[Tutor] How to set up an Array?

[John Fouhy]

2009/5/12 nickel flipper <nickelflipper at yahoo.com>:
> sfr (key=PORTA addr=0xf80 size=1 access='rw rw rw u rw rw rw rw')
>    reset (por='xxxxxxxx' mclr='uuuuuuuu')
>    bit (names='RA7 RA6 RA5 - RA3 RA2 RA1 RA0' width='1 1 1 1 1 1 1 1')
>    bit (tag=scl names='RA' width='8')
>    bit (names='OSC1 OSC2 AN4 - AN3 AN2 AN1 AN0' width='1 1 1 1 1 1 1 1')
>    bit (names='CLKI CLKO nSS1 - VREF_PLUS VREF_MINUS C2INA C1INA' width='1 1 1 1 1 1 1 1')
>    bit (names='- - LVDIN - C1INB CVREF_MINUS PMPA7 PMPA6' width='1 1 1 1 1 1 1 1')
>    bit (names='- - RCV - - C2INB RP1 RP0' width='1 1 1 1 1 1 1 1')
>    bit (names='- - RP2 - - - - -' width='1 1 1 1 1 1 1 1')

Hmm, hairy!

You could try looking into regular expressions.  Let's see:

    \b\w+='.+?'

That should be a regular expression matching the beginning of a word,
followed by one or more word characters, an =, a ', some characters,
up to the first '.

i.e. we're trying to match things like "names='RA7 RA6 RA5 - RA3 RA2 RA1 RA0'"

So, here we go:

>>> import re
>>> s = "   bit (names='RA7 RA6 RA5 - RA3 RA2 RA1 RA0' width='1 1 1 1 1 1 1 1')"
>>> rex = re.compile(r"\b\w+='.+?'")

(note the r before the string.. that makes it a "raw" string and
prevents backslashes from causing problems, mostly)

>>> def to_dict(s):
...   res = {}
...   for token in rex.findall(s):
...     word, rest = token.split('=')
...     rest = rest.replace("'", '')
...     res[word] = rest.split()
...   return res
...

(note that my regular expression, rex, is here effectively a global
variable.  That's convenient in the interactive interpreter, but you
may want to do that a little differently in your final script)

>>> to_dict(s)
{'width': ['1', '1', '1', '1', '1', '1', '1', '1'], 'names': ['RA7',
'RA6', 'RA5', '-', 'RA3', 'RA2', 'RA1', 'RA0']}

If regular expressions aren't powerful enough, you could look into a
full-fledged parsing library.  There are two I know of: simpleparse
and pyparsing.  SimpleParse works well if you're familiar with writing
grammars: you write a grammar in the usual style, and simpleparse
makes a parser out of it.  pyparsing is more OO.

-- 
John.