- patch for bad web structure dumps

- added automatic slow down of accessed to specific domains when access to a web page fails

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@5765 6c8d7289-2bf4-0310-a012-ef5d649a1542

source/de/anomic/crawler/Latency.java

@@ -45,6 +45,17 @@ public class Latency {
         }
     }
     
+    public static void slowdown(String hosthash, String host) {
+        assert hosthash.length() == 6;
+        Host h = map.get(hosthash);
+        if (h == null) {
+            h = new Host(host, 3000);
+            map.put(hosthash, h);
+        } else {
+            h.slowdown();
+        }
+    }
+    
     public static Host host(String hosthash) {
         assert hosthash.length() == 6;
         return map.get(hosthash);
@@ -172,6 +183,11 @@ public class Latency {
             this.timeacc += time;
             this.count++;
         }
+        public void slowdown() {
+            this.lastacc = System.currentTimeMillis();
+            this.timeacc = Math.min(60000, average() * 5);
+            this.count = 1;
+        }
         public int count() {
             return this.count;
         }

source/de/anomic/http/httpClient.java

@@ -59,8 +59,10 @@ import org.apache.commons.httpclient.params.HttpMethodParams;
 import org.apache.commons.httpclient.protocol.Protocol;
 import org.apache.commons.httpclient.protocol.ProtocolSocketFactory;
 
+import de.anomic.crawler.Latency;
 import de.anomic.kelondro.order.Base64Order;
 import de.anomic.kelondro.util.Log;
+import de.anomic.yacy.yacyURL;
 
 /**
  * HttpClient implementation which uses Jakarta Commons HttpClient 3.x {@link http://hc.apache.org/httpclient-3.x/}
@@ -448,14 +450,20 @@ public class httpClient {
             }
         } catch (final IllegalThreadStateException e) {
         	// cleanUp statistics
+            yacyURL url = new yacyURL(method.getURI().toString(), null);
+            Latency.slowdown(url.hash().substring(6), url.getHost());
             HttpConnectionInfo.removeConnection(generateConInfo(method));
             throw e;
         } catch (final IOException e) {
             // cleanUp statistics
+            yacyURL url = new yacyURL(method.getURI().toString(), null);
+            Latency.slowdown(url.hash().substring(6), url.getHost());
             HttpConnectionInfo.removeConnection(generateConInfo(method));
             throw e;
         } catch (final IllegalStateException e) {
             // cleanUp statistics
+            yacyURL url = new yacyURL(method.getURI().toString(), null);
+            Latency.slowdown(url.hash().substring(6), url.getHost());
             HttpConnectionInfo.removeConnection(generateConInfo(method));
             throw new IOException(e.getMessage());
         }

source/de/anomic/plasma/plasmaWebStructure.java

@@ -184,9 +184,15 @@ public class plasmaWebStructure {
         final Map<String, Integer> map = new HashMap<String, Integer>();
         String c;
         final int refsc = refstr2count(refs);
+        int d;
         for (int i = 0; i < refsc; i++) {
             c = refs.substring(8 + i * 10, 8 + (i + 1) * 10);
-            map.put(c.substring(0, 6), Integer.valueOf(c.substring(6), 16));
+            try {
+                d = Integer.valueOf(c.substring(6), 16);
+            } catch (NumberFormatException e) {
+                d = 1;
+            }
+            map.put(c.substring(0, 6), d);
         }
         return map;
     }