quiz about symbolic manipulation

[Michael Hudson]

bokr at oz.net (Bengt Richter) writes:

> Well, for a toy problem, parsing according to python grammar is probably
> overkill. I.e.,
> 
> 'square(square(x+y)+z)+square(x+w)'
> 
> becomes
> 
> ['eval_input',
>  ['testlist',
>   ['test',
>    ['and_test',
>     ['not_test',
>      ['comparison',
>       ['expr',
>        ['xor_expr',
[... and on and on and on ...]

Bengt probably knows this, but you can use the compiler package to
make this very much pleasanter...

>>> compiler.transformer.parse('square(square(x+y)+z)+square(x+w)')
Module(None, Stmt([Discard(Add((CallFunc(Name('square'),
[Add((CallFunc(Name('square'), [Add((Name('x'), Name('y')))], None,
None), Name('z')))], None, None), CallFunc(Name('square'),
[Add((Name('x'), Name('w')))], None, None))))]))
>>> _.node.nodes[0].expr
Add((CallFunc(Name('square'), [Add((CallFunc(Name('square'),
[Add((Name('x'), Name('y')))], None, None), Name('z')))], None, None),
CallFunc(Name('square'), [Add((Name('x'), Name('w')))], None, None)))

Hmm, well the fact that pprint.pprint doesn't know how to deal with
these thigns doesn't help, but these really are nicer to work with :)

Cheers,
M.

-- 
  Famous remarks are very seldom quoted correctly.
                                                    -- Simeon Strunsky