Web scraping
You are encouraged to solve this task according to the task description, using any language you may know.
Create a program that downloads the time from this URL: http://tycho.usno.navy.mil/cgi-bin/timer.pl and then prints the current UTC time by extracting just the UTC time from the web page's HTML.
If possible, only use libraries that come at no extra monetary cost with the programming language and that are widely available and popular such as CPAN for Perl or Boost for C++.
Ada
<lang Ada>with AWS.Client, AWS.Response, AWS.Resources, AWS.Messages; with Ada.Text_IO, Ada.Strings.Fixed; use Ada, AWS, AWS.Resources, AWS.Messages;
procedure Get_UTC_Time is
Page : Response.Data; File : Resources.File_Type; Buffer : String (1 .. 1024); Position, Last : Natural := 0; S : Messages.Status_Code;
begin
Page := Client.Get ("http://tycho.usno.navy.mil/cgi-bin/timer.pl");
S := Response.Status_Code (Page);
if S not in Success then
Text_IO.Put_Line
("Unable to retrieve data => Status Code :" & Image (S) &
" Reason :" & Reason_Phrase (S));
return;
end if;
Response.Message_Body (Page, File);
while not End_Of_File (File) loop
Resources.Get_Line (File, Buffer, Last);
Position :=
Strings.Fixed.Index
(Source => Buffer (Buffer'First .. Last),
Pattern => "UTC");
if Position > 0 then
Text_IO.Put_Line (Buffer (5 .. Position + 2));
return;
end if;
end loop;
end Get_UTC_Time;</lang>
AutoHotkey
<lang AutoHotkey>UrlDownloadToFile, http://tycho.usno.navy.mil/cgi-bin/timer.pl, time.html FileRead, timefile, time.html pos := InStr(timefile, "UTC") msgbox % time := SubStr(timefile, pos - 9, 8)</lang>
AWK
This is inspired by GETURL example in the manual for gawk.
<lang awk>#! /usr/bin/awk -f
BEGIN {
purl = "/inet/tcp/0/tycho.usno.navy.mil/80"
ORS = RS = "\r\n\r\n"
print "GET /cgi-bin/timer.pl HTTP/1.0" |& purl
purl |& getline header
while ( (purl |& getline ) > 0 )
{
split($0, a, "\n")
for(i=1; i <= length(a); i++)
{
if ( a[i] ~ /UTC/ )
{
sub(/^
/, "", a[i])
printf "%s\n", a[i]
}
}
}
close(purl)
}</lang>
ALGOL 68
<lang algol68>STRING
domain="tycho.usno.navy.mil", page="cgi-bin/timer.pl";
STRING # search for the needle in the haystack #
needle = "UTC", hay stack = "http://"+domain+"/"+page,
re success="^HTTP/[0-9.]* 200", re result description="^HTTP/[0-9.]* [0-9]+ [a-zA-Z ]*", re doctype ="\s\s<![Dd][Oo][Cc][Tt][Yy][Pp][Ee] [^>]+>\s+";
PROC raise error = (STRING msg)VOID: ( put(stand error, (msg, new line)); stop);
PROC is html page = (REF STRING page) BOOL: (
BOOL out=grep in string(re success, page, NIL, NIL) = 0;
IF INT start, end;
grep in string(re result description, page, start, end) = 0
THEN
page:=page[end+1:];
IF grep in string(re doctype, page, start, end) = 0
THEN page:=page[start+2:]
ELSE raise error("unknown format retrieving page")
FI
ELSE raise error("unknown error retrieving page")
FI;
out
);
STRING reply; INT rc = http content (reply, domain, haystack, 0); IF rc = 0 AND is html page (reply) THEN
STRING line; FILE freply; associate(freply, reply); on logical file end(freply, (REF FILE freply)BOOL: (done; SKIP)); DO get(freply,(line, new line)); IF string in string(needle, NIL, line) THEN print((line, new line)) FI OD; done: SKIP
ELSE raise error (strerror (rc)) FI</lang> Sample output:
<BR>Sep. 26, 21:51:17 UTC Universal Time
BBC BASIC
Note that the URL cache is cleared so the code works correctly if run more than once. <lang bbcbasic> SYS "LoadLibrary", "URLMON.DLL" TO urlmon%
SYS "GetProcAddress", urlmon%, "URLDownloadToFileA" TO UDTF%
SYS "LoadLibrary", "WININET.DLL" TO wininet%
SYS "GetProcAddress", wininet%, "DeleteUrlCacheEntryA" TO DUCE%
url$ = "http://tycho.usno.navy.mil/cgi-bin/timer.pl"
file$ = @tmp$+"navytime.txt"
SYS DUCE%, url$
SYS UDTF%, 0, url$, file$, 0, 0 TO result%
IF result% ERROR 100, "Download failed"
file% = OPENIN(file$)
REPEAT
text$ = GET$#file%
IF INSTR(text$, "UTC") PRINT MID$(text$, 5)
UNTIL EOF#file%
CLOSE #file%</lang>
C
There's no any proper error handling.
