Pyparsing...

[Larry Bates]

Note: The Cellheader layout and the cell data
layout don't appear to match properly (when
compared to the data you show in your sample
dictionary).  My solution follows the cellheader
layout.

The format of your data file is perfect to be
parsed with ConfigParser with [QC1] as section
name and Type, NumberCells, etc. as options.

import ConfigParser

inputfilename='data.ini' # Insert input filename
INI=ConfigParser.ConfigParser()
INI.read(inputfilename)
data={'type': None, 'numbercells': None, 'table':{}}

section='QC1'
option='type'
try: data['type']=INI.getint(section, option)
except:
    #
    # Insert code to handle missing type option
    #
    pass

option='numbercells'
try: data['numbercells']=INI.getint(section, option)
except:
    #
    # Insert code to handle missing numbercells option
    #
    pass

option='cellheader'
try: data['cellheader']=INI.get(section, option)
except:
    #
    # Insert code to handle missing numbercells option
    #
    pass

CELLS=[x for x in INI.options(section)
       if x.startswith('cell')]

#
# Must get rid of 'cellheader' or maybe change the key name?
#
CELLS=[x for x in CELLS if x != 'cellheader']

celldatalist=[]
for CELL in CELLS:
    celldata={}
    x, y, probe, plen, atom, index=INI.get(section, CELL).split(' ')
    celldata['cell']=int(CELL[4:])
    celldata['x']=int(x)
    celldata['y']=int(y)
    celldata['plen']=plen
    celldata['atom']=int(atom)
    celldata['index']=int(index)
    celldatalist.append(celldata)

data['table']=celldatalist


This is tested so it should be close (I do this quite a lot in
my code).

You could wrap a loop around the outside of this if you have
multiple QC instances.

Hope it helps.
Larry Bates



"Raoul" <raoulsam at yahoo.com> wrote in message 
news:7b22ae5b.0409201639.7837fca9 at posting.google.com...
>I am futzing with pyparsing and for the most part enjoying it.
> However, I'm running into trouble with whitespace delimited lists. I
> get data in blocks like this:
>
> [QC1]
> Type=15
> NumberCells=1925
> CellHeader=X Y PROBE PLEN ATOM INDEX
> Cell1=132 0 N 25 0 132
> Cell2=652 0 N 25 0 652
> Cell3=648 0 N 25 0 648
> ...
>
> I'd like to be able to parse this structure.
>
> Ideally, I'd like for a QC node, to have a dictionary with
> {'number':1
> 'type' : 15
> 'NumberCells' : 1925
> 'Table' : [{'cell':1,'x':132,'y':0,'probe':25,'plen':0,'atom':132',
> 'index':None}, {'cell':2 ....
>
> I'm running into the following problems:
>
> 1. I can't seem to use delimitList() to define a rule that parses the
> right hand side of the table into
> ['x','y','probe','plen','atom','index']. I think it's because my lists
> are whitespace delimited.
>
> 2. I can't seem to convert value into an integer, for example, I can
> parse each row in the table to :
>  ['Cell','2','=', '652 0 N 25 0 652']
> but am unable to get the setParseAction(see below) to convert and
> substitute in the right value.
>
> Any hints will help a great deal. Thanks...
>
> Raoul-Sam
>
>
> I have some ugly non functional code below..
>
> def cdffile_BNF():
>    global cdfbnf
>
>    if not cdfbnf:
>        makeint = Word(nums).setParseAction( lambda s,l,t:[int(t[0])])
>        equals = Literal("=").suppress()
>        nonequals = "".join( [ c for c in printables if c != "=" ] ) +
> " \t"
>
>        key = Word(nonequals)
>        value = Word(nonequals)
>        kvp = Group(key + equals + restOfLine)
>        kvpBlk = OneOrMore(kvp)
>
>        headerCell = delimitedList(Word(alphanums)," ")
>        rowHeader = Combine( Literal("CellHeader") + equals +
> headerCell)
>        row = Combine(Literal("Cell").suppress() + restOfLine)
>        rows = OneOrMore(row)
>
>        CDF = Literal("[CDF]")
>        CDFBlk = Group(CDF + kvpBlk)
>
>        CHIP = Literal("[CHIP]")
>        CHIPBlk = Group(CHIP + kvpBlk)
>        CHIPBlk.setResultsName("chip")
>
>        QC = Combine( Literal("[QC").suppress() + Word(nums) +
> Literal("]").suppress())
>        QCBlk = QC + kvp + kvp + rowHeader + rows
>
>        cdfbnf = CDFBlk + CHIPBlk + QCBlk
>
>    return cdfbnf