--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/progs/crawler.scala	Sat Jun 15 09:23:18 2013 -0400
@@ -0,0 +1,144 @@
+import io.Source
+import scala.util.matching.Regex
+
+// gets the first ~10K of a page
+def get_page(url: String) : String = { 
+  try {
+    Source.fromURL(url).take(10000).mkString  
+  }
+  catch {
+    case e => {
+      println("  Problem with: " + url)
+      ""
+    }
+  }
+}
+
+// non-existing page -> returns the empty string
+get_page("""http://www.foobar.com""")
+
+
+// staring URL for the crawler
+val startURL = """http://www.inf.kcl.ac.uk/staff/urbanc/"""
+
+// starts with an "
+// then either http or https
+// then ://
+// then any character that is not "
+// finally "
+val http_pattern = """\"((?:http|https)://(?:[^\"])*)\"""".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
+}
+
+// get all urls in startURL
+get_all_URLs(get_page(startURL))
+
+// number of all urls in startURL 
+get_all_URLs(get_page(startURL)).toList.length
+
+
+// naive version - seraches until a given depth
+// visits pages potentially more than once
+def crawl(url: String, n: Int) : Unit = {
+  if (n == 0) ()
+  else {
+    println("Visiting: " + n + " " + url)
+    for (u <- get_all_URLs(get_page(url))) crawl(u, n - 1)
+  }
+}
+
+crawl(startURL, 2)
+
+
+//breadth-first version without visiting 
+//pages twice
+def bf_crawl(todo: Set[String], visited: Set[String], n: Int) : Unit = {
+  if (n == 0) ()
+  else {
+    val new_todo = todo.flatMap { 
+      url => {
+        if (visited.contains(url)) Set[String]()
+        else {
+          println("Visiting: " + n + " " + url)
+          get_all_URLs(get_page(url))
+        }
+      }
+    } 
+    bf_crawl(new_todo, visited union todo, n - 1)
+  }
+}
+
+bf_crawl(Set(startURL1), Set(), 2)
+
+
+//breadth-first version without visiting 
+//pages twice and only in "my" domain
+val my_pattern = """urbanc""".r
+
+// breadth first search avoiding double searches
+def bf_crawl2(todo: Set[String], visited: Set[String], n: Int) : Unit = {
+  if (n == 0) ()
+  else {
+    val new_todo = todo.flatMap { 
+      url => {
+        if (visited.contains(url)) Set[String]()
+        else if (my_pattern.findFirstIn(url) == None) Set[String]()
+        else {
+          println("Visiting: " + n + " " + url);
+          get_all_URLs(get_page(url))
+        }
+      }
+    } 
+    bf_crawl2(new_todo, visited union todo, n - 1)
+  }
+}
+
+bf_crawl2(Set(startURL1), Set(), 5)
+
+// email harvester
+// from 
+// http://net.tutsplus.com/tutorials/other/8-regular-expressions-you-should-know/
+
+val email_pattern = """([a-z0-9_\.-]+)@([\da-z\.-]+)\.([a-z\.]{2,6})""".r
+
+def bf_crawl3(todo: Set[String], visited: Set[String], n: Int) : Unit = {
+  if (n == 0) ()
+  else {
+    val new_todo = todo.flatMap { 
+      url => {
+        if (visited.contains(url)) Set[String]()
+        else {
+          println("Visiting: " + n + " " + url);
+          val page = get_page(url)
+          println(email_pattern.findAllIn(page).mkString("\n"))
+          get_all_URLs(get_page(url))
+        }
+      }
+    } 
+    bf_crawl3(new_todo, visited union todo, n - 1)
+  }
+}
+
+bf_crawl3(Set(startURL1), Set(), 3)
+
+
+// depth-first version does not work,
+// because it might visit pages at depth 1
+// while it still wants to visit them at 
+// depth 2 
+var visited = Set("")
+
+def crawl(url: String, n: Int) : Unit = {
+  if (n == 0) ()
+  else if (visited.contains(url)) () //println("Already visited: " + n + " " + url)
+  else {
+    println("Visiting: " + n + " " + url);
+    visited += url
+    for (u <- getAllURLs(getURLpage(url))) crawl(u, n - 1);
+  }
+}