<lang c>#include <stdio.h>
- include <string.h>
- include <curl/curl.h>
- include <sys/types.h>
- include <regex.h>
- define BUFSIZE 16384
size_t lr = 0;
size_t filterit(void *ptr, size_t size, size_t nmemb, void *stream) {
if ( (lr + size*nmemb) > BUFSIZE ) return BUFSIZE; memcpy(stream+lr, ptr, size*nmemb); lr += size*nmemb; return size*nmemb;
}
int main() {
CURL *curlHandle; char buffer[BUFSIZE]; regmatch_t amatch; regex_t cregex;
curlHandle = curl_easy_init(); curl_easy_setopt(curlHandle, CURLOPT_URL, "http://tycho.usno.navy.mil/cgi-bin/timer.pl"); curl_easy_setopt(curlHandle, CURLOPT_FOLLOWLOCATION, 1); curl_easy_setopt(curlHandle, CURLOPT_WRITEFUNCTION, filterit); curl_easy_setopt(curlHandle, CURLOPT_WRITEDATA, buffer); int success = curl_easy_perform(curlHandle); curl_easy_cleanup(curlHandle);
buffer[lr] = 0; regcomp(&cregex, " UTC", REG_NEWLINE); regexec(&cregex, buffer, 1, &amatch, 0); int bi = amatch.rm_so; while ( bi-- > 0 ) if ( memcmp(&buffer[bi], "
", 4) == 0 ) break;
buffer[amatch.rm_eo] = 0;
printf("%s\n", &buffer[bi+4]);
regfree(&cregex); return 0;
}</lang>
C++
<lang cpp>#include <iostream>
- include <string>
- include <boost/asio.hpp>
- include <boost/regex.hpp>
int main() {
boost::asio::ip::tcp::iostream s("tycho.usno.navy.mil", "http");
if(!s)
std::cout << "Could not connect to tycho.usno.navy.mil\n";
s << "GET /cgi-bin/timer.pl HTTP/1.0\r\n"
<< "Host: tycho.usno.navy.mil\r\n"
<< "Accept: */*\r\n"
<< "Connection: close\r\n\r\n" ;
for(std::string line; getline(s, line); )
{
boost::smatch matches;
if(regex_search(line, matches, boost::regex("
(.+\\s+UTC)") ) )
{
std::cout << matches[1] << '\n';
break;
}
}
}</lang>
C#
<lang csharp>class Program
{
static void Main(string[] args)
{
WebClient wc = new WebClient();
Stream myStream = wc.OpenRead("http://tycho.usno.navy.mil/cgi-bin/timer.pl");
string html = "";
using (StreamReader sr = new StreamReader(myStream))
{
while (sr.Peek() >= 0)
{
html = sr.ReadLine();
if (html.Contains("UTC"))
{
break;
}
}
}
Console.WriteLine(html.Remove(0, 4));
Console.ReadLine();
}
}
</lang>
Clojure
Clojure 1.2:
<lang clojure> (second (re-find #" (\d{1,2}:\d{1,2}:\d{1,2}) UTC" (slurp "http://tycho.usno.navy.mil/cgi-bin/timer.pl"))) </lang>
Common Lisp
<lang lisp>BOA> (let* ((url "http://tycho.usno.navy.mil/cgi-bin/timer.pl")
(regexp (load-time-value
(cl-ppcre:create-scanner "(?m)^.{4}(.+? UTC)")))
(data (drakma:http-request url)))
(multiple-value-bind (start end start-regs end-regs)
(cl-ppcre:scan regexp data)
(declare (ignore end))
(when start
(subseq data (aref start-regs 0) (aref end-regs 0)))))
"Aug. 12, 04:29:51 UTC"</lang>
D
<lang d>import std.stdio, std.string, std.socket;
string loadPage(string url, string host) {
InternetAddress addr = new InternetAddress(url, 80); TcpSocket sock = new TcpSocket(addr);
sock.send("GET " ~ host ~ r" HTTP/1.0\r\n");
sock.send("Host: " ~ host ~ r"\r\n");
sock.send("Connection: close\r\n");
sock.send("\r\n");
char[1024] buffer;
string msg;
uint buflen = 1024;
while (buflen == 1024) {
buflen = sock.receive(buffer);
msg ~= buffer[0 .. buflen];
}
sock.shutdown(SocketShutdown.BOTH); sock.close(); return msg;
}
void main() {
string page = loadPage(r"tycho.usno.navy.mil", r"/cgi-bin/timer.pl");
foreach (line; page.splitlines())
if (line.length > 4 && indexOf(line, " UTC") != -1)
writeln(line[4 .. $]);
}</lang>
Delphi
There are a number of ways to do this with Delphi using any one of a number of free/open source TCP/IP component suites such as, for example, ICS, Synapse and Indy (which ships with Delphi anyway). However, I thought it would be interesting to do this using the Winsock API direct.
