Left Brain Tinkering: Regex Filter Craigslist Search Results

How much does it suck to search for something on Craigslist (especially when looking for cars) only to get a bunch of completely irrelevant results?

In the automotive postings there are a bunch of jackasses who feel that they need to list their vehicle with a bunch of 'keywords' so that their vehicle is displayed to people who MIGHT be interested in it.

Consider this lovely F-150. The owner thought that someone interested in a Ford Probe might consider buying a pickup:

This guy was really clever to add all of these keywords to his ad.

Because of these people, your search results will often look like this:

I'm trying to find Jeep Wranglers for sale and I don't care about this crap.

I thought, "Gee, it would be sweet if I could use regex to filter these results."

Well, fortunately if you have Greasemonkey, there is a way to do this. I searched userscripts.org and found the Craigslist Live Filter by Sam Rawlins. Sam has posted the source at Github also, which is awesome as I'll explain in a minute.

After installing Sam's script, the first thing I noticed was that it was written to EXCLUDE results. That is, what you are searching for can be minimalized in one of two ways:

1. The script will make it gray and in a smaller font
2. The result will simply be hidden from view.

This *will* do what I want but it makes my search a little more tedious. In this case I'd need to have a regex to get rid of the crap above, something maybe like:

Focus|sand|explorer|cherokee|silverado|expedition

In this case what I'm searching for will be REMOVED from the results by Sam's script.

Instead, I want the tool to show me only what I WANT and exclude the other crap. So, I've added an 'invert' checkbox to his script and I checked in the result at Github.

Now, since I'm interested in only 1997+ Jeep Wranglers, I can enter a regex like this:

199[789]|20\d\d.*wrangler

If you're not familiar with regular expressions, I could go on for another hour about how cool this little meta language is... Instead I'll just explain that the above awesomeness means that I want to look for the text "199" followed by a 7, 8, or 9. This means the following will produce a positive hit:

1997
1998
1999

Further, the pipe (|) means OR, and I continue by saying that I'm also interested in the text "20" followed by two digits (\d means any digit character). Then I say that the script is allowed to match ANYTHING (.) as MANY TIMES AS IT WANTS (*).

So at the end of the day, I'm looking for 1997-1999 OR 20xx with the word wrangler somewhere in the title.

Here is what the script looks like in action: