如何在Python中从字符串创建Google Like文本片段？

Question

I am trying to build something similar to Google's text snippet. The Google snippet contains highlighted keywords and "shifts" the text nicely in case a keyword does not appear right at the beginning of the analyzed string.

For example:

keyword "nike"

haystack string "lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor it is no wonder that nike is one of the largest brands in the world is not lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor

should become this snippet:

... lorem ipsum dorlor it is no wonder that nike is one of the largest brands in the world is not lorem ipsum dorlor lorem dorlor lorem ipsum dorlor loremdorlor lorem ipsum dorlor loremipsum dorlor lorem ipsum dorlor lorem ...

This is what I have so far as an idea:

keywordPosition = haystack.lower().index(keyword.lower())
snippetStart = keywordPosition - 100
snippetEnd = keywordPosition + 200
haystack = " ..." + haystack[snippetStart:snippetEnd] + " ..."

Is there an elegant way in python to dynamically adjust snippetStart and snippetEnd? In many cases the above approach obviously throws an exception since the haystrack slice indices are out of range.

Answer 1

I created a little example with comments for you here.

http://pythonfiddle.com/google-snippet

haystack = "lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor it is no wonder that nike is one of the largest brands in the world is not lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor lorem ipsum dorlor"

needle = "nike342"

lookahead = 7  # Number of tokens to show before "nike"

tokens = haystack.split(" ")  # Split string into a list of tokens

found_index = -1  #  Represents the index of the token.  Initialize to -1 and assume it doesn't exist.

# Loop through tokens and compare each to the needle.  If we find the needle, rememeber the index and break out of the loop

found_index = tokens.index(needle)        

try:
    found_index = tokens.index(needle)
    # Get the max of the found index minus the number of words to show before the needle, and 0
    found_index = max(found_index - lookahead, 0)        

    # Create a sub list of the tokens from the found_index and end, then join those terms back together with a space.
    snippet = " ".join(tokens[found_index:len(tokens)])

except ValueError:
    snippet = ""  # No snippet or whatever error handling you are going to do

print snippet