<lang Delphi>
program WebScrape;
{$APPTYPE CONSOLE}
{.$DEFINE DEBUG}
uses
Classes, Winsock;
{ Function to connect to host, send HTTP request and retrieve response }
function DoHTTPGET(const hostName: PAnsiChar; const resource: PAnsiChar; HTTPResponse: TStrings): Boolean;
const
Port: integer = 80; CRLF = #13#10; // carriage return/line feed
var
WSAData: TWSAData; Sock: TSocket; SockAddrIn: TSockAddrIn; IPAddress: PHostEnt; bytesIn: integer; inBuffer: array [0..1023] of char; Req: string;
begin
Result := False; HTTPResponse.Clear;
{ Initialise use of the Windows Sockets DLL.
Older Windows versions support Winsock 1.1 whilst newer Windows
include Winsock 2 but support 1.1. Therefore, we'll specify
version 1.1 ($101) as being the highest version of Windows Sockets
that we can use to provide greatest flexibility.
WSAData receives details of the Windows Sockets implementation }
Winsock.WSAStartUp($101, WSAData);
try
{ Create a socket for TCP/IP usage passing in
Address family spec: AF_INET (TCP, UDP, etc.)
Type specification: SOCK_STREAM
Protocol: IPPROTO_TCP (TCP) }
Sock := WinSock.Socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
try
// Check we have a valid socket
if (Sock <> INVALID_SOCKET) then
begin
// Populate socket address structure
with SockAddrIn do
begin
// Address family specification
sin_family := AF_INET;
// Port
sin_port := htons(Port);
// Address
sin_addr.s_addr := inet_addr(hostName);
end;
if (SockAddrIn.sin_addr.s_addr = INADDR_NONE) then
begin
{ As we're using a domain name instead of an
IP Address, we need to resolve the domain name }
IPAddress := Winsock.gethostbyname(hostName);
// Quit if we didn't get an IP Address
if (IPAddress = nil) then
Exit;
// Update the structure with the IP Address
SockAddrIn.sin_addr.s_addr := PLongint(IPAddress^.h_addr_list^)^;
end;
// Try to connect to host
if (Winsock.connect(Sock, SockAddrIn, SizeOf(SockAddrIn)) <> SOCKET_ERROR) then
begin
// OK - Connected
// Compose our request
// Each line of the request must be terminated with a carriage return/line feed
{ The First line specifies method (e.g. GET, POST), path to required resource,
and the HTTP version being used. These three fields are space separated. }
Req := 'GET '+resource+' HTTP/1.1' + CRLF +
// Host: is the only Required header in HTTP 1.1
'Host: '+hostName + CRLF +
{ Persistent connections are the default in HTTP 1.1 but, as we don't want
or need one for this exercise, we must include the "Connection: close"
header in our request }
'Connection: close' + CRLF +
CRLF; // Request must end with an empty line!
// Try to send the request to the host
if (Winsock.send(Sock,Req[1],Length(Req),0) <> SOCKET_ERROR) then
begin
// Initialise incoming data buffer (i.e. fill array with nulls)
FillChar(inBuffer,SizeOf(inBuffer),#0);
// Loop until nothing left to read
repeat
// Read incoming data from socket
bytesIn := Winsock.recv(Sock, inBuffer, SizeOf(inBuffer), 0);
// Assign buffer to Stringlist
HTTPResponse.Text := HTTPResponse.Text + Copy(string(inBuffer),1,bytesIn);
until
(bytesIn <= 0) or (bytesIn = SOCKET_ERROR);
{ Our list of response strings should
contain at least 1 line }
Result := HTTPResponse.Count > 0;
end;
end;
end;
finally
// Close our socket
Winsock.closesocket(Sock);
end;
finally
{ This causes our application to deregister itself from this
Windows Sockets implementation and allows the implementation
to free any resources allocated on our behalf. }
Winsock.WSACleanup;
end;
end;
{ Simple function to locate and return the UTC time from the
request sent to http://tycho.usno.navy.mil/cgi-bin/timer.pl The HTTPResponse parameter contains both the HTTP Headers and the HTML served up by the requested resource. }
function ParseResponse(HTTPResponse: TStrings): string; var
i: Integer;
begin
Result := ;
{ Check first line for server response code
We want something like this: HTTP/1.1 200 OK }
if Pos('200',HTTPResponse[0]) > 0 then
begin
for i := 0 to Pred(HTTPResponse.Count) do
begin
{ The line we're looking for is something like this:
May. 04. 21:55:19 UTC Universal Time }
// Check each line
if Pos('UTC',HTTPResponse[i]) > 0 then
begin
Result := Copy(HTTPResponse[i],5,Pos('UTC',HTTPResponse[i])-1);
Break;
end;
end; end else Result := 'HTTP Error: '+HTTPResponse[0];
end;
const
host: PAnsiChar = 'tycho.usno.navy.mil'; res : PAnsiChar = '/cgi-bin/timer.pl';
var
Response: TStrings;
begin
{ A TStringList is a TStrings descendant class
that is used to store and manipulate a list
of strings.
