honeymoose/OpenSearch
1
0
Fork 0
mirror of https://github.com/honeymoose/OpenSearch.git synced 2025-02-22 21:05:23 +00:00
Code Issues Packages Projects Releases Wiki Activity

[ML] fixing and unmuting testHRDSplit test (#55349) (#55393)

Browse Source 
This fixes the long muted testHRDSplit. Some minor adjustments for modern day elasticsearch changes :). 

The cause of the failure is that a new `by` field entering the model with an exceptionally high count does not cause an anomaly. We have since stopped combining the `rare` and `by` in this manner. New entries in a `by` field are not anomalous because we have no history on them yet. 

closes https://github.com/elastic/elasticsearch/issues/32966
This commit is contained in:
Benjamin Trent 2020-04-17 09:55:52 -04:00 committed by GitHub
parent 13426ce92f
commit 8c581c3388
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 64 additions and 46 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

110
x-pack/plugin/ml/qa/single-node-tests/src/test/java/org/elasticsearch/xpack/ml/transforms/PainlessDomainSplitIT.java
View File

@ -14,7 +14,7 @@ import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.test.rest.ESRestTestCase;
import org.elasticsearch.xpack.ml.MachineLearning;
import java.time.Clock;
import java.time.ZoneOffset;
import java.time.ZonedDateTime;
import java.time.format.DateTimeFormatter;
import java.util.ArrayList;
@ -25,6 +25,7 @@ import java.util.regex.Matcher;
import java.util.regex.Pattern;
import static org.elasticsearch.common.xcontent.XContentFactory.jsonBuilder;
import static org.hamcrest.Matchers.containsString;
import static org.hamcrest.Matchers.equalTo;
public class PainlessDomainSplitIT extends ESRestTestCase {
@ -239,7 +240,6 @@ public class PainlessDomainSplitIT extends ESRestTestCase {
}
}
@AwaitsFix(bugUrl = "https://github.com/elastic/elasticsearch/issues/32966")
public void testHRDSplit() throws Exception {
// Create job
Request createJobRequest = new Request("PUT", MachineLearning.BASE_PATH + "anomaly_detectors/hrd-split-job");
@ -268,7 +268,7 @@ public class PainlessDomainSplitIT extends ESRestTestCase {
"\"time\": { \"type\": \"date\" } }");
// Index some data
ZonedDateTime baseTime = ZonedDateTime.now(Clock.systemDefaultZone()).minusYears(1);
ZonedDateTime baseTime = ZonedDateTime.now(ZoneOffset.UTC).minusYears(1);
TestConfiguration test = tests.get(randomInt(tests.size()-1));
// domainSplit() tests had subdomain, testHighestRegisteredDomainCases() did not, so we need a special case for sub
@ -276,20 +276,18 @@ public class PainlessDomainSplitIT extends ESRestTestCase {
String expectedHRD = test.domainExpected.replace(".", "\\.");
Pattern pattern = Pattern.compile("domain_split\":\\[\"(" + expectedSub + "),(" + expectedHRD +")\"[,\\]]");
for (int i = 0; i < 100; i++) {
for (int i = 1; i <= 100; i++) {
ZonedDateTime time = baseTime.plusHours(i);
if (i == 64) {
String formattedTime = time.format(DateTimeFormatter.ISO_DATE_TIME);
if (i % 50 == 0) {
// Anomaly has 100 docs, but we don't care about the value
for (int j = 0; j < 100; j++) {
String formattedTime = time.format(DateTimeFormatter.ISO_DATE_TIME);
Request createDocRequest = new Request("PUT", "/painless/_doc/" + formattedTime + "_" + j);
Request createDocRequest = new Request("POST", "/painless/_doc");
createDocRequest.setJsonEntity("{\"domain\": \"" + "bar.bar.com\", \"time\": \"" + formattedTime + "\"}");
client().performRequest(createDocRequest);
}
} else {
// Non-anomalous values will be what's seen when the anomaly is reported
String formattedTime = time.format(DateTimeFormatter.ISO_DATE_TIME);
Request createDocRequest = new Request("PUT", "/painless/_doc/" + formattedTime);
createDocRequest.setJsonEntity("{\"domain\": \"" + test.hostName + "\", \"time\": \"" + formattedTime + "\"}");
client().performRequest(createDocRequest);
@ -304,7 +302,6 @@ public class PainlessDomainSplitIT extends ESRestTestCase {
"{\n" +
" \"job_id\":\"hrd-split-job\",\n" +
" \"indexes\":[\"painless\"],\n" +
" \"types\":[\"_doc\"],\n" +
" \"script_fields\": {\n" +
" \"domain_split\": {\n" +
" \"script\": \"return domainSplit(doc['domain'].value, params);\"\n" +
@ -313,46 +310,67 @@ public class PainlessDomainSplitIT extends ESRestTestCase {
