Skip to content

Commit

Permalink
Update to current version of this test-runner.
Browse files Browse the repository at this point in the history
  • Loading branch information
@vstange
vstange committed Jun 19, 2017
1 parent 3e67433 commit e040f6a
Show file tree
Hide file tree
Showing 6 changed files with 238 additions and 115 deletions.
3 changes: 3 additions & 0 deletions src/main/java/de/vstange/ApplicationStart.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,9 @@
import org.springframework.data.jpa.repository.config.EnableJpaRepositories;

/**
* Just simple start up class for the spring boot environment.
* Why spring boot? I just wanted a simple db/hibernate setup.
*
* @author Vincent Stange
*/
@SpringBootApplication
Expand Down
266 changes: 154 additions & 112 deletions src/main/java/de/vstange/controller/Runner.java
View file
Original file line number Diff line number Diff line change
@@ -1,6 +1,7 @@
package de.vstange.controller;

import com.formulasearchengine.mathmlsim.similarity.MathPlag;
import com.formulasearchengine.mathmltools.mml.CMMLInfo;
import de.vstange.entity.*;
import de.vstange.query.QueryCollect;
import org.apache.commons.io.IOUtils;
Expand All @@ -9,82 +10,184 @@
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.boot.CommandLineRunner;
import org.springframework.stereotype.Component;
import org.springframework.transaction.annotation.Transactional;

import java.io.IOException;
import java.sql.Timestamp;
import java.util.Date;
import java.util.List;
import java.util.Map;
import java.util.*;
import java.util.concurrent.atomic.AtomicInteger;
import java.util.concurrent.atomic.AtomicLong;

/**
* TODO
* Test class for reproducing
*
* @author Vincent Stange
*/
@Component
public class Runner implements CommandLineRunner {

@Autowired

private FormulaRepository formulaRepository;

@Autowired
private ResultRepository resultRepository;

private final Log logger = LogFactory.getLog(getClass());

private List<Exception> errorlist = new ArrayList<>();

@Override
@Transactional
public void run(String... strings) throws Exception {
foo();
// qvar_single_test();

// testRun();
// deleteResults();
// compareRun();

deleteResults();
compareRun();
}

public void compareRun() throws IOException {
logger.info("collecting and preparing queries");
List<Query> queries = getQueries();
queries.parallelStream().forEach(q -> {
try {
CMMLInfo queryDoc = new CMMLInfo(q.getMathml());
q.setCmmlInfo(queryDoc);
q.setQuery(queryDoc.getXQuery());
q.setAbstractQuery(new CMMLInfo(q.getMathml()).toStrictCmml().abstract2CDs().getXQuery());
q.setDataQuery(new CMMLInfo(q.getMathml()).toStrictCmml().abstract2DTs().getXQuery());
} catch (Exception e) {
e.printStackTrace();
}
});
logger.info("queries collected");

logger.info("collecting formulas");
List<Formula> formulas = formulaRepository.findAll();
// filtering
// formulas = formulas.stream()
// .filter(f -> f.getName().equals("S7.Ex1.m1.1"))
// .filter(f -> f.getSectionname().equals("math0506287_1_147"))
// .collect(Collectors.toList());
logger.info("formulas collected");

logger.info("begin streaming");
AtomicInteger counter = new AtomicInteger(0);
AtomicLong totalTime = new AtomicLong(0L);
formulas.forEach(f -> {
long time = System.currentTimeMillis();

compareFormula(f, queries);
// compute avg time and print out
long tookTime = System.currentTimeMillis() - time;
long avgTime = totalTime.accumulateAndGet(tookTime, (l, r) -> l + r) / counter.incrementAndGet();
System.out.println("=== time (ms): " + tookTime + " (avg: " + avgTime + ") " + counter.get() + " | " + f.getName());
f = null;
});
logger.info("finished (total: " + totalTime.get() + ")");

