author | Christian Urban <christian dot urban at kcl dot ac dot uk> |
Thu, 26 Sep 2013 11:05:54 +0100 | |
changeset 104 | ffde837b1db1 |
parent 97 | 60a3ba90dd53 |
child 254 | dcd4688690ce |
permissions | -rw-r--r-- |
val http_pattern = """\"https?://[^\"]*\"""".r val my_urls = """urbanc""".r val email_pattern = """([a-z0-9_\.-]+)@([\da-z\.-]+)\.([a-z\.]{2,6})""".r def crawl(url: String, n: Int) : Unit = { if (n == 0) () else { println(s"Visiting: $n $url") val page = get_page(url) println(email_pattern.findAllIn(page).mkString("\n")) for (u <- get_all_URLs(page)) crawl(u, n - 1) } }