boxtec
/
yacy_search_server
mirror of https://github.com/yacy/yacy_search_server
4
1
Fork 0
Code Issues Projects Releases Wiki Activity

show IP of crawled host and country in CrawlResults

Browse Source 
git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@7981 6c8d7289-2bf4-0310-a012-ef5d649a1542
pull/1/head
orbiter 13 years ago
parent b250e6466d
commit 9c131adeb6
2 changed files with 60 additions and 38 deletions
Show Stats Download Patch File Download Diff File

5
htroot/CrawlResults.html
Unescape View File

@ -117,6 +117,8 @@
#(showDate)#::<td><strong>Modified</strong></td>#(/showDate)# #(showDate)#::<td><strong>Modified</strong></td>#(/showDate)#
#(showWords)#::<td><strong>Words</strong></td>#(/showWords)# #(showWords)#::<td><strong>Words</strong></td>#(/showWords)#
#(showTitle)#::<td><strong>Title</strong></td>#(/showTitle)# #(showTitle)#::<td><strong>Title</strong></td>#(/showTitle)#
#(showCountry)#::<td><strong>Country</strong></td>#(/showCountry)#
#(showIP)#::<td><strong>IP of Host</strong></td>#(/showIP)#
#(showURL)#::<td><strong>URL</strong></td>#(/showURL)# #(showURL)#::<td><strong>URL</strong></td>#(/showURL)#
</tr> </tr>
#{indexed}# #{indexed}#
@ -147,6 +149,9 @@
</td> </td>
#(/showTitle)# #(/showTitle)#
#(showCountry)#::<td>#[country]#</td>#(/showCountry)#
#(showIP)#::<td>#[ip]#</td>#(/showIP)#
#(showURL)# #(showURL)#
:: ::
<td> <td>