// print errors
errorlist.forEach(Throwable::printStackTrace);
}

private void compareFormula(Formula formula, List<Query> queries) {
// prep work
try {
formula.setCmmlInfo(new CMMLInfo(formula.getValue()));
formula.setAbstractCmmlInfo(new CMMLInfo(formula.getValue()).toStrictCmml().abstract2CDs());
formula.setDataCmmlInfo(new CMMLInfo(formula.getValue()).toStrictCmml().abstract2DTs());
} catch (Exception e) {
e.printStackTrace();
errorlist.add(e);
return;
}

List<Result> resultSet = Collections.synchronizedList(new ArrayList<>(queries.size()));
queries.parallelStream().forEach(query -> {
Map<String, Object> oldFactors = compareOriginalFactors(query, formula);
Double coverage = (Double) oldFactors.get("coverage");
Integer depth = (Integer) oldFactors.get("depth");
Boolean structureMatch = (Boolean) oldFactors.get("structure match");
Boolean dataMatch = (Boolean) oldFactors.get("data match");
Boolean isFormula = (Boolean) oldFactors.get("equation");

Result result = new Result();
result.setQuerynum(query.getQueryNum());
result.setQueryformulaid(query.getQueryFormulaId());
result.setPatternname(formula.getName());
result.setPageid(formula.getSectionname());
result.setCdMatch(structureMatch == null ? 0 : structureMatch ? 1 : 0);
result.setDataMatch(dataMatch == null ? 0 : dataMatch ? 1 : 0);
result.setQueryCoverage(coverage);
result.setMatchDepth(depth == null ? -1 : depth);
result.setIsFormulae(isFormula == null ? 0 : isFormula ? 1 : 0);

result.setTimestamp(new Timestamp(new Date().getTime()));

resultSet.add(result);
});
resultRepository.save(resultSet);

// clean up - free space
formula.setCmmlInfo(null);
formula.setAbstractCmmlInfo(null);
formula.setDataCmmlInfo(null);
resultSet.clear();
}

/**
* Old score - this is basically a print out for comparison and be deleted later on
*
* @param query Reference MathML string (must contain pMML and cMML)
* @param formula Comparison MathML string (must contain pMML and cMML)
* @return ap of all found factors
*/
private Map<String, Object> compareOriginalFactors(Query query, Formula formula) {
try {
CMMLInfo compDoc = formula.getCmmlInfo();

final Integer depth = compDoc.getDepth(query.getQuery());
final Double coverage = compDoc.getCoverage(query.getCmmlInfo().getElements(true));
Boolean equation = compDoc.isEquation(true);
Boolean match = formula.getAbstractCmmlInfo().isMatch(query.getAbstractQuery());
Boolean dataMatch = formula.getDataCmmlInfo().isMatch(query.getDataQuery());

HashMap<String, Object> result = new HashMap<>();
result.put("depth", depth);
result.put("coverage", coverage);
result.put("structure match", match);
result.put("data match", dataMatch);
result.put("equation", equation);
return result;
} catch (Exception e) {
System.err.println(formula.getName() + " : " + query.getQueryNum());
e.printStackTrace();
return null;
}
}

private Result createRecord(Result result) {
return resultRepository.save(result);
}

private void testRun() throws IOException {
logger.info("Formulas: " + formulaRepository.count());
logger.info("Results: " + resultRepository.count());
logger.info("queries: " + getQueries());
}

private List<Query> getQueries() throws IOException {
String s = IOUtils.toString(getClass().getClassLoader().getResourceAsStream("fQuery.xml"), "UTF-8");
return new QueryCollect().getQueries(s);
}