Instantiate a stringlist class to
hold the results of our HTTP GET }
Response := TStringList.Create;
try
// Try an HTTP GET request
if DoHTTPGET(host,res,Response) then
begin
{$IFDEF DEBUG}
{ Write the entire response to
the console window }
Writeln(Response.text);
{$ELSE}
{ Parse the response and write the
result to the console window }
Writeln(ParseResponse(Response));
{$ENDIF DEBUG}
end
else
Writeln('Error retrieving data');
finally Response.Free; end;
// Keep console window open Readln;
end.
</lang>
Example using Indy's IdHTTP component.
<lang Delphi>program ReadUTCTime;
{$APPTYPE CONSOLE}
uses SysUtils, Classes, IdHTTP;
var
s: string; lHTTP: TIdHTTP; lReader: TStringReader;
begin
lHTTP := TIdHTTP.Create(nil);
try
lReader := TStringReader.Create(lHTTP.Get('http://tycho.usno.navy.mil/cgi-bin/timer.pl'));
while lReader.Peek > 0 do
begin
s := lReader.ReadLine;
if Pos('UTC', s) > 0 then
begin
Writeln(s);
Break;
end;
end;
finally
lHTTP.Free;
lReader.Free;
end;
end.</lang>
E
<lang e>interp.waitAtTop(when (def html := <http://tycho.usno.navy.mil/cgi-bin/timer.pl>.getText()) -> {
def rx`(?s).*>(@time.*? UTC).*` := html println(time)
})</lang>
Erlang
Using regular expressions:
<lang erlang>-module(scraping).
-export([main/0]).
-define(Url, "http://tycho.usno.navy.mil/cgi-bin/timer.pl").
-define(Match, "
(.+ UTC)").
main() -> inets:start(), {ok, {_Status, _Header, HTML}} = http:request(?Url), {match, [Time]} = re:run(HTML, ?Match, [{capture, all_but_first, binary}]), io:format("~s~n",[Time]).</lang>
F#
This code is asynchronous - it will not block any threads while it waits on a response from the remote server. <lang fsharp> open System open System.Net open System.Text.RegularExpressions
async {
use wc = new WebClient()
let! html = wc.AsyncDownloadString(Uri("http://tycho.usno.navy.mil/cgi-bin/timer.pl"))
return Regex.Match(html, @"
(.+ UTC)").Groups.[1].Value
} |> Async.RunSynchronously |> printfn "%s" </lang>
Factor
<lang factor>USING: http.get io sequences ;
"http://tycho.usno.navy.mil/cgi-bin/timer.pl" http-get nip [ "UTC" swap start [ 9 - ] [ 1 - ] bi ] keep subseq print</lang>
Forth
<lang forth>include unix/socket.fs
- extract-time ( addr len type len -- time len )
dup >r search 0= abort" that time not present!" dup >r begin -1 /string over 1- c@ [char] > = until \ seek back to
at start of line r> - r> + ;
s" tycho.usno.navy.mil" 80 open-socket dup s\" GET /cgi-bin/timer.pl HTTP/1.0\n\n" rot write-socket dup pad 4096 read-socket s\" \r\n\r\n" search 0= abort" can't find headers!" \ skip headers s" UTC" extract-time type cr close-socket</lang>
Go
<lang go>package main
import (
"bytes" "fmt" "http" "os" "regexp" "time" "xml"
)
func main() {
resp, err := http.Get("http://tycho.usno.navy.mil/cgi-bin/timer.pl")
if err != nil {
fmt.Println(err) // connection or request fail
return
}
defer resp.Body.Close()
var us string
var ux int
utc := []byte("UTC")
for p := xml.NewParser(resp.Body); ; {
t, err := p.RawToken()
switch err {
case nil:
case os.EOF:
fmt.Println("UTC not found")
return
default:
fmt.Println(err) // read or parse fail
return
}
if ub, ok := t.(xml.CharData); ok {
if ux = bytes.Index(ub, utc); ux != -1 {
// success: found a line with the string "UTC"
us = string([]byte(ub))
break
}
}
}
// first thing to try: parsing the expected date format
if t, err := time.Parse("Jan. 2, 15:04:05 UTC", us[:ux+3]); err == nil {
fmt.Println("parsed UTC:", t.Format("January 2, 15:04:05"))
return
}
// fallback: search for anything looking like a time and print that
tx := regexp.MustCompile("[0-2]?[0-9]:[0-5][0-9]:[0-6][0-9]")
if justTime := tx.FindString(us); justTime > "" {
fmt.Println("found UTC:", justTime)
return
}
// last resort: just print the whole element containing "UTC" and hope
// there is a human readable time in there somewhere.
fmt.Println(us)
}</lang> Output:
parsed UTC: May 23, 00:44:13
Haskell
Using package HTTP-4000.0.8 from HackgageDB <lang Haskell>import Data.List import Network.HTTP (simpleHTTP, getResponseBody, getRequest)
tyd = "http://tycho.usno.navy.mil/cgi-bin/timer.pl"
readUTC = simpleHTTP (getRequest tyd)>>=
fmap ((!!2).head.dropWhile ("UTC"`notElem`).map words.lines). getResponseBody>>=putStrLn</lang>
Usage in GHCi: <lang Haskell>*Main> readUTC 08:30:23</lang>
Icon and Unicon
Icon
Icon has capability to read web pages using the external function cfunc. The Unicon messaging extensions are more succinct.
