympy
> namespace is handled specially ('C' is probably the only significant
> exception, for obscure reasons), so IIUC the list you want is
> dir(sympy). The way parsing works now is hackish and probably unsafe,
> so don't count on it remaining the same.
Ah, thanks.
dir(sympy) looks applicable at least on first glance.
I think you did understand correctly - the aim is to obtain a list of
reserved strings which sympify() will parse into something else than a
generic Symbol. This list is then used to ensure that any user-defined
quantities (in the FEM solver) will parse as expected, before the
user-given expressions are given to SymPy.
It seems, though, that just the string is not enough information. For
example, in 0.7.x there is the LT() function (to get the leading term),
but e.g. the following is nevertheless valid:
expr = sy.sympify("LT*a") # no error, expr will have Symbols "LT" and "a"
# artificial example: let's replace the symbol "LT" by sin(x)
my_LT = sy.symbols("LT") # refer to the Symbol "LT", not function LT()
expr.subs( {my_LT : sy.sympify("sin(x)")} )
So maybe functions should be ignored? But on the third hand,
expr = sy.sympify("sin*a")
triggers a TypeError; "sin" is recognized as the standard sine function
even if it has no parentheses to denote a function call. (At least in
0.6.7; I have 0.7.1 only at home, so can't check it right now.)
So it seems there is no general rule?
(The way I ran into this was that when developing with 0.6.x, I happened
to decide that I would use the prefix "L" to denote the Laplacian of a
quantity in my equation processor. Then, running the code on 0.7.1
later, with a quantity named "T", produced a surprise...)
>>>> And finally, a related question: is there a way to get a
>>>> human-readable
>>>> description for a reserved name (e.g. "zoo" -> "complex infinity")? I
>>>> didn't find anything in the API for this.
>>> Not sure how to go further than
>>>
>>>>>> type(oo)
>>> <class 'sympy.core.numbers.Infinity'>
>> Aaaahh, the type! Nice.
>>
>> Thanks!
> You can also access the attributes of the type, e.g. type(oo).__name__
Ok. Thanks for the tip.
>> The final code that I ended up using (modulo variable naming and
>> comments) is:
>>
>> ---8<---8<---8<---
>>
>> rawnames = filter( lambda name: re.match("__", name) is None,
>> dir(sy.S) )
>> objs = map( lambda name: eval("sy.S.%s" % name), rawnames )
>> objs = filter( lambda obj: not obj.is_Number, objs )
>> strs = map( lambda obj: str(obj), objs )
> Using eval() is dangerous, and should be avoided. Here you just need
> getattr(sy.S, name).
Ah, of course. Thanks.
> BTW, whenever you want to put a lambda inside a map, you should really
> use a list comprehension instead, as it's much more readable, and can
> be combined with a filter, as in:
>
> objs = [getattr(sy.S, name) for name in dis(sy.S) if not
> name.startswith("__")]
> strs = [str(obj) for obj in objs if not obj.is_Number]
>
> (NB: I don't think this is exactly the right code, cf. my first comment)
Ok. I think both ways are equally readable, but if the list
comprehension is more Pythonic, then why not.
Thanks for this tip, too. I've been programming for a long time, but
only started Python recently :)
-J
--
You received this message because you are subscribed to the Google Groups
"sympy" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/sympy?hl=en.
