Skip to content

Commit

Permalink
Merge pull request #31 from hobbit-project/develop
Browse files Browse the repository at this point in the history 
add query type inclusion/exclusion feature
  • Loading branch information
@vpapako
vpapako authored Jun 15, 2018
2 parents 4314104 + 3fefdc5 commit afae5d1
Show file tree
Hide file tree
Showing 4 changed files with 124 additions and 88 deletions.
1 change: 1 addition & 0 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@ The data generator of SPVB uses the data generator of Linked Data Benchmark Coun
* **Version insertion ratio** defines the proportion of added triples between two consecutive versions. The default value is `15%`.
* **Version deletion ratio** defines the proportion of deleted triples between two consecutive versions. The default value is `10%`.
* **Generated data form**: Since each system implements a different versioning strategy, it requires the generated data in a specific form. SPVB’s data generator can output the data as i) an *Independent Copy* (suitable for systems that implement the full materialization strategy), ii) as a *Changeset* − set of added and deleted triples (suitable for systems implementing the delta-based or annotated triples versioning strategies) or iii) both as an independent copy and changeset (suitable for systems implementing a hybrid strategy). The default value is `Independent Copies (IC)`.
* **Enable/Disable Query Types** let the system under test to determine the query types for which will be tested in.
* **Generator seed** is used to set the random seed for the data generator. This seed is used to control all random data generation happening in SPVB. The default value is `100`.

The generated SPARQL queries are of different types and are partially based on a subset of the 25 query templates defined in the context of DBpedia SPARQL Benchmark ([DBPSB](http://aksw.org/Projects/DBPSB.html)).
Expand Down
11 changes: 7 additions & 4 deletions src/main/java/org/hobbit/benchmark/versioning/components/VersioningBenchmarkController.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -23,9 +23,9 @@ public class VersioningBenchmarkController extends AbstractBenchmarkController {

private static final Logger LOGGER = LoggerFactory.getLogger(VersioningBenchmarkController.class);

private static final String DATA_GENERATOR_CONTAINER_IMAGE = "git.project-hobbit.eu:4567/papv/versioningdatagenerator:2.1.0";
private static final String TASK_GENERATOR_CONTAINER_IMAGE = "git.project-hobbit.eu:4567/papv/versioningtaskgenerator:2.1.0";
private static final String EVALUATION_MODULE_CONTAINER_IMAGE = "git.project-hobbit.eu:4567/papv/versioningevaluationmodule:2.1.0";
private static final String DATA_GENERATOR_CONTAINER_IMAGE = "git.project-hobbit.eu:4567/papv/versioningdatagenerator:2.2.0";
private static final String TASK_GENERATOR_CONTAINER_IMAGE = "git.project-hobbit.eu:4567/papv/versioningtaskgenerator:2.2.0";
private static final String EVALUATION_MODULE_CONTAINER_IMAGE = "git.project-hobbit.eu:4567/papv/versioningevaluationmodule:2.2.0";

private static final String PREFIX = "http://w3id.org/hobbit/versioning-benchmark/vocab#";

Expand Down Expand Up @@ -62,6 +62,7 @@ public void init() throws Exception {
int insRatio = (Integer) getPropertyOrDefault(PREFIX + "versionInsertionRatio", 5);
int delRatio = (Integer) getPropertyOrDefault(PREFIX + "versionDeletionRatio", 3);
String dataForm = (String) getPropertyOrDefault(PREFIX + "generatedDataForm", "ic");
String enabledQueries = (String) getPropertyOrDefault(PREFIX + "enableDisableQT", "QT1=1;QT2=1;QT3=1;QT4=1;QT5=1;QT6=1;QT7=1;QT8=1");

loadingTimes = new long[numOfVersions];
triplesToBeAdded = new AtomicIntegerArray(numOfVersions);
Expand All @@ -76,8 +77,10 @@ public void init() throws Exception {
VersioningConstants.NUMBER_OF_VERSIONS + "=" + numOfVersions,
VersioningConstants.VERSION_INSERTION_RATIO + "=" + insRatio,
VersioningConstants.VERSION_DELETION_RATIO + "=" + delRatio,
VersioningConstants.SENT_DATA_FORM + "=" + dataForm
VersioningConstants.SENT_DATA_FORM + "=" + dataForm,
VersioningConstants.ENABLED_QUERY_TYPES + "=" + enabledQueries
};
LOGGER.info(enabledQueries);