83
htroot/CrawlResults.java
Unescape View File

@ -40,7 +40,6 @@ import net.yacy.kelondro.logging.Log;
import net.yacy.peers.yacySeed; import net.yacy.peers.yacySeed;
import net.yacy.search.Switchboard; import net.yacy.search.Switchboard;
import net.yacy.search.index.Segments; import net.yacy.search.index.Segments;
import de.anomic.crawler.ResultURLs; import de.anomic.crawler.ResultURLs;
import de.anomic.crawler.ResultURLs.EventOrigin; import de.anomic.crawler.ResultURLs.EventOrigin;
import de.anomic.crawler.ResultURLs.InitExecEntry; import de.anomic.crawler.ResultURLs.InitExecEntry;
@ -56,12 +55,14 @@ public class CrawlResults {
final serverObjects prop = new serverObjects(); final serverObjects prop = new serverObjects();
int lines = 500; int lines = 500;
boolean showInit = env.getConfigBool("IndexMonitorInit", false); boolean showInit = env.getConfigBool("IndexMonitorInit", false);
boolean showExec = env.getConfigBool("IndexMonitorExec", false); boolean showExec = env.getConfigBool("IndexMonitorExec", false);
boolean showDate = env.getConfigBool("IndexMonitorDate", true); boolean showDate = env.getConfigBool("IndexMonitorDate", true);
boolean showWords = env.getConfigBool("IndexMonitorWords", true); boolean showWords = env.getConfigBool("IndexMonitorWords", true);
boolean showTitle = env.getConfigBool("IndexMonitorTitle", true); boolean showTitle = env.getConfigBool("IndexMonitorTitle", true);
boolean showURL = env.getConfigBool("IndexMonitorURL", true); boolean showCountry = env.getConfigBool("IndexMonitorCountry", true);
boolean showIP = env.getConfigBool("IndexMonitorIP", true);
boolean showURL = env.getConfigBool("IndexMonitorURL", true);
if (post == null) { if (post == null) {
post = new serverObjects(); post = new serverObjects();
@ -128,7 +129,7 @@ public class CrawlResults {
try { try {
sb.indexSegments.urlMetadata(Segments.Process.LOCALCRAWLING).deleteDomain(hashpart); sb.indexSegments.urlMetadata(Segments.Process.LOCALCRAWLING).deleteDomain(hashpart);
ResultURLs.deleteDomain(tabletype, domain, hashpart); ResultURLs.deleteDomain(tabletype, domain, hashpart);
} catch (IOException e) { } catch (final IOException e) {
Log.logException(e); Log.logException(e);
} }
} }
@ -138,12 +139,14 @@ public class CrawlResults {
lines = post.getInt("showIndexed", 500); lines = post.getInt("showIndexed", 500);
} }
if (post.get("si") != null) showInit = !("0".equals(post.get("si"))); if (post.get("si") != null) showInit = !("0".equals(post.get("si")));
if (post.get("se") != null) showExec = !("0".equals(post.get("se"))); if (post.get("se") != null) showExec = !("0".equals(post.get("se")));
if (post.get("sd") != null) showDate = !("0".equals(post.get("sd"))); if (post.get("sd") != null) showDate = !("0".equals(post.get("sd")));
if (post.get("sw") != null) showWords = !("0".equals(post.get("sw"))); if (post.get("sw") != null) showWords = !("0".equals(post.get("sw")));
if (post.get("st") != null) showTitle = !("0".equals(post.get("st"))); if (post.get("st") != null) showTitle = !("0".equals(post.get("st")));
if (post.get("su") != null) showURL = !("0".equals(post.get("su"))); if (post.get("sc") != null) showCountry = !("0".equals(post.get("sc")));
if (post.get("sp") != null) showIP = !("0".equals(post.get("sp")));
if (post.get("su") != null) showURL = !("0".equals(post.get("su")));
} // end != null } // end != null
// create table // create table
@ -164,12 +167,14 @@ public class CrawlResults {
prop.putHTML("table_feedbackpage", "CrawlResults.html"); prop.putHTML("table_feedbackpage", "CrawlResults.html");
prop.put("table_tabletype", tabletype.getCode()); prop.put("table_tabletype", tabletype.getCode());
prop.put("table_showInit", (showInit) ? "1" : "0"); prop.put("table_showInit", (showInit) ? "1" : "0");
prop.put("table_showExec", (showExec) ? "1" : "0"); prop.put("table_showExec", (showExec) ? "1" : "0");
prop.put("table_showDate", (showDate) ? "1" : "0"); prop.put("table_showDate", (showDate) ? "1" : "0");
prop.put("table_showWords", (showWords) ? "1" : "0"); prop.put("table_showWords", (showWords) ? "1" : "0");
prop.put("table_showTitle", (showTitle) ? "1" : "0"); prop.put("table_showTitle", (showTitle) ? "1" : "0");
prop.put("table_showURL", (showURL) ? "1" : "0"); prop.put("table_showCountry", (showCountry) ? "1" : "0");
prop.put("table_showIP", (showIP) ? "1" : "0");
prop.put("table_showURL", (showURL) ? "1" : "0");
boolean dark = true; boolean dark = true;
String urlstr, urltxt; String urlstr, urltxt;
@ -229,27 +234,39 @@ public class CrawlResults {
if (showTitle) { if (showTitle) {
prop.put("table_indexed_" + cnt + "_showTitle", (showTitle) ? "1" : "0"); prop.put("table_indexed_" + cnt + "_showTitle", (showTitle) ? "1" : "0");
prop.put("table_indexed_" + cnt + "_showTitle_available", "1"); prop.put("table_indexed_" + cnt + "_showTitle_available", "1");
if (metadata == null || metadata.dc_title() == null || metadata.dc_title().trim().length() == 0) if (metadata == null || metadata.dc_title() == null || metadata.dc_title().trim().length() == 0)
prop.put("table_indexed_" + cnt + "_showTitle_available_nodescr", "0"); prop.put("table_indexed_" + cnt + "_showTitle_available_nodescr", "0");
else { else {
prop.put("table_indexed_" + cnt + "_showTitle_available_nodescr", "1"); prop.put("table_indexed_" + cnt + "_showTitle_available_nodescr", "1");
prop.putHTML("table_indexed_" + cnt + "_showTitle_available_nodescr_urldescr", metadata.dc_title()); prop.putHTML("table_indexed_" + cnt + "_showTitle_available_nodescr_urldescr", metadata.dc_title());
} }
prop.put("table_indexed_" + cnt + "_showTitle_available_urlHash", entry.getKey()); prop.put("table_indexed_" + cnt + "_showTitle_available_urlHash", entry.getKey());
prop.putHTML("table_indexed_" + cnt + "_showTitle_available_urltitle", urlstr); prop.putHTML("table_indexed_" + cnt + "_showTitle_available_urltitle", urlstr);
} else } else
prop.put("table_indexed_" + cnt + "_showTitle", "0"); prop.put("table_indexed_" + cnt + "_showTitle", "0");
if (showCountry && urle != null) {
prop.put("table_indexed_" + cnt + "_showCountry", "1");
prop.put("table_indexed_" + cnt + "_showCountry_country", metadata.url().getLocale().getCountry());
} else
prop.put("table_indexed_" + cnt + "_showCountry", "0");
if (showIP && urle != null) {
prop.put("table_indexed_" + cnt + "_showIP", "1");
prop.put("table_indexed_" + cnt + "_showIP_ip", metadata.url().getInetAddress().getHostAddress());
} else
prop.put("table_indexed_" + cnt + "_showIP", "0");
if (showURL) { if (showURL) {
prop.put("table_indexed_" + cnt + "_showURL", "1"); prop.put("table_indexed_" + cnt + "_showURL", "1");
prop.put("table_indexed_" + cnt + "_showURL_available", "1"); prop.put("table_indexed_" + cnt + "_showURL_available", "1");
prop.put("table_indexed_" + cnt + "_showURL_available_urlHash", entry.getKey()); prop.put("table_indexed_" + cnt + "_showURL_available_urlHash", entry.getKey());
prop.putHTML("table_indexed_" + cnt + "_showURL_available_urltitle", urlstr); prop.putHTML("table_indexed_" + cnt + "_showURL_available_urltitle", urlstr);
prop.put("table_indexed_" + cnt + "_showURL_available_url", urltxt); prop.put("table_indexed_" + cnt + "_showURL_available_url", urltxt);
} else } else
prop.put("table_indexed_" + cnt + "_showURL", "0"); prop.put("table_indexed_" + cnt + "_showURL", "0");

Write Preview
Loading…
Cancel
Save