7
|
1 |
val http_pattern = """\"https?://[^\"]*\"""".r
|
|
2 |
val my_urls = """urbanc""".r
|
|
3 |
val email_pattern =
|
|
4 |
"""([a-z0-9_\.-]+)@([\da-z\.-]+)\.([a-z\.]{2,6})""".r
|
|
5 |
|
|
6 |
def crawl(url: String, n: Int) : Unit = {
|
|
7 |
if (n == 0) ()
|
|
8 |
else {
|
96
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
9 |
println(s"Visiting: $n $url")
|
7
|
10 |
val page = get_page(url)
|
|
11 |
println(email_pattern.findAllIn(page).mkString("\n"))
|
|
12 |
for (u <- get_all_URLs(page)) crawl(u, n - 1)
|
|
13 |
}
|
|
14 |
}
|