// evaluation module environmental values
evalModuleEnvVariables = new String[] {
Expand Down
196 changes: 113 additions & 83 deletions src/main/java/org/hobbit/benchmark/versioning/components/VersioningDataGenerator.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@
import java.io.IOException;
import java.io.InputStreamReader;
import java.io.OutputStreamWriter;
import java.io.StringReader;
import java.lang.reflect.Constructor;
import java.nio.ByteBuffer;
import java.nio.charset.StandardCharsets;
Expand All @@ -18,6 +19,7 @@
import java.util.List;
import java.util.Locale;
import java.util.Map;
import java.util.Properties;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Semaphore;
Expand All @@ -27,10 +29,8 @@
import org.apache.commons.io.FileUtils;
import org.apache.commons.io.filefilter.RegexFileFilter;
import org.apache.commons.lang3.SerializationUtils;
import org.apache.jena.query.Query;
import org.apache.jena.query.QueryExecution;
import org.apache.jena.query.QueryExecutionFactory;
import org.apache.jena.query.QueryFactory;
import org.apache.jena.query.ResultSet;
import org.apache.jena.query.ResultSetFactory;
import org.apache.jena.query.ResultSetFormatter;
Expand Down Expand Up @@ -83,6 +83,7 @@ public class VersioningDataGenerator extends AbstractDataGenerator {
private int versionDeletionRatio;
private String serializationFormat;
private String sentDataForm;
private String enabledQueryTypesParam;
private String generatedDatasetPath = "/versioning/data";
private String initialVersionDataPath = generatedDatasetPath + File.separator + "v0";
private String ontologiesPath = "/versioning/ontologies";
Expand All @@ -94,6 +95,8 @@ public class VersioningDataGenerator extends AbstractDataGenerator {
private int[] triplesExpectedToBeLoaded;
private int[] cwsToBeLoaded;

private Properties enabledQueryTypes = new Properties();

private AtomicInteger numberOfmessages = new AtomicInteger(0);

private Configuration configuration = new Configuration();
Expand Down Expand Up @@ -147,6 +150,9 @@ public void init() throws Exception {
triplesExpectedToBeLoaded = new int[numberOfVersions];
cwsToBeLoaded = new int[numberOfVersions];

// load the enabled queries
enabledQueryTypes.load(new StringReader(enabledQueryTypesParam.replaceAll(";", "\n")));

// Given the above input, update configuration files that are necessary for data generation
reInitializeSPBProperties();

Expand Down Expand Up @@ -320,17 +326,18 @@ public void run() {
}

public void initFromEnv() {
LOGGER.info("Getting Data Generator's properites from the environment...");
LOGGER.info("Getting data generator configuration parameters...");

Map<String, String> env = System.getenv();
// Assume that in v0Size the 40362 triples of DBpedia initial dataset
// plus the 8135 ontologies triples are included
v0SizeInTriples = (Integer) getFromEnv(env, VersioningConstants.V0_SIZE_IN_TRIPLES, 0) - 48497 ;
numberOfVersions = (Integer) getFromEnv(env, VersioningConstants.NUMBER_OF_VERSIONS, 0);
subGeneratorSeed = (Integer) getFromEnv(env, VersioningConstants.DATA_GENERATOR_SEED, 0) + getGeneratorId();
versionInsertionRatio = (Integer) getFromEnv(env, VersioningConstants.VERSION_INSERTION_RATIO, 5);
versionDeletionRatio = (Integer) getFromEnv(env, VersioningConstants.VERSION_DELETION_RATIO, 3);
versionInsertionRatio = (Integer) getFromEnv(env, VersioningConstants.VERSION_INSERTION_RATIO, 0);
versionDeletionRatio = (Integer) getFromEnv(env, VersioningConstants.VERSION_DELETION_RATIO, 0);
sentDataForm = (String) getFromEnv(env, VersioningConstants.SENT_DATA_FORM, "");
enabledQueryTypesParam = (String) getFromEnv(env, VersioningConstants.ENABLED_QUERY_TYPES, "");
}

