On Wednesday, June 13, 2018 at 6:30:45 AM UTC+5:30, Cameron Simpson wrote:
> On 11Jun2018 13:48, Subhabrata Banerjee wrote:
> >I have the following sentence,
> >
> >"Donald Trump is the president of United States of America".
> >
> >I am trying to extract the index 'of', not only for single but also
> >for its multi-occurance (if they occur), from the list of words of the
> >string, made by simply splitting the sentence.
> > index1=[index for index, value in enumerate(words) if value == "of"],
> >where words=sentence.split()
> >
> >I could do this part more or less nicely.
> >
> >But I am trying to say if the list of words has the words "United"
> >and "States" and it has "of " in the sentence then the previous
> >word before of is, president.
> >
> >I am confused how may I write this, if any one may help it.
> You will probably have to drop the list comprehension and go with something 
> more elaborate.
> Also, lists have an "index" method:
>   >>> L = [4,5,6]
>   >>> L.index(5)
>   1
> though it doesn't solve your indexing problems on its own.
> I would be inclined to deconstuct the sentence into a cross linked list of 
> elements. Consider making a simple class to encapsulate the knowledge about 
> each word (totally untested):
>   class Word:
>     def __init__(word):
>       self.word = word
>   words = []
>   for index, word in sentence.split():
>     W = Word(word)
>     W.index = index
>     words.append(W)
>     W.wordlist = words
> Now you have a list of Word objects, each of which knows its list position 
> _and_ also knows about the list itself, _and_ you have the list of Word 
> objects 
> correspnding to your sentence words.
> You'll notice we can just hang whatever attributes we like off these "Word" 
> objects: we added a .wordlist and .index on the fly. It isn't great formal 
> object design, but it makes building things up very easy.
> You can add methods or properties to your class, such as ".next":
>   @property
>   def next(self):
>     return self.wordlist[self.index - 1]
> and so forth. That will let you write expressions about Words:
>   for W in wordlist:
>     if W.word == 'of' and W.next.word == 'the' and W.next.next.word == 
> 'United' ...:
>       if W.previous.word != 'president':
>         ... oooh, unexpected preceeding word! ...
> You can see that you could also write methods like "is_preceeded_by":
>   def is_preceed_by(self, word2):
>     return self.previous.word == word2
> and test "W.is_preceeded_by('president')".
> In short, write out what you would like to express. Then write methods that 
> implement the smaller parts of what you just wrote.
> Cheers,
> Cameron Simpson <c

Dear Sir, 
Thank you for your kind reply. I am trying in few days time and getting back. I 
made a small fix of my own and I would discuss it, too. Thank you for your kind 
words, but I ignore unnecessary remarks. So please do not worry. Meanwhile, 
please see as you are posting your mail id is coming here, it may be misused. 
If possible, please omit it as you post next time. 


Reply via email to