author | Christian Urban <christian dot urban at kcl dot ac dot uk> |
Sat, 31 Oct 2015 11:37:55 +0000 | |
changeset 364 | 50ce3667c190 |
parent 254 | dcd4688690ce |
permissions | -rw-r--r-- |
val http_pattern = """"https?://[^"]*"""".r def unquote(s: String) = s.drop(1).dropRight(1) def get_all_URLs(page: String) : Set[String] = http_pattern.findAllIn(page).map(unquote).toSet def crawl(url: String, n: Int) : Unit = { if (n == 0) () else { println(s"Visiting: $n $url") for (u <- get_all_URLs(get_page(url))) crawl(u, n - 1) } } crawl(some_start_URL, 2)