Unicon
<lang Unicon>procedure main() m := open(url := "http://tycho.usno.navy.mil/cgi-bin/timer.pl","m") | stop("Unable to open ",url) every (p := "") ||:= |read(m) # read the page into a single string close(m)
map(p) ? ( tab(find("
")), ="
", write("UTC time=",p[&pos:find(" utc")])) # scrape and show
end</lang>
J
<lang j> require 'web/gethttp'
_8{. ' UTC' taketo gethttp 'http://tycho.usno.navy.mil/cgi-bin/timer.pl'
04:32:44</lang>
The web/gethttp addon uses Wget on Linux or Windows (J ships with Wget on Windows) and cURL on the Mac.
(A sockets solution is also possible. But, while basic HTTP support is trivial to implement, a full standards compliant implementation and can involve a lot of code to deal with rare corner cases, and the time required to complete a web request is often significantly longer than the time to invoke an external program. This would imply a fair bit of maintenance and testing overhead to deal with issues which rarely matter, if a direct sockets implementation were used.)
Java
<lang java>import java.io.BufferedReader; import java.io.IOException; import java.io.InputStreamReader; import java.net.URL; import java.net.URLConnection;
public class WebTime{
public static void main(String[] args){
try{
URL address = new URL(
"http://tycho.usno.navy.mil/cgi-bin/timer.pl");
URLConnection conn = address.openConnection();
BufferedReader in = new BufferedReader(
new InputStreamReader(conn.getInputStream()));
String line;
while(!(line = in.readLine()).contains("UTC"));
System.out.println(line.substring(4));
}catch(IOException e){
System.err.println("error connecting to server.");
e.printStackTrace();
}
}
}</lang>
Liberty BASIC
<lang lb>if DownloadToFile("http://tycho.usno.navy.mil/cgi-bin/timer.pl", DefaultDir$ + "\timer.htm") = 0 then
open DefaultDir$ + "\timer.htm" for input as #f
html$ = lower$(input$(#f, LOF(#f)))
close #f
a= instr( html$, "utc" )-1
print "UTC";mid$( html$, a-9,9)
end if
end
function DownloadToFile(urlfile$, localfile$)
open "URLmon" for dll as #url calldll #url, "URLDownloadToFileA",_ 0 as long,_ 'null urlfile$ as ptr,_ 'url to download localfile$ as ptr,_ 'save file name 0 as long,_ 'reserved, must be 0 0 as long,_ 'callback address, can be 0 DownloadToFile as ulong '0=success close #url
end function</lang>
Mathematica
<lang mathematica> test = StringSplit[Import["http://tycho.usno.navy.mil/cgi-bin/timer.pl"], "\n"]; Extract[test, Flatten@Position[StringFreeQ[test, "UTC"], False]] </lang>
Objeck
<lang objeck> use Net; use IO; use Structure;
bundle Default {
class Scrape {
function : Main(args : String[]) ~ Nil {
client := HttpClient->New("http://tycho.usno.navy.mil/cgi-bin/timer.pl", 80);
lines := client->Get();
i := 0;
found := false;
while(found <> true & i < lines->Size()) {
line := lines->Get(i)->As(String);
index := line->Find("UTC");
if(index > -1) {
time := line->SubString(index - 9, 9)->Trim();
time->PrintLine();
found := true;
};
i += 1;
};
}
}
} </lang>
OCaml
<lang ocaml>let () =
let _,_, page_content = make_request ~url:Sys.argv.(1) ~kind:GET () in
let lines = Str.split (Str.regexp "\n") page_content in
let str =
List.find
(fun line ->
try ignore(Str.search_forward (Str.regexp "UTC") line 0); true
with Not_found -> false)
lines
in
let str = Str.global_replace (Str.regexp "
") "" str in
print_endline str;
- </lang>
There are libraries for this, but it's rather interesting to see how to use a socket to achieve this, so see the implementation of the above function make_request on this page.
Oz
<lang oz>declare
[Regex] = {Module.link ['x-oz://contrib/regex']}
fun {GetPage Url}
F = {New Open.file init(url:Url)}
Contents = {F read(list:$ size:all)}
in
{F close}
Contents
end
fun {GetDateString Doc}
case {Regex.search "
([A-Za-z0-9:., ]+ UTC)" Doc}
of match(1:S#E ...) then {List.take {List.drop Doc S} E-S+1}
end
end
Url = "http://tycho.usno.navy.mil/cgi-bin/timer.pl"
in
{System.showInfo {GetDateString {GetPage Url}}}</lang>
Perl
<lang perl>use LWP::Simple;
my $url = 'http://tycho.usno.navy.mil/cgi-bin/timer.pl';
get($url) =~ /
(.+? UTC)/
and print "$1\n";</lang>
Perl 6
<lang perl6>use HTTP::Client; # http://github.com/carlins/http-client/
my $site = "http://tycho.usno.navy.mil/cgi-bin/timer.pl";
HTTP::Client.new.get($site).match(/'
'( .+? <ws> UTC )/)[0].say</lang>
Note that the string between '<' and '>' refers to regex tokens, so to match a literal '<BR>' you need to quote it, while <ws> refers to the built-in token whitespace. Also, whitespace is ignored by default in Perl 6 regexes.
