Thread: [Pyparsing] pyparsing svg - noob
Brought to you by:
ptmcg
Menu
▾
▴
pyparsing-users
|
From: Donn I. <don...@gm...> - 2007-11-08 09:55:20
|
Hi - I have been trying, but I need some help:
Here's my test code to parse an SVG path element, can anyone steer me right?
d1="""
M 209.12237 , 172.2415
L 286.76739 , 153.51369
L 286.76739 , 275.88534
L 209.12237 , 294.45058
L 209.12237 , 172.2415
z """
#Try it with no enters
d1="""M 209.12237,172.2415 L 286.76739,153.51369 L 286.76739,275.88534 L
209.12237,294.45058 L 209.12237,172.2415 z """
#For later, has more commands
d2="""
M 269.78326 , 381.27104
C 368.52151 , 424.27023
90.593578 , -18.581883
90.027729 , 129.28708
C 89.461878 , 277.15604
171.04501 , 338.27184
269.78326 , 381.27104
z """
## word :: M, L, C, Z
## number :: group of numbers
## dot :: "."
## comma :: ","
## couple :: number dot number comma number dot number
## command :: word
##
## phrase :: command couple
from pyparsing import Word, Literal, alphas, nums, Optional, OneOrMore
command = Word("MLCZ")
comma = Literal(",")
dot = Literal(".")
float = nums + dot + nums
couple = float + comma + float
phrase = OneOrMore(command + OneOrMore(couple | ( couple + couple ) ) )
print phrase
print phrase.parseString(d1.upper())
Thanks
\d
|
|
From: Donn <don...@gm...> - 2007-11-08 10:24:31
|
Paul - thanks again for the accurate help.
> phrase = OneOrMore(command("command") + Group(OneOrMore(couple))("coords")
> )
Gives me an error:
phrase = OneOrMore(command("command") + Group(OneOrMore(couple))
("coords") )
TypeError: 'MatchFirst' object is not callable
But I have not had a chance to get the latest pyparsing installed yet. Using
your other suggestions seems to be working, so I'll keep hacking.
Best,
\d
|
|
From: Ralph C. <ra...@in...> - 2007-11-08 10:29:21
|
Hi Donn,
> from pyparsing import Word, Literal, alphas, nums, Optional, OneOrMore
>
> command = Word("MLCZ")
> comma = Literal(",")
> dot = Literal(".")
> float = nums + dot + nums
> couple = float + comma + float
>
> phrase = OneOrMore(command + OneOrMore(couple | ( couple + couple ) ) )
nums is just str('0123456789') so you're saying that you literally want
'0123456789.0123456789' for every float. Change it to
float = Word(nums) + dot + Word(nums)
and all will be fine.
You may find it's better to check you've the right number of couples
depending on which command it is though, e.g. the grammar says
move = Literal('M') + couple
command = ... | move | ...
Cheers,
Ralph.
|
|
From: Donn I. <don...@gm...> - 2007-11-08 10:43:45
|
Thanks for the help all. I've got a working version now. \d |
|
From: Donn <don...@gm...> - 2007-11-08 11:34:35
|
Paul,
Sorry to bug you. Two things:
1. I am having a conceptual issue with the "list" returned by parseString. I
can't seem to use it as a normal list. Mainly I like to while len(l)>0: and
then l.pop(0) to process it but I can't seem to pop this list.
2. It seems the optional closing "Z" is being ignored.
dot = Literal(".")
float = Combine(Optional("-") + Word(nums) + dot + Word(nums))
command = oneOf("M L C Z")
comma = Literal(",").suppress()
couple = Group(float + comma + float)
phrase = OneOrMore(command + Group(OneOrMore(couple)) )
#No C commands in this one...
d1="""M 209.12237,172.2415 L 286.76739,153.51369 L 286.76739,275.88534 L
209.12237,294.45058 L 209.12237,172.2415 z """
paths = []
pl = phrase.parseString(d1.upper())
while len(pl) > 0:
something = pl.pop(0)
print something
(I get : TypeError: 'str' object is not callable)
I could use <for something in pl:> but since some of the coord lists are
longer than others and not grouped ideally, it gets hairy.
The ideal grouping would be something like this (I mix C and L commands)
['M', [float,float],
'C',[float,float,float,float,float,float],
'L',[float,float],
'Z']
So I can just pop them off one by one.
