next »

[taltas]

Hi all,

I am trying to the following but I really struggling.
I am using python 2.6 I cannot import ngrams as well.

My problem is i have string
st = 'A quick brown fox jumped over 007 wink* wink*'

I am trying to get bigrams using regular expressions ie:

'qu' 'ui' 'ic' 'ck' 'br' 'ro' 'ow' 'wn' 'fo' 'ox' 'ju' 'um' 'mp' 'pe' 'ed' 'ov' 've' 'er' '00' '07' 'wi' 'in' 'nk' 'wi' 'in' 'nk'

basically i thought of:

(\w{2})+

but this gave me:
'ic', 'ow', 'fo', 'ed', 'er', '00', 'nk', 'nk'

which is obviously not what i need...
any suggestions?

[dirk]

Hi taltas,

don't know the Python code but this is working in Perl:

Code:

$string = 'A quick brown fox jumped over 007 wink* wink*';

while ( $string =~ m{ (\w{2}) }xmsg ) {
    pos $string = ( pos $string ) - 1;
    print "'$1' ";
}

It's a bit tricky because you always have to go one step back.

"pos" returns the position of the last search. If you substract 1 you
can go one step back.

[Bompa]

That's a good one Dirk.

Code:

$st = 'A quick brown fox jumped over 007 wink* wink*';

@chars = split(//, $st);  # SPLIT STRING INTO A LIST

for($x=0; $x<@chars; ++$x) {
  if($chars[$x] =~ /\w/ and $chars[$x+1] =~ /\w/) {
    print "$chars[$x]$chars[$x+1]\n";
  }
}

Split the string into an array, iterate the array, if current character
is a valid character and is followed by a valid character, print both. 
In my demo code, I just used \w as a validation to keep it simple in my mind.

At TheCache, you get a choice of solutions.
Bompa

[dirk]

That's Perl ;-)

"There's more than one way to do it", commonly known as TMTOWTDI.

Or the second slogan which is often forgotten:

"Easy things should be easy and hard things should be possible".

[vsloathe]

Python's slogan is "Perl done right", but I'm more of a Ruby man these days.

[dirk]

Python is interesting, I will use it too for our projects.

Anyway, a slogan shouldn't look at other languages.

Python is "Python done right" would be a better slogan ;-)