From 1e14a83ee1f657477ab8ede7720ff8ffbab59d2e Mon Sep 17 00:00:00 2001
From: Doug Coleman <doug.coleman@gmail.com>
Date: Fri, 19 Jun 2009 14:41:48 -0500
Subject: [PATCH] allow robot-identifiers to be set for robots.txt

---
 extra/robots/robots.factor | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)

diff --git a/extra/robots/robots.factor b/extra/robots/robots.factor
index 3c0eb045f7..af039ef8c4 100644
--- a/extra/robots/robots.factor
+++ b/extra/robots/robots.factor
@@ -1,15 +1,18 @@
 ! Copyright (C) 2009 Doug Coleman.
 ! See http://factorcode.org/license.txt for BSD license.
-USING: accessors http.client kernel unicode.categories
-sequences urls splitting combinators splitting.monotonic
-combinators.short-circuit assocs unicode.case arrays
-math.parser calendar.format make fry present globs
-multiline regexp.combinators regexp ;
+USING: accessors arrays assocs calendar.format combinators
+combinators.short-circuit fry globs http.client kernel make
+math.parser multiline namespaces present regexp
+regexp.combinators sequences sets splitting splitting.monotonic
+unicode.case unicode.categories urls ;
 IN: robots
 
 ! visit-time is GMT, request-rate is pages/second 
 ! crawl-rate is seconds
 
+SYMBOL: robot-identities
+robot-identities [ { "FactorSpider" } ] initialize
+
 TUPLE: robots site sitemap rules rules-quot ;
 
 : <robots> ( site sitemap rules -- robots )
@@ -80,6 +83,13 @@ visit-time request-rate crawl-delay unknowns ;
         derive-urls [ <glob> ] map <and> <not>
     ] bi 2array <or> '[ _ matches? ] ;
 
+: relevant-rules ( robots -- rules )
+    [
+        user-agents>> [
+            robot-identities get [ swap glob-matches? ] with any?
+        ] any?
+    ] filter ;
+
 PRIVATE>
 
 : parse-robots.txt ( string -- sitemaps rules-seq )