使用Python正则表达式过滤推文

Question

I'm trying to create a query that filters tweets by the @ or # tags.

So I just want the results for either @Obama or #Obama but not Obama. This is what I have so far:

re.compile(r'\b(?:#|@|)*%s*\b' % re.escape(obama), re.IGNORECASE)

Thanks for the replies....I tries both answers and what seems to work in my situation is:

 re.compile(r'\b[#@]*%s\b' % re.escape(term), re.IGNORECASE)  

'term' is an element in a list which I iterate over. This then returns tweets that has either a # or @ pre-pended to the 'term'. Itried not using '*' but It was giving out exceptions.

Thanks

Answer 1

Try using this regular expression:

r'\b[#@]{name}\b'.format(name=re.escape('Obama'))

Character class [%@] works faster then choice group (?:#|@) .

So, we begin with word boundary \\b , then follows # or @ . Then goes substitute from obama variable. Then goes the trailing boundary.

In the question you used * quantifiers which repeat the previous expression from 0 to infinity times. There is no reason to repeat # and @ symbols. Also, the last sybmol of obama shouldn't be repeated either.

Answer 2

If this is purely to do with regexes, and has nothing to do with Twitter per se (aside from the fact that you're filtering tweets), then the regex you want is this:

compiled = re.compile(r'\b[#@]obama\b', re.IGNORECASE)

If you want an example of some code doing something similar to what you're doing, take a look at this as it might be a worthwhile example:

https://github.com/kgaughan/is-on-a-train/blob/master/isonatrain.py

That code tracks a bunch of users, looking for certain trigger phrases, and writes out a HTML file based on what they say.