mirror of
https://github.com/internetarchive/heritrix3.git
synced 2026-11-09 07:24:41 +00:00
Report total bytes, as well as breakdown, in xml version of job page, to match
html.
* CrawlJob.java
sizeTotalsReportData() - add "total" (bytes) and "total-count" (urls) to map
* CrawledBytesHistotable.java
new method getTotalUrls() to complement getTotalBytes()
This commit is contained in:
1 parent
f58fdcdba4
commit
9468cb0c1f
2 files changed
+11
-1
No files matched your search
@@ -31,6 +31,7 @@ import java.util.LinkedHashMap;
|
||||
import java.util.LinkedList;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
import java.util.TreeMap;
|
||||
import java.util.logging.FileHandler;
|
||||
import java.util.logging.Formatter;
|
||||
import java.util.logging.Handler;
|
||||
@@ -797,7 +798,12 @@ public class CrawlJob implements Comparable<CrawlJob>, ApplicationListener {
|
||||
if(stats==null) {
|
||||
return null;
|
||||
}
|
||||
return stats.getCrawledBytes();
|
||||
|
||||
// stats.crawledBytesSummary() also includes totals, so add those in here
|
||||
TreeMap<String, Long> map = new TreeMap<String,Long>(stats.getCrawledBytes());
|
||||
map.put("total", stats.getCrawledBytes().getTotalBytes());
|
||||
map.put("total-count", stats.getCrawledBytes().getTotalUrls());
|
||||
return map;
|
||||
}
|
||||
|
||||
public String sizeTotalsReport() {
|
||||
|
||||
@@ -79,4 +79,8 @@ implements CoreAttributeConstants {
|
||||
public long getTotalBytes() {
|
||||
return get(NOVEL) + get(DUPLICATE) + get(NOTMODIFIED);
|
||||
}
|
||||
|
||||
public long getTotalUrls() {
|
||||
return get(NOVELCOUNT) + get(DUPLICATECOUNT) + get(NOTMODIFIEDCOUNT);
|
||||
}
|
||||
}
|
||||
Reference in new issue
Block a user