jenkins-bot has submitted this change and it was merged.
Change subject: add log analysis scripts
......................................................................
add log analysis scripts
For internal use only
Change-Id: I0c1303b6eae1f1bae6e94bca1baa4c2893f51ba4
---
A log-analysis/analyze.py
A log-analysis/datediff.pl
A log-analysis/getq.pl
A log-analysis/getsparql.pl
A log-analysis/split.pl
5 files changed, 173 insertions(+), 0 deletions(-)
Approvals:
Smalyshev: Looks good to me, approved
jenkins-bot: Verified
diff --git a/log-analysis/analyze.py b/log-analysis/analyze.py
new file mode 100644
index 0000000..2c8a805
--- /dev/null
+++ b/log-analysis/analyze.py
@@ -0,0 +1,91 @@
+import sys
+import os
+import subprocess
+import os.path
+import re
+
+def getOut(cmd, file=None):
+ if file == None:
+ file = mainfile
+ return subprocess.check_output(cmd % file, shell =
True).rstrip('0').replace('V1', '')
+
+def getSummary(cmd, file=None):
+ return getOut("cat %%s | %s | ~/bin/r summary -" % cmd, file)
+
+meanre = re.compile('\s*Mean\s*:\s*([\d.]+)', re.MULTILINE)
+def getAvg(summary):
+ m = meanre.search(summary)
+ #print m, summary
+ return float(m.group(1))
+
+logfile = sys.argv[1]
+outfile = open(logfile + ".out.md", "w")
+mainfile = logfile + ".part"
+
+if not os.path.exists(mainfile):
+ print "Splitting %s into parts" % logfile
+ os.system("perl split.pl %s" % logfile)
+
+print ("Main file %s " % mainfile)
+
+batches = int(getOut("cat %s | grep 'changes in batch' | wc -l"))
+print "%d batches in file" % batches
+outfile.write("Logfile %s contains %d batches\n\n" % (logfile, batches))
+outfile.write("= Batch Stats =\n")
+
+modified_triples = getSummary("grep 'and modified' | cut -d ' ' -f 13")
+modified_avg = getAvg(modified_triples)
+
+changes_per_batch = getSummary("grep 'changes in batch' | cut -d ' ' -f 6")
+print "%f changes per batch" % getAvg(changes_per_batch)
+outfile.write("Changes per batch:\n%s\n" % changes_per_batch)
+
+filtered_per_batch = getSummary("grep 'Filtered' | cut -d ' ' -f 9")
+print "%f real changes per batch" % getAvg(filtered_per_batch)
+outfile.write("Filtered changes per batch:\n%s\n" % filtered_per_batch)
+filter_avg = getAvg(filtered_per_batch)
+
+time_per_batch = getSummary("grep 'Polled up to' | cut -d ' ' -f 1 | perl -ne
'/(\d+):(\d+):(\d+)\.(\d+)/; $ms = $4+1000*($3+60*($2+60*$1)); if($lastms) {
print $ms-$lastms, chr(0xA); } $lastms = $ms;'")
+outfile.write("Time per batch (ms):\n%s\n" % time_per_batch)
+time_avg = getAvg(time_per_batch)
+
+advance_per_batch = getSummary("grep 'Polled up to' | cut -d ' ' -f 10 | perl
-ne '/T(\d+):(\d+):(\d+)Z/; $ms = ($3+60*($2+60*$1)); if($lastms) { print
$ms-$lastms,chr(0xA); } $lastms = $ms;'")
+outfile.write("Advance per batch (s):\n%s\n" % advance_per_batch)
+adv_avg = getAvg(advance_per_batch)
+
+outfile.write("Modified triples:\n%s\n" % modified_triples)
+
+outfile.write("Modified triples - %f per entity\n" % (modified_avg/filter_avg))
+
+outfile.write("= Processing Stats =\n")
+
+outfile.write("Processing time per change: %f ms, %f updates/second\n"
%(time_avg/filter_avg, filter_avg/time_avg*1000))
+print ("Processing time per change: %f ms, %f updates/second\n"
%(time_avg/filter_avg, filter_avg/time_avg*1000))
+
+print "Stream is %f updates/s\n" % (filter_avg/adv_avg)
+outfile.write("\nStream is %f updates/s\n" % (filter_avg/adv_avg))
+outfile.write("\nUpdate/stream ratio is %f\n\n" % (adv_avg/time_avg*1000))
+
+update_preparation = getSummary("grep 'Preparing update' | cut -d ' ' -f 10")
+outfile.write("Update preparation time (ms):\n%s\n" % update_preparation)
