app4.scala
author Christian Urban <christian dot urban at kcl dot ac dot uk>
Fri, 23 Nov 2012 14:08:31 +0000
changeset 71 7717f20f0504
parent 13 6d554a18b81e
permissions -rw-r--r--
updated
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
7
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     1
val http_pattern = """\"https?://[^\"]*\"""".r
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     2
val my_urls = """urbanc""".r
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     3
val email_pattern = 
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     4
  """([a-z0-9_\.-]+)@([\da-z\.-]+)\.([a-z\.]{2,6})""".r
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     5
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     6
def crawl(url: String, n: Int) : Unit = {
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     7
  if (n == 0) ()
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     8
  else {
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
     9
    println("Visiting: " + n + " " + url)
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    10
    val page = get_page(url)
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    11
    println(email_pattern.findAllIn(page).mkString("\n"))
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    12
    for (u <- get_all_URLs(page)) crawl(u, n - 1)
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    13
  }
73cf4406b773 updated
Christian Urban <urbanc@in.tum.de>
parents:
diff changeset
    14
}