"}");
client().performRequest(createFeedRequest);
client().performRequest(new Request("POST", MachineLearning.BASE_PATH + "datafeeds/hrd-split-datafeed/_start"));
Request startDatafeedRequest = new Request("POST", MachineLearning.BASE_PATH + "datafeeds/hrd-split-datafeed/_start");
startDatafeedRequest.addParameter("start", baseTime.format(DateTimeFormatter.ISO_DATE_TIME));
startDatafeedRequest.addParameter("end", ZonedDateTime.now(ZoneOffset.UTC).format(DateTimeFormatter.ISO_DATE_TIME));
client().performRequest(startDatafeedRequest);
try {
assertBusy(() -> {
client().performRequest(new Request("POST", "/_refresh"));
waitUntilDatafeedIsStopped("hrd-split-datafeed");
waitUntilJobIsClosed("hrd-split-job");
Response response = client().performRequest(new Request("GET",
MachineLearning.BASE_PATH + "anomaly_detectors/hrd-split-job/results/records"));
String responseBody = EntityUtils.toString(response.getEntity());
client().performRequest(new Request("POST", "/.ml-anomalies-*/_refresh"));
if (responseBody.contains("\"count\":2")) {
Matcher m = pattern.matcher(responseBody);
Response records = client().performRequest(new Request("GET",
MachineLearning.BASE_PATH + "anomaly_detectors/hrd-split-job/results/records"));
String responseBody = EntityUtils.toString(records.getEntity());
assertThat("response body [" + responseBody + "] did not contain [\"count\":2]",
responseBody,
containsString("\"count\":2"));
String actualSubDomain = "";
String actualDomain = "";
if (m.find()) {
actualSubDomain = m.group(1).replace("\"", "");
actualDomain = m.group(2).replace("\"", "");
}
String expectedTotal = "[" + test.subDomainExpected + "," + test.domainExpected + "]";
String actualTotal = "[" + actualSubDomain + "," + actualDomain + "]";
// domainSplit() tests had subdomain, testHighestRegisteredDomainCases() do not
if (test.subDomainExpected != null) {
assertThat("Expected subdomain [" + test.subDomainExpected + "] but found [" + actualSubDomain
+ "]. Actual " + actualTotal + " vs Expected " + expectedTotal, actualSubDomain,
equalTo(test.subDomainExpected));
}
assertThat("Expected domain [" + test.domainExpected + "] but found [" + actualDomain + "]. Actual "
+ actualTotal + " vs Expected " + expectedTotal, actualDomain, equalTo(test.domainExpected));
} else {
logger.error(responseBody);
fail("Response body didn't contain [\"count\":2]");
}
}, 5, TimeUnit.SECONDS);
} catch (Exception e) {
logger.error(e.getMessage());
fail("Anomaly records were not found within 5 seconds");
Matcher m = pattern.matcher(responseBody);
String actualSubDomain = "";
String actualDomain = "";
if (m.find()) {
actualSubDomain = m.group(1).replace("\"", "");
actualDomain = m.group(2).replace("\"", "");
}
String expectedTotal = "[" + test.subDomainExpected + "," + test.domainExpected + "]";
String actualTotal = "[" + actualSubDomain + "," + actualDomain + "]";
// domainSplit() tests had subdomain, testHighestRegisteredDomainCases() do not
if (test.subDomainExpected != null) {
assertThat("Expected subdomain [" + test.subDomainExpected + "] but found [" + actualSubDomain
+ "]. Actual " + actualTotal + " vs Expected " + expectedTotal, actualSubDomain,
equalTo(test.subDomainExpected));
}
assertThat("Expected domain [" + test.domainExpected + "] but found [" + actualDomain + "]. Actual "
+ actualTotal + " vs Expected " + expectedTotal, actualDomain, equalTo(test.domainExpected));
}
private void waitUntilJobIsClosed(String jobId) throws Exception {
assertBusy(() -> {
try {
Response jobStatsResponse = client().performRequest(new Request("GET",
MachineLearning.BASE_PATH + "anomaly_detectors/" + jobId + "/_stats"));
assertThat(EntityUtils.toString(jobStatsResponse.getEntity()), containsString("\"state\":\"closed\""));
} catch (Exception e) {
throw new RuntimeException(e);
}
});
}
private void waitUntilDatafeedIsStopped(String dfId) throws Exception {
assertBusy(() -> {
try {
Response datafeedStatsResponse = client().performRequest(new Request("GET",
MachineLearning.BASE_PATH + "datafeeds/" + dfId + "/_stats"));
assertThat(EntityUtils.toString(datafeedStatsResponse.getEntity()),
containsString("\"state\":\"stopped\""));
} catch (Exception e) {
throw new RuntimeException(e);
}
}, 60, TimeUnit.SECONDS);
}
}