public void foo() {
// String math1 = "<m:math xmlns:m=\"http://www.w3.org/1998/Math/MathML\">\n" +
// " <m:semantics xml:id=\"m1.1a\" xref=\"m1.1.pmml\">\n" +
// " <m:mrow xml:id=\"m1.1.8.pmml\" xref=\"m1.1.8\">\n" +
// " <m:mrow xml:id=\"m1.1.8.1.pmml\" xref=\"m1.1.8.1\">\n" +
// " <mws:qvar xmlns:mws=\"http://search.mathweb.org/ns\" name=\"\uD835\uDC60\uD835\uDC5E\uD835\uDC62\uD835\uDC4E\uD835\uDC5F\uD835\uDC52\"/>\n" +
// " <m:mo xml:id=\"m1.1.8.1.1.pmml\" xref=\"m1.1.8.1.1\">\u2062</m:mo>\n" +
// " <m:mrow xml:id=\"m1.1.3.pmml\" xref=\"m1.1.3\">\n" +
// " <m:mo xml:id=\"m1.1.3a.pmml\" xref=\"m1.1.3\">(</m:mo>\n" +
// " <mws:qvar xmlns:mws=\"http://search.mathweb.org/ns\" name=\"\uD835\uDC5Dℎ\uD835\uDC56\"/>\n" +
// " <m:mo xml:id=\"m1.1.3c.pmml\" xref=\"m1.1.3\">)</m:mo>\n" +
// " </m:mrow>\n" +
// " </m:mrow>\n" +
// " <m:mo xml:id=\"m1.1.5.pmml\" xref=\"m1.1.5\">=</m:mo>\n" +
// " <m:mrow xml:id=\"m1.1.8.2.pmml\" xref=\"m1.1.8.2\">\n" +
// " <m:mi xml:id=\"m1.1.6.pmml\" xref=\"m1.1.6\">i</m:mi>\n" +
// " <m:mo xml:id=\"m1.1.8.2.1.pmml\" xref=\"m1.1.8.2.1\">\u2062</m:mo>\n" +
// " <m:mi xml:id=\"m1.1.7.pmml\" xref=\"m1.1.7\">d</m:mi>\n" +
// " </m:mrow>\n" +
// " </m:mrow>\n" +
// " <m:annotation-xml encoding=\"MathML-Content\" xml:id=\"m1.1.pmml\" xref=\"m1.1\">\n" +
// "\t\t\t<m:apply xml:id=\"m1.1.8\" xref=\"m1.1.8.pmml\">\n" +
// " <m:eq xml:id=\"m1.1.5\" xref=\"m1.1.5.pmml\"/>\n" +
// " <m:apply xml:id=\"m1.1.8.1\" xref=\"m1.1.8.1.pmml\">\n" +
// " <m:times xml:id=\"m1.1.8.1.1\" xref=\"m1.1.8.1.1.pmml\"/>\n" +
// " <m:ci xml:id=\"m1.1.6\" xref=\"m1.1.6.pmml\">a</m:ci>\n" +
// " <m:ci xml:id=\"m1.1.6\" xref=\"m1.1.6.pmml\">b</m:ci>\n" +
// " </m:apply>\n" +
// " <m:apply xml:id=\"m1.1.8.2\" xref=\"m1.1.8.2.pmml\">\n" +
// " <m:times xml:id=\"m1.1.8.2.1\" xref=\"m1.1.8.2.1.pmml\"/>\n" +
// " <m:ci xml:id=\"m1.1.6\" xref=\"m1.1.6.pmml\">i</m:ci>\n" +
// " <m:ci xml:id=\"m1.1.7\" xref=\"m1.1.7.pmml\">d</m:ci>\n" +
// " </m:apply>\n" +
// " </m:apply>\n" +
// " \n" +
// " </m:annotation-xml>\n" +
// " <m:annotation encoding=\"application/x-tex\" xml:id=\"m1.1b\" xref=\"m1.1.pmml\">\n" +
// " \\qvar{square}(\\qvar{phi})=id\n" +
// " </m:annotation>\n" +
// " </m:semantics>\n" +
// " </m:math>";
private void deleteResults() {
logger.info("delete results");
resultRepository.deleteAll();
logger.info("results deleted");
}