/*
Expand Down Expand Up @@ -588,27 +595,34 @@ public void buildSPRQLQueries() {
String queryString;

// QT1
queryString = compileMustacheTemplate(1, queryIndex++, 0);
tasks.add(new Task(1, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery1.1.1.sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
if(enabledQueryTypes.getProperty("QT1").equals("1")) {
queryString = compileMustacheTemplate(1, queryIndex, 0);
tasks.add(new Task(1, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery1.1.1.sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
}
}
queryIndex++;

// QT2
querySubstParamCount = 5;
for (int querySubType = 1; querySubType <= Statistics.VERSIONING_SUB_QUERIES_COUNT; querySubType++) {
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount; querySubstParam++) {
queryString = compileMustacheTemplate(2, queryIndex, querySubstParam);
tasks.add(new Task(2, querySubType, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery2." + querySubType + "." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
if(enabledQueryTypes.getProperty("QT2").equals("1")) {
querySubstParamCount = 5;
for (int querySubType = 1; querySubType <= Statistics.VERSIONING_SUB_QUERIES_COUNT; querySubType++) {
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount; querySubstParam++) {
queryString = compileMustacheTemplate(2, queryIndex, querySubstParam);
tasks.add(new Task(2, querySubType, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery2." + querySubType + "." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
}
}
queryIndex++;
}
queryIndex++;
} else {
queryIndex += Statistics.VERSIONING_SUB_QUERIES_COUNT;
}

// if there is only one version return
Expand All @@ -617,33 +631,39 @@ public void buildSPRQLQueries() {

// QT3
// if the total number of versions is lower than 1 there are no historical versions
querySubstParamCount = 3;
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount && querySubstParam < numberOfVersions ; querySubstParam++) {
queryString = compileMustacheTemplate(3, queryIndex, querySubstParam);
tasks.add(new Task(3, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery3.1." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
if(enabledQueryTypes.getProperty("QT3").equals("1")) {
querySubstParamCount = 3;
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount && querySubstParam < numberOfVersions ; querySubstParam++) {
queryString = compileMustacheTemplate(3, queryIndex, querySubstParam);
tasks.add(new Task(3, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery3.1." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
}
}
}
}
queryIndex++;

// QT4
// if the total number of versions is lower than 1 there are no historical versions
querySubstParamCount = 3;
for (int querySubType = 1; querySubType <= Statistics.VERSIONING_SUB_QUERIES_COUNT; querySubType++) {
// if there are less than four versions take all the historical ones
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount && querySubstParam < numberOfVersions; querySubstParam++) {
queryString = compileMustacheTemplate(4, queryIndex, querySubstParam);
tasks.add(new Task(4, querySubType, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery4." + querySubType + "." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
if(enabledQueryTypes.getProperty("QT4").equals("1")) {
querySubstParamCount = 3;
for (int querySubType = 1; querySubType <= Statistics.VERSIONING_SUB_QUERIES_COUNT; querySubType++) {
// if there are less than four versions take all the historical ones
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount && querySubstParam < numberOfVersions; querySubstParam++) {
queryString = compileMustacheTemplate(4, queryIndex, querySubstParam);
tasks.add(new Task(4, querySubType, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery4." + querySubType + "." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
}
}
queryIndex++;
}
queryIndex++;
} else {
queryIndex += Statistics.VERSIONING_SUB_QUERIES_COUNT;
}