\d
|
|
From: Donn I. <don...@gm...> - 2007-11-08 13:46:04
|
Yeah, related to prev post, still no wiser.
I'm having trouble getting a float out of a variable:
It contains:
Combine:({["-"] W:(0123...) "." W:(0123...)})
And it's type is:
<class 'pyparsing.Combine'>
But float(v) does not return 123.456 it returns the above.
/d
|
|
From: Paul M. <pa...@al...> - 2007-11-08 15:01:59
|
Donn -
1. Here is why you are having trouble getting floats out of a parsed string:
float = ... whatever ...
You are redefining the global float!
Change
float = Combine(Optional("-") + Word(nums) + dot + Word(nums))
to
float_ = Combine(Optional("-") + Word(nums) + dot + Word(nums))
or
realNum = Combine(Optional("-") + Word(nums) + dot + Word(nums))
or
floatValue = Combine(Optional("-") + Word(nums) + dot + Word(nums))
Call it anything but float!
2. It seems the optional closing "Z" is being ignored.
dot = Literal(".")
float = Combine(Optional("-") + Word(nums) + dot + Word(nums))
command = oneOf("M L C Z")
comma = Literal(",").suppress()
couple = Group(float + comma + float)
phrase = OneOrMore(command + Group(OneOrMore(couple)) )
Well this is because a phrase is made up of a repetition of (command +
couple...), and there are no couples after the 'Z'.
You later wrote:
The ideal grouping would be something like this (I mix C and L commands)
['M', [float,float],
'C',[float,float,float,float,float,float],
'L',[float,float],
'Z']
Well why didn't you say so? :) In that case, Ralph Corderoy's suggestion
works better:
M_command = "M" + Group(couple)
C_command = "C" + Group(couple + couple + couple)
L_command = "L" + Group(couple)
Z_command = "Z"
command = M_command | C_command | L_command | Z_command
phrase = OneOrMore(command) # what would generate your preferred list
phrase = OneOrMore(Group(command)) # what I would recommend instead
My recommended version of phrase makes each command its own nested list,
like:
[['M', [float,float]],
['C',[float,float,float,float,float,float]],
['L',[float,float]],
['Z']]
Now you *can* iterate cleanly over this list, since the variable-length
couple lists are enclosed in a list with their respective command letter.
for command,couples in tokens[:-1]:
print command
print couples.asList()
-- Paul
|
|
From: Donn I. <don...@gm...> - 2007-11-08 13:50:05
|
> 2. It seems the optional closing "Z" is being ignored. Fixed this with + Optional(command) on the end. \d |
|
From: Paul M. <pa...@al...> - 2007-11-08 15:12:59
|
You are right, parseString really returns an object, ParseResults.
ParseResults tries to behave much like a list, but adds some dict-like and
object-like behavior too. I see I overlooked pop() in ParseResults'
methods, I'll try to remember to include that in the next version. (You may
have to resort to reading some documentation soon...)
If you really, really, really want to pop your way through the tokens, you
can convert the results to a list by using asList(), and then pop() away!
I forgot to mention in the other e-mail:
Another recommended technique is to do conversions like string-to-float
during the parsing process itself, by adding a parse action. Let's assume
you changed the variable name float to floatNum:
floatNum = Combine(Optional("-") + Word(nums) + dot + Word(nums))
Now add:
def convertToFloat(tokens):
return float(tokens[0])
floatNum.setParseAction(convertToFloat)
Or if you want to make this a one-liner (more compact, but a lot to take in
at once when first learning):
floatNum.setParseAction(lambda toks:float(toks[0]))
Now when you process the resulting tokens, the floats will already be
floating point numbers instead of strings, no need to convert them
after-the-fact.
-- Paul
-----Original Message-----
From: pyp...@li...
[mailto:pyp...@li...] On Behalf Of Donn
Sent: Thursday, November 08, 2007 5:39 AM
To: Paul McGuire
Cc: pyp...@li...
Subject: Re: [Pyparsing] pyparsing svg - noob
Paul,
Sorry to bug you. Two things:
1. I am having a conceptual issue with the "list" returned by parseString. I
can't seem to use it as a normal list. Mainly I like to while len(l)>0: and
then l.pop(0) to process it but I can't seem to pop this list.
