summaryrefslogtreecommitdiff
path: root/src/crawl/crawl.conf
diff options
context:
space:
mode:
Diffstat (limited to 'src/crawl/crawl.conf')
-rw-r--r--src/crawl/crawl.conf12
1 files changed, 7 insertions, 5 deletions
diff --git a/src/crawl/crawl.conf b/src/crawl/crawl.conf
index 8169ae1..25e4003 100644
--- a/src/crawl/crawl.conf
+++ b/src/crawl/crawl.conf
@@ -13,6 +13,13 @@ logger = {
level = "DEBUG"
}
+modules = {
+ urlnormalizers = {
+ "mod_normalizer_simple",
+ "mod_normalizer_google"
+ }
+}
+
-- seeds: URLS which are fed in the beginning to the URL frontier
seeds = {
@@ -20,11 +27,6 @@ seeds = {
"http://wolframe.net"
}
-urlnormalizers = {
- "mod_normalizer_simple",
- "mod_normalizer_google"
-}
-
filters = {
-- allowed protocols to be fetched
protocols = {