PHP
By iterating over each line:
<lang PHP><?
$contents = file('http://tycho.usno.navy.mil/cgi-bin/timer.pl'); foreach ($contents as $line){
if (($pos = strpos($line, ' UTC')) === false) continue; echo subStr($line, 4, $pos - 4); //Prints something like "Dec. 06, 16:18:03" break;
}</lang>
By regular expressions (
):
<lang PHP><?
echo preg_replace(
"/^.*
(.*) UTC.*$/su",
"\\1",
file_get_contents('http://tycho.usno.navy.mil/cgi-bin/timer.pl')
);
</lang>
PicoLisp
<lang PicoLisp>(load "@lib/http.l")
(client "tycho.usno.navy.mil" 80 "cgi-bin/timer.pl"
(when (from "
") (pack (trim (till "U"))) ) )</lang>
Output:
-> "Feb. 19, 18:11:37"
PowerShell
<lang powershell>$wc = New-Object Net.WebClient $html = $wc.DownloadString('http://tycho.usno.navy.mil/cgi-bin/timer.pl') $html -match ', (.*) UTC' | Out-Null Write-Host $Matches[1]</lang>
Protium
English dialect, short form, using integrated Rexx pattern matcher:
<lang html><@ DEFAREPRS>Rexx Parse</@>
<@ DEFPRSLIT>Rexx Parse|'
' UTCtime 'UTC'</@>
<@ LETVARURL>timer|http://tycho.usno.navy.mil/cgi-bin/timer.pl</@>
<@ ACTRPNPRSVAR>Rexx Parse|timer</@>
<@ SAYVAR>UTCtime</@></lang>
English dialect, padded variable-length form:
<lang html><# DEFINE WORKAREA PARSEVALUES>Rexx Parse</#>
<# DEFINE PARSEVALUES LITERAL>Rexx Parse|'
' UTCtime 'UTC'</#>
<# LET VARIABLE URLSOURCE>timer|http://tycho.usno.navy.mil/cgi-bin/timer.pl</#>
<# ACT REPLACEBYPATTERN PARSEVALUES VARIABLE>Rexx Parse|timer</#>
<# SAY VARIABLE>UTCtime</#></lang>
PureBasic
<lang Purebasic>URLDownloadToFile_( #Null, "http://tycho.usno.navy.mil/cgi-bin/timer.pl", "timer.htm", 0, #Null) ReadFile(0, "timer.htm") While Not Eof(0) : Text$ + ReadString(0) : Wend MessageRequester("Time", Mid(Text$, FindString(Text$, "UTC", 1) - 9 , 8))</lang>
Python
<lang python>import urllib page = urllib.urlopen('http://tycho.usno.navy.mil/cgi-bin/timer.pl') for line in page:
if ' UTC' in line:
print line.strip()[4:]
break
page.close()</lang> Sample output:
Aug. 12, 15:22:08 UTC Universal Time
R
There are two ways of scraping data from webpages. You can either convert the page into an array of strings containing each line of HTML and use regular expressions to locate the useful data, or you can parse the HTML and use xPath to locate them. The first method is quicker for simpler pages, but may become more difficult for more complicated ones.
Regex method
Read the page as lines, find the line containing the string "UTC", then extract the portion of that string that is the date.
<lang R> all_lines <- readLines("http://tycho.usno.navy.mil/cgi-bin/timer.pl") utc_line <- grep("UTC", all_lines, value = TRUE) matched <- regexpr("(\\w{3}.*UTC)", utc_line) utc_time_str <- substring(line, matched, matched + attr(matched, "match.length") - 1L) </lang>
The last three lines can be made simpler by using
<lang R> library(stringr) utc_line <- all_lines[str_detect(all_lines, "UTC")] utc_time_str <- str_extract(utc_line, "\\w{3}.*UTC") </lang>
Finally, the date and time must be parsed and printed in the desired format.
<lang R> utc_time <- strptime(utc_time_str, "%b. %d, %H:%M:%S UTC") strftime(utc_time, "%A, %d %B %Y, %H:%M:%S") </lang>
Friday, 13 May 2011, 15:12:20
Parsing method
First, retrieve the web page. See HTTP_Request for more options with this.
<lang R> library(RCurl) web_page <- getURL("http://tycho.usno.navy.mil/cgi-bin/timer.pl") </lang>
Now parse the html code into a tree and retrieve the pre node that contains interesting bit. Without xPath, the syntax is quite clunky.