+
+sparqlname = logfile + ".q"
+if not os.path.exists(sparqlname):
+ os.system("cat %s | perl getsparql.pl | sort > %s" %(logfile, sparqlname))
+print "Queries in %s" % sparqlname
+
+filter_times = getSummary("grep 'SELECT DISTINCT ?s WHERE' | cut -d ' ' -f 7",
sparqlname)
+update_times = getSummary("grep 'Clear' | cut -d ' ' -f 9", sparqlname)
+tsupdate_times = getSummary("grep 'DELETE {' | cut -d ' ' -f 4", sparqlname)
+values_times = getSummary("grep 'SELECT DISTINCT ?entity ?s' | cut -d ' ' -f
6", sparqlname)
+
+outfile.write("== Query Stats ==\n")
+outfile.write("FILTER query (ms):\n%s\n" % filter_times)
+outfile.write("SELECT values query (ms):\n%s\n" % values_times)
+outfile.write("UPDATE query (ms):\n%s\n" % update_times)
+outfile.write("Timestamp UPDATE query (ms):\n%s\n" % tsupdate_times)
+
+update_avg = getAvg(update_times)
+
+outfile.write("Update query speed is %f entities/s, %f triples/s\n" %
(filter_avg/update_avg*1000, modified_avg/update_avg*1000))
+os.system("cat %s.out.md | pbcopy" % logfile)
diff --git a/log-analysis/datediff.pl b/log-analysis/datediff.pl
new file mode 100644
index 0000000..b8b2db4
--- /dev/null
+++ b/log-analysis/datediff.pl
@@ -0,0 +1,7 @@
+while(<>) {
+if(/(\d+):(\d+):(\d+)\.(\d+)/) {
+ $ms = $4+1000*($3+60*($2+60*$1));
+ if($lastms) { print $ms-$lastms, "\n"; }
+ $lastms = $ms;
+}
+}
diff --git a/log-analysis/getq.pl b/log-analysis/getq.pl
new file mode 100644
index 0000000..b6dd906
--- /dev/null
+++ b/log-analysis/getq.pl
@@ -0,0 +1,32 @@
+$qtype = undef;
+while(<>) {
+ if(/Running SPARQL/) {
+ $qtype = 'other';
+ $cnt = 1;
+ }
+ if(/Completed in (\d+)/) {
+ if($qtype) {
+ print "$1 $qtype $cnt\n";
+ }
+ $qtype = undef;
+ next;
+ }
+ next unless $qtype;
+ $cnt++;
+ if(/ASK {/) {
+ $qtype = 'version';
+ }
+ if(/SELECT DISTINCT \?s/) {
+ $qtype = 'fetch'
+ }
+ if($qtype eq 'fetch' && m|<http://www.w3.org/ns/prov#wasDerivedFrom>|) {
+ $qtype = 'refs';
+ }
+ if($qtype eq 'refs' && /} UNION {/) {
+ $qtype = 'values';
+ }
+ if(/INSERT {/) {
+ $qtype = "update";
+ }
+}
+
diff --git a/log-analysis/getsparql.pl b/log-analysis/getsparql.pl
new file mode 100644
index 0000000..57b45b7
--- /dev/null
+++ b/log-analysis/getsparql.pl
@@ -0,0 +1,15 @@
+$sparql = 0;
+$name = "wtf";
+while(<>) {
+ if(!$sparql) {
+ if(/Running SPARQL: (.+)/) {
+ $name = $1;
+ $sparql = 1;
+ }
+ } else {
+ if(/- Completed in (\d+) ms/) {
+ $sparql = 0;
+ print "$name => $1\n";
+ }
+ }
+}
\ No newline at end of file
diff --git a/log-analysis/split.pl b/log-analysis/split.pl
new file mode 100644
index 0000000..5251d7b
--- /dev/null
+++ b/log-analysis/split.pl
@@ -0,0 +1,28 @@
+
+$f = $ARGV[0];
+
+open $in, $ARGV[0] or die $!;
+%FILES = {};
+$out = '';
+$outf = STDOUT;
+
+while(<$in>) {
+ if(/\[(main|update (\d+))\]/) {
+ $newout = $1;
+ if($out ne $newout) {
+ $outf = switchout($newout, $2);
+ $out = $newout;
+ }
+ }
+ print $outf $_;
+}
+
+sub switchout {
+ my ($newout, $name) = @_;
+ if(!defined($FILES{$newout})) {
+ my $f = $ARGV[0].".part".$name;
+ print "Opening $f for $newout\n";
+ open $FILES{$newout}, ">", $f;
+ }
+ return $FILES{$newout};
+}
--
To view, visit https://gerrit.wikimedia.org/r/240769
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: merged
Gerrit-Change-Id: I0c1303b6eae1f1bae6e94bca1baa4c2893f51ba4
Gerrit-PatchSet: 1
Gerrit-Project: wikidata/query/rdf
Gerrit-Branch: master
Gerrit-Owner: Smalyshev <[email protected]>
Gerrit-Reviewer: Smalyshev <[email protected]>
Gerrit-Reviewer: jenkins-bot <>
_______________________________________________
MediaWiki-commits mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits
