Maximum PC

Many of us have used them - and those who haven't will truly appreciate them in the future - so using whatever language you like that supports RE's give this simple exercise a shot. =)

Houses are either: new, old or neither (ie null value)
Houses can only come in the following colors: red, blue, and green

#1 Just matching - return either a true or false. Write a RE that matches all of the possible colors of houses and color types:

My house is red.
My house is green.
My house is blue.
My house is black. => returns false

#2 Now do add a replace to #1. No one wants an old looking house, so change any occurance of old to new, like so...
My old house is red. => My new house is red.
My house is blue. => My house is blue.

#3 The previous, add the color orange, and use several colors per sentence.

Match combinations like...
My house has blue walls and red trim.
My house has green walls and blue trim.
etc.

In any order...
My house has red trim and orange walls.
etc.

Still replace... everywhere!
My house has an old red fence. => My house has a new red fence.

(if you're having trouble at this point, it may help to break it down into multiple simpler RE's that solve part of the problem and then think about forming them into more complex RE's).

#3 Match for any number of colors/fixtures (using simple english grammar)....
My house has blue walls, red trim and a red fence.
My house has green walls, blue trim, a red fence, and an orange fence.
etc.

My house has an old red tree house and blue fence. => My house has a new red tree house and blue fence.

#1 and #2 are simple... and good practice. I'll post a solution in Java either T or W depending on my free time. Maybe we'll parse a log file or something more useful next time.

I just found out about them about a week ago. They seem pretty obscure still since when I asked 2 friends majoring in CS they had no idea.

I have absolutely no idea.. Maybe I'll look them up and try it out.

I'm told lycos uses them: http://en.wikipedia.org/wiki/Regular_expression

Looks like a good test for my syntax tree and parser. I'll try it out when I get a chance.

Oh, you'll be amazed how handy they are... =)

Did you get a chance? I've done the first two - I'll leave the third for now in case someone else still wants to try...



And the output (I ran it in NetBeans):

init:
deps-jar:
Compiling 1 source file to C:\Documents and Settings\zd253c\ICD Checker Tools\build\classes
compile-single:
run-single:
Sentence: My house is red. matches => return true
Sentence: My house is green. matches => return true
Sentence: My house is blue. matches => return true
Sentence: My house is black. does not match => return false
before: My old house is red.
after: My new house is red.
before: My house is blue.
after: My house is blue.
BUILD SUCCESSFUL (total time: 2 seconds)

Nah, the only chances I've had I was working on a gametype extractor for Halo 2 so I can edit my gametypes.

I'll try to do it tomorrow.

Sorry, couldn't access my computer and get the source I have for my parser today, and I don't feel like rewriting 1200 lines right now, so maybe tomorrow.

EDIT: I finally got the files.

Damnit just when I started the tutorial I had to get off.

The RSS importer for azureus supports them so I actually have a (weak) reason to learn them.

Give us examples of what you would want to match (and not match).

Grep'ing for files across the internet... nice. =)

I love RegExs and I hate RegExs.

Well I want to setup my RSSFeed plugin for Azureus to grab the new torrents of <> show... unfortunately I can't just put the name in there because there are releases in several languages and I only want the english one. Luckily the torrent files themselves are named with the following nomenclature and as long as it persists a regex should work. The naming style goes as follows:

n - prefixes all the torrents for this series, assumed to be part of the torrent
133 - a 3 digit number for the episodes number, guess which episode number the example is for
.torrent - obviously at the end of all torrents.

A full blown example of a desireable torrent: n124.torrent

now the releases in the languages I DONT want would look the same, only with an added 2 letter suffix to the ep number, as exampled:

n124br.torrent - the br is for brazillian subtitles. I do NOT want this as I don't read brazillian.

Heres the regex I cooked up to do this for me: ... which works at least in editpad where I tested it.

Unfortunately the RSS feed only uses the title of the shows release and its own sites php index # of the torrent (which is likely to be something like 523373) until you download the file where it sends you "n124.torrent".

I've kinda lost interest in the show so I haven't bothered to try to find a work around or a regex the will properly filter by title name (though that might be hard as I'm not sure the RSSFeed plugin allows you to use regex's for title searching in addition to the fact that the titles of the releases vary in nomenclature).

...so, do I pass regex's 101?

I'm not sure. Based on your description, I would have written the following regex...



The difference being that mine will match something like n333.torrent - yours won't. They may have been what you meant though.

They are only on episode 135, it would take a year to get past the 200 mark. If it matches 300+ it might pull down manga releases.

Looks like you pass. =)