At least once it was stated that the preferred way to get a named
object is to subclass from Symbol and basically use Symbol as a string
class.
This is so because many recursion algorithms in SymPy assume that
every element of args subclasses Basic. This is a reasonable
requirement for all objects except the leafs of the tree that have
names. A hack around this is the way that Symbol does not contain its
name string in its args and thus it fails on obj.func(*obj.args)==obj.
And now everybody who needs a name string resorts to encapsulating it
in a Symbol.
The root of the problem is that there is no notion of an atom or a
leaf **consistently** used in sympy. There are the classes Atom and
the property is_atom, however very few of the algorithms in sympy
actually check those.
I have raised this issue a few times already, however it got stalled.
Can we settle on a definition for an atom that works including for
named objects (i.e. objects that need to have strings in their args).
What about defining a function atom() in utilities that does something like:
def atom(obj):
try:
return bool(obj.args) and obj.is_Atom
# captures Atom instances and
# shortcircuits on empty args
except AttributeError:
return False # strings work
Then each time an issue arises, we find the offending recursive
algorithm and reimplement it with atom(). Step by step all of sympy
will have consistent recursion routines.
--
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.
