How to replace substring that's not enclosed in tags with regex in python

Question

I have sentences.

text="The president of America is <PERSON>Barack Obama</PERSON>. He was born on August 4, 1961. Obama was reelected president in November 2012".

I want to put <PERSON></PERSON> tag in "Obama", so the result will be like this:
The president of America is <PERSON>Barack Obama</PERSON>. He was born on August 4, 1961. <PERSON>Obama</PERSON> was reelected president in November 2012".

I want to find substring(example: Obama) that there is no tag <PERSON> before the substring and there is no tag </PERSON> after the substring, but I don't know the right syntax for regex in python.
**I'm new to python :''

With simple regex re.sub(namedEntity, "<PERSON>"+namedEntity+"</PERSON>", text) will give an output
The president of America is <PERSON>Barack <PERSON>Obama</PERSON></PERSON>. He was born on August 4, 1961. <PERSON>Obama</PERSON> was reelected president in November 2012".

this is my code(using python2.7)

import re

result=re.sub(r"((?!<PERSON>).*"+namedEntity+".*(?!</PERSON>))","<PERSON>"+namedEntity+"</PERSON>",text)

print "result: "+result

The output
result: <PERSON>Obama</PERSON>
And I don't know that is the first "Obama" or the second one.

Thanks for your help before

Did you copied the code from somewhere ? Do you understand what you are doing in that regex ? — user2705585
– user2705585, Commented Mar 6, 2016 at 18:18
I tried the regex in regex101.com/#python by learn from this answer stackoverflow.com/questions/6259443/… . Maybe I'm wrong because I assume that ?!regex means "not contain regex" :'' — khusnanadia
– khusnanadia, Commented Mar 6, 2016 at 18:34

Holloway · Accepted Answer · 2016-03-06 18:30:54Z

2

You are very close. In your new regex r"((?!<PERSON>).*"+namedEntity+".*(?!</PERSON>))", you have .* before and after which matches 'Obama' with any characters before and after it and the lookarounds are ignored because the tags are in the matched group. If you remove them, you get the results you're after.

>>> import re
>>> text = "The president of America is <PERSON>Barack Obama</PERSON>. He was born on August 4, 1961. Obama was reelected president in November 2012"
>>> namedEntity = 'Obama'
>>> result = re.sub(r"((?!<PERSON>)"+namedEntity+"(?!</PERSON>))","<PERSON>"+namedEntity+"</PERSON>",text)
>>> print result
'The president of America is <PERSON>Barack Obama</PERSON>. He was born on August 4, 1961. <PERSON>Obama</PERSON> was reelected president in November 2012'

For future regex testing, regex101 works well to check how things work as you change them live. For your case this shows what's happening.

edited Mar 6, 2016 at 18:30

answered Mar 6, 2016 at 18:19

Holloway

7,5371 gold badge29 silver badges37 bronze badges

Sign up to request clarification or add additional context in comments.

7 Comments

user2705585 Over a year ago

Shouldn't it be (?<!<PERSON>) i.e negative lookbehind ? I actually got confused there.

Holloway Over a year ago

@noob, I don't think so. You want to ignore the matches that already have the tags around them.

Jan Over a year ago

Just an additional ideone.com demo to show that this is the correct answer (+1).

khusnanadia Over a year ago

Is that ?! and ?<! applied for look ahead and look behind exactly before and after the substring? Because went I try to "The president of America is <PERSON>Barack Obama Aaa</PERSON>. He was born on August 4, 1961. Obama Aaa was reelected president in November 2012". It gives the result "The president of America is <PERSON>Barack <PERSON>Obama</PERSON> Aaa</PERSON>. He was born on August 4, 1961. <PERSON>Obama</PERSON> Aaa was reelected president in November 2012" cc @Holloway

Holloway Over a year ago

@KhusnaNadia, that's true, it relies on the tags being around the name.

|

riteshtch · Accepted Answer · 2016-03-06 18:21:43Z

just remove the .* part in your regex-lookarounds.

>>>text="The president of America is <PERSON>Barack Obama</PERSON>. He was born on August 4, 1961. Obama was reelected president in November 2012"
>>> surname=re.search(r'<PERSON>(.*)</PERSON>', text).group(1).split()[1]
>>> print surname
Obama
>>> re.sub(r'(?<!<PERSON>)'+surname+'(?!</PERSON>)', '<PERSON>'+surname+'</PERSON>', text)'  
The president of America is <PERSON>Barack Obama</PERSON>. He was born on August 4, 1961. <PERSON>Obama</PERSON> was reelected president in November 2012'
>>>

Note: you can also extract the surname of the person using regex and capture groups which i have captured in surname variable. You can use (?<!regex) to assert negative lookbehind and (?!regex) to assert negative lookahead

Collectives™ on Stack Overflow

How to replace substring that's not enclosed in tags with regex in python

2 Answers 2

7 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

7 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related