<lang R> library(XML) page_tree <- htmlTreeParse(webpage) times_node <- page_tree$children$html$children$body$children$h3$children$pre$children times_node <- times_node[names(times_node) == "text"] time_lines <- sapply(times_node, function(x) x$value) </lang>
Here, xPath simplifies things a little bit.
<lang R> page_tree <- htmlTreeParse(web_page, useInternalNodes = TRUE) times_node <- xpathSApply(page_tree, "//pre")1 times_node <- times_node[names(times_node) == "text"] time_lines <- sapply(times_node, function(x) as(x, "character")) </lang>
Either way, the solution proceeds from here as in the regex method. <lang R> utc_line <- time_lines[str_detect(time_lines, "UTC")]
- etc.
</lang>
REBOL
<lang REBOL>REBOL [ Title: "Web Scraping" Author: oofoe Date: 2009-12-07 URL: http://rosettacode.org/wiki/Web_Scraping ]
- Notice that REBOL understands unquoted URL's
service: http://tycho.usno.navy.mil/cgi-bin/timer.pl
- The 'read' function can read from any data scheme that REBOL knows
- about, which includes web URLs. NOTE
- Depending on your security
- settings, REBOL may ask you for permission to contact the service.
html: read service
- I parse the HTML to find the first
(note the unquoted HTML tag - -- REBOL understands those too), then copy the current time from
- there to the "UTC" terminator.
- I have the "to end" in the parse rule so the parse will succeed.
- Not strictly necessary once I've got the time, but good practice.
parse html [thru
copy current thru "UTC" to end]
print ["Current UTC time:" current]</lang>
Ruby
A verbose example for comparison
<lang ruby>require "open-uri"
open('http://tycho.usno.navy.mil/cgi-bin/timer.pl') do |p|
p.each_line do |line|
if line =~ /UTC/
puts line.match(/ (\d{1,2}:\d{1,2}:\d{1,2}) /)
break
end
end
end </lang>
A more concise example
<lang ruby>require 'open-uri' puts URI.parse('http://tycho.usno.navy.mil/cgi-bin/timer.pl').read.match(/ (\d{1,2}:\d{1,2}:\d{1,2}) UTC/)[1] </lang>
Scala
<lang scala> import scala.io.Source
object WebTime extends Application { val text = Source.fromURL("http://tycho.usno.navy.mil/cgi-bin/timer.pl") val utc = text.getLines.find(_.contains("UTC")) utc match { case Some(s) => println(s.substring(4)) case _ => println("error") } } </lang>
Scheme
<lang scheme>; Use the regular expression module to parse the url (use-modules (ice-9 regex) (ice-9 rdelim))
- Variable to store result
(define time "")
- Set the url and parse the hostname, port, and path into variables
(define url "http://tycho.usno.navy.mil/cgi-bin/timer.pl") (define r (make-regexp "^(http://)?([^:/]+)(:)?(([0-9])+)?(/.*)?" regexp/icase)) (define host (match:substring (regexp-exec r url) 2)) (define port (match:substring (regexp-exec r url) 4)) (define path (match:substring (regexp-exec r url) 6))
- Set port to 80 if it wasn't set above and convert from a string to a number
(if (eq? port #f) (define port "80")) (define port (string->number port))
- Connect to remote host on specified port
(let ((s (socket PF_INET SOCK_STREAM 0)))
(connect s AF_INET (car (hostent:addr-list (gethostbyname host))) port)
- Send a HTTP request for the specified path
(display "GET " s)
(display path s)
(display " HTTP/1.0\r\n\r\n" s)
(set! r (make-regexp "
(.+? UTC)")) (do ((line (read-line s) (read-line s))) ((eof-object? line)) (if (regexp-match? (regexp-exec r line)) (set! time (match:substring (regexp-exec r line) 1)))))
- Display result
(display time) (newline)</lang>
Seed7
<lang seed7>$ include "seed7_05.s7i";
include "gethttp.s7i";
const proc: main is func
local
var string: pageWithTime is "";
var integer: posOfUTC is 0;
var integer: posOfBR is 0;
var string: timeStri is "";
begin
pageWithTime := getHttp("tycho.usno.navy.mil/cgi-bin/timer.pl");
posOfUTC := pos(pageWithTime, "UTC");
if posOfUTC <> 0 then
posOfBR := rpos(pageWithTime, "
", posOfUTC);
if posOfBR <> 0 then
timeStri := pageWithTime[posOfBR + 4 .. pred(posOfUTC)];
writeln(timeStri);
end if;
end if;
end func;</lang>
Tcl
<lang tcl>package require http
set request [http::geturl "http://tycho.usno.navy.mil/cgi-bin/timer.pl"]
if {[regexp -line {
(.* UTC)} [http::data $request] --> utc]} {
puts $utc
}</lang>
TUSCRIPT
<lang tuscript> $$ MODE TUSCRIPT SET time = REQUEST ("http://tycho.usno.navy.mil/cgi-bin/timer.pl") SET utc = FILTER (time,":*UTC*:",-) </lang>
TXR
Robust
Large amounts of the document are matched (in fact the entire thing!), rather than blindly looking for some small amount of context.