// if the total number of versions is lower than 5 read deltas for all the historical ones
Expand All @@ -654,65 +674,75 @@ public void buildSPRQLQueries() {
querySubstParamCount = 4;
}
// QT5
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount; querySubstParam++) {
queryString = compileMustacheTemplate(5, queryIndex, querySubstParam);
tasks.add(new Task(5, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery5.1." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
if(enabledQueryTypes.getProperty("QT5").equals("1")) {
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount; querySubstParam++) {
queryString = compileMustacheTemplate(5, queryIndex, querySubstParam);
tasks.add(new Task(5, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery5.1." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
}
}
}
queryIndex++;

// QT6
// same querySubstParamCount as QT5
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount; querySubstParam++) {
queryString = compileMustacheTemplate(6, queryIndex, querySubstParam);
tasks.add(new Task(6, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery6.1." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
if(enabledQueryTypes.getProperty("QT6").equals("1")) {
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount; querySubstParam++) {
queryString = compileMustacheTemplate(6, queryIndex, querySubstParam);
tasks.add(new Task(6, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery6.1." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
}
}
}
queryIndex++;

// QT7
// can not be supported when we have 2 or less total versions, as there cannot exist cross-deltas
querySubstParamCount = 3;
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount && querySubstParam < numberOfVersions - 1; querySubstParam++) {
queryString = compileMustacheTemplate(7, queryIndex, querySubstParam);
tasks.add(new Task(7, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery7.1." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
}
}
queryIndex++;

// QT8
if (numberOfVersions == 2) {
querySubstParamCount = 1;
} else if(numberOfVersions == 3) {
if(enabledQueryTypes.getProperty("QT7").equals("1")) {
querySubstParamCount = 3;
} else if(numberOfVersions == 4) {
querySubstParamCount = 5;
} else {
querySubstParamCount = 6;
}
for (int querySubType = 1; querySubType <= Statistics.VERSIONING_SUB_QUERIES_COUNT; querySubType++) {
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount; querySubstParam++) {
queryString = compileMustacheTemplate(8, queryIndex, querySubstParam);
tasks.add(new Task(8, querySubType, Integer.toString(taskId++), queryString, null));
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount && querySubstParam < numberOfVersions - 1; querySubstParam++) {
queryString = compileMustacheTemplate(7, queryIndex, querySubstParam);
tasks.add(new Task(7, 1, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery8." + querySubType + "." + querySubstParam + ".sparql"), queryString);
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery7.1." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
}
}
queryIndex++;
}
queryIndex++;

// QT8
if(enabledQueryTypes.getProperty("QT8").equals("1")) {
if (numberOfVersions == 2) {
querySubstParamCount = 1;
} else if(numberOfVersions == 3) {
querySubstParamCount = 3;
} else if(numberOfVersions == 4) {
querySubstParamCount = 5;
} else {
querySubstParamCount = 6;
}
for (int querySubType = 1; querySubType <= Statistics.VERSIONING_SUB_QUERIES_COUNT; querySubType++) {
for(int querySubstParam = 1; querySubstParam <= querySubstParamCount; querySubstParam++) {
queryString = compileMustacheTemplate(8, queryIndex, querySubstParam);
tasks.add(new Task(8, querySubType, Integer.toString(taskId++), queryString, null));
try {
FileUtils.writeStringToFile(new File(queriesPath + "versioningQuery8." + querySubType + "." + querySubstParam + ".sparql"), queryString);
} catch (IOException e) {
LOGGER.error("Exception caught during saving of generated task : ", e);
}
}
queryIndex++;
}
} else {
queryIndex += Statistics.VERSIONING_SUB_QUERIES_COUNT;
}
}

Expand Down
4 changes: 3 additions & 1 deletion src/main/java/org/hobbit/benchmark/versioning/properties/VersioningConstants.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,9 @@ public final class VersioningConstants {
public static final String NUMBER_OF_VERSIONS = "number_of_versions";

public static final String SENT_DATA_FORM = "sent_data_form";


public static final String ENABLED_QUERY_TYPES = "enabled_query_types";

// =============== TASK GENERATOR CONSTANTS ===============

public static final String NUMBER_OF_TASK_GENERATORS= "task_generators_number";
Expand Down

0 comments on commit afae5d1

Please sign in to comment.