2. It seems the optional closing "Z" is being ignored.
dot = Literal(".")
float = Combine(Optional("-") + Word(nums) + dot + Word(nums))
command = oneOf("M L C Z")
comma = Literal(",").suppress()
couple = Group(float + comma + float)
phrase = OneOrMore(command + Group(OneOrMore(couple)) )
#No C commands in this one...
d1="""M 209.12237,172.2415 L 286.76739,153.51369 L 286.76739,275.88534 L
209.12237,294.45058 L 209.12237,172.2415 z """
paths = []
pl = phrase.parseString(d1.upper())
while len(pl) > 0:
something = pl.pop(0)
print something
(I get : TypeError: 'str' object is not callable)
I could use <for something in pl:> but since some of the coord lists are
longer than others and not grouped ideally, it gets hairy.
The ideal grouping would be something like this (I mix C and L commands)
['M', [float,float],
'C',[float,float,float,float,float,float],
'L',[float,float],
'Z']
So I can just pop them off one by one.
\d
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Pyparsing-users mailing list
Pyp...@li...
https://lists.sourceforge.net/lists/listinfo/pyparsing-users
|
|
From: Donn <don...@gm...> - 2007-11-08 15:22:11
|
> object-like behavior too. I see I overlooked pop() in ParseResults' > methods, I'll try to remember to include that in the next version. Thanks. > (You may have to resort to reading some documentation soon...) I hear you. You've been magnificently patient with me :) > If you really, really, really want to pop your way through the tokens, you > can convert the results to a list by using asList(), and then pop() away! Great! It's an approach my mind likes - it lets me process things logically while in a loop that simply gets shorter as it loops! > floatNum.setParseAction(lambda toks:float(toks[0])) This and your other example are great ideas, I'll give 'em a whirl. I hope I don't hit many more walls, I'm actually aiming at other things and all this parsing kind of hit me over the head. I'll have to change gears and slow down. Thanks again. \d |
|
From: Donn I. <don...@gm...> - 2007-11-08 15:18:37
|
> float = ... whatever ... AAARRRGGGHHH! :) A Homer moment. > phrase = OneOrMore(Group(command)) # what I would recommend instead > My recommended version of phrase makes each command its own nested list, > like: > [['M', [float,float]], > ['C',[float,float,float,float,float,float]], > ['L',[float,float]], > ['Z']] > Your recc. phrase is producing this: M [['269.78326', '381.27104']] This is ok. C [['368.52151', '424.27023'], ['90.593578', '-18.581883'], ['90.027729', '129.28708']] This is 3 lists of 2 coords. Not a train smash, but not a single list of 6 coords. It's to do with the nature of a 'couple' I think. > for command,couples in tokens[:-1]: Now that is cool, but strange. I printed tokens[:-1] out and, oh boy, there's tupels in lists and dics too! Heck, as long as it works. \d |
|
From: Paul M. <pa...@al...> - 2007-11-08 15:24:43
|
I assumed that the couple always referred to an x-y coordinate pair, which is why I grouped them. If you don't wan the grouping, just make couple read: couple = floatNum + comma + floatNum and those inner groups will go away. -- Paul -----Original Message----- From: pyp...@li... [mailto:pyp...@li...] On Behalf Of Donn Ingle Sent: Thursday, November 08, 2007 9:23 AM To: pyp...@li... Subject: Re: [Pyparsing] pyparsing svg - noob > float = ... whatever ... AAARRRGGGHHH! :) A Homer moment. > phrase = OneOrMore(Group(command)) # what I would recommend instead > My recommended version of phrase makes each command its own nested > list, > like: > [['M', [float,float]], > ['C',[float,float,float,float,float,float]], > ['L',[float,float]], > ['Z']] > Your recc. phrase is producing this: M [['269.78326', '381.27104']] This is ok. C [['368.52151', '424.27023'], ['90.593578', '-18.581883'], ['90.027729', '129.28708']] This is 3 lists of 2 coords. Not a train smash, but not a single list of 6 coords. It's to do with the nature of a 'couple' I think. > for command,couples in tokens[:-1]: Now that is cool, but strange. I printed tokens[:-1] out and, oh boy, there's tupels in lists and dics too! Heck, as long as it works. \d ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Pyparsing-users mailing list Pyp...@li... https://lists.sourceforge.net/lists/listinfo/pyparsing-users |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X