If the web page changes too much, the query will fail to match. TXR will print the word "false" and terminate with a failed exit status. This is preferrable to finding a false positive match and printing a wrong result. (E.g. any random garbage that happened to be in a line of HTML accidentally containing the string UTC).
<lang txr>@(next `!wget -c http://tycho.usno.navy.mil/cgi-bin/timer.pl -O - 2> /dev/null`) <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final"//EN> <html> <body> <TITLE>What time is it?</TITLE>
@(collect :vars (MO DD HH MM SS (PM " ") TZ TZNAME))
<BR>@MO. @DD, @HH:@MM:@SS @(maybe)@{PM /PM/} @(end)@TZ@/\t+/@TZNAME
@ (until)
@/.*/
@(end)
<A HREF="http://www.usno.navy.mil"> US Naval Observatory</A>
</body></html> @(output) @ (repeat) @MO-@DD @HH:@MM:@SS @PM @TZ @ (end) @(end)</lang>
Sample run:
$ txr navytime.txr Nov-22 22:49:41 UTC Nov-22 05:49:41 PM EST Nov-22 04:49:41 PM CST Nov-22 03:49:41 PM MST Nov-22 02:49:41 PM PST Nov-22 01:49:41 PM AKST Nov-22 12:49:41 PM HAST
Get just the UTC time:
$ txr -DTZ=UTC navytime.txr Nov-22 22:50:16 UTC
Naive
Skip stuff until a line beginning with <BR> has some stuff before "UTC", and capture that stuff:
<lang txr>@(next `!wget -c http://tycho.usno.navy.mil/cgi-bin/timer.pl -O - 2> /dev/null`)
@(skip)
@time@\ UTC@(skip)
@(output)
@time
@(end)</lang>
UNIX Shell
This solution uses curl, which can be downloaded for free, and the popular (at least in the unix world) utilities programs grep and sed.
<lang bash>#!/bin/sh curl -s http://tycho.usno.navy.mil/cgi-bin/timer.pl |
grep ' UTC' | sed -e 's/^
//;s/ UTC.*$//'</lang>
Ursala
This works by launching the wget command in a separate process and capturing its output. The program is compiled to an executable command. <lang Ursala>#import std
- import cli
- executable ('parameterized',)
whatime =
<.file$[contents: --<>]>+ -+
@hm skip/*4+ ~=(9%cOi&)-~l*+ *~ ~&K3/'UTC', (ask bash)/0+ -[wget -O - http://tycho.usno.navy.mil/cgi-bin/timer.pl]-!+-</lang>
Here is a bash session.
$ whatime Jun. 26, 20:49:52 UTC
Visual Basic .NET
New, .NET way with StringReader: <lang vbnet>Imports System.Net Imports System.IO
Dim client As WebClient = New WebClient()
Dim content As String = client.DownloadString("http://tycho.usno.navy.mil/cgi-bin/timer.pl")
Dim sr As New StringReader(content)
While sr.peek <> -1
Dim s As String = sr.ReadLine
If s.Contains("UTC") Then
Dim time As String() = s.Substring(4).Split(vbTab)
Console.WriteLine(time(0))
End If
End While</lang>
Alternative, old fashioned way using VB "Split" function: <lang vbnet>Imports System.Net
Dim client As WebClient = New WebClient()
Dim content As String = client.DownloadString("http://tycho.usno.navy.mil/cgi-bin/timer.pl")
Dim lines() As String = Split(content, vbLf) 'may need vbCrLf
For Each line In lines
If line.Contains("UTC") Then
Dim time As String() = line.Substring(4).Split(vbTab)
Console.WriteLine(time(0))
End If
Next</lang>
- Programming Tasks
- Networking and Web Interaction
- Ada
- AWS
- AutoHotkey
- AWK
- ALGOL 68
- BBC BASIC
- C
- Libcurl
- C++
- Boost
- C sharp
- Clojure
- Common Lisp
- Cl-ppcre
- DRAKMA
- D
- Delphi
- E
- Erlang
- F Sharp
- Factor
- Forth
- Go
- Haskell
- Icon
- Unicon
- J
- Java
- Liberty BASIC
- Mathematica
- Objeck
- OCaml
- Oz
- Perl
- LWP
- Perl 6
- PHP
- PicoLisp
- PowerShell
- Protium
- PureBasic
- Python
- R
- Stringr
- RCurl
- XML
- REBOL
- Ruby
- Scala
- Scheme
- Seed7
- Tcl
- TUSCRIPT
- TXR
- UNIX Shell
- Ursala
- TI-83 BASIC/Omit
- TI-89 BASIC/Omit
- Visual Basic .NET
- Input Output
- Batch File/Omit
- Locomotive Basic/Omit
- Lotus 123 Macro Scripting/Omit
- M4/Omit
- Openscad/Omit
- PARI/GP/Omit
- PostScript/Omit
- Retro/Omit
- ZX Spectrum Basic/Omit