public void qvar_single_test() {
String math1 = "<m:math xmlns:m=\"http://www.w3.org/1998/Math/MathML\">\n" +
" <m:semantics xml:id=\"m1.1a\" xref=\"m1.1.pmml\">\n" +
" <m:apply xml:id=\"m1.1.8\" xref=\"m1.1.8.pmml\">\n" +
Expand Down Expand Up @@ -212,65 +315,4 @@ public void foo() {
Map<String, Object> stringObjectMap = MathPlag.compareOriginalFactors(math1, math2);
System.out.println(stringObjectMap);
}

public void compareRun() throws IOException {
String s = IOUtils.toString(getClass().getClassLoader().getResourceAsStream("fQuery.xml"), "UTF-8");
List<Query> queries = new QueryCollect().getQueries(s);
logger.info("collected queries");
logger.info("begin streaming");
formulaRepository.streamAll().filter(f -> f.getName().equals("Thmdefinition1.p1.1.1.m8.1")).forEach(f -> compareFormula(f, queries));
}

private void compareFormula(Formula formula, List<Query> queries) {
int i = 0;
System.out.println(i++ + ": " + formula.getName());
for (Query query : queries) {

Map<String, Object> oldFactors = MathPlag.compareOriginalFactors(query.getMathml(), formula.getValue());
Double coverage = (Double) oldFactors.get("coverage");
Integer depth = (Integer) oldFactors.get("depth");
Boolean structureMatch = (Boolean) oldFactors.get("structure match");
Boolean isFormula = (Boolean) oldFactors.get("formula");

Result result = new Result();
result.setQuerynum(query.getQueryNum());
result.setQueryformulaid(query.getQueryFormulaId());
result.setPatternname(formula.getName());
result.setPageid(formula.getSectionname());
//
result.setCdMatch(structureMatch == null ? 0 : structureMatch ? 1 : 0);
// result.setDataMatch();
result.setQueryCoverage(coverage);
result.setMatchDepth(depth);
result.setIsFormulae(isFormula == null ? 0 : isFormula ? 1 : 0);

result.setTimestamp(new Timestamp(new Date().getTime()));

Result record = createRecord(result);

System.out.println(record);

// System.out.println(i++ + ": " + formula.getName() + "_" + query.getQueryNum());
}
}

private Result createRecord(Result result) {
return resultRepository.save(result);
}

private void testRun() throws IOException {
logger.info("Formulas: " + formulaRepository.count());
logger.info("Results: " + resultRepository.count());

QueryCollect queryCollect = new QueryCollect();
String s = IOUtils.toString(getClass().getClassLoader().getResourceAsStream("fQuery.xml"), "UTF-8");
List<Query> queries = queryCollect.getQueries(s);

logger.info("queries: " + queries.size());
}

private void deleteResults() {
resultRepository.deleteAll();
}

}
37 changes: 36 additions & 1 deletion src/main/java/de/vstange/entity/Formula.java
View file
Original file line number Diff line number Diff line change
@@ -1,5 +1,8 @@
package de.vstange.entity;

import com.formulasearchengine.mathmltools.mml.CMMLInfo;
import net.sf.saxon.s9api.XQueryExecutable;

import javax.persistence.*;

/**
Expand All @@ -15,7 +18,6 @@
@IdClass(FormulaId.class)
public class Formula {


@Id
@Column(name = "formula_name", nullable = false)
private String name;
Expand All @@ -27,6 +29,15 @@ public class Formula {
@Column(name = "value")
private String value;

@Transient
private CMMLInfo cmmlInfo;

@Transient
private CMMLInfo abstractCmmlInfo;

@Transient
private CMMLInfo dataCmmlInfo;

public String getName() {
return name;
}
Expand All @@ -50,4 +61,28 @@ public String getValue() {
public void setValue(String value) {
this.value = value;
}

public CMMLInfo getCmmlInfo() {
return cmmlInfo;
}

public void setCmmlInfo(CMMLInfo cmmlInfo) {
this.cmmlInfo = cmmlInfo;
}

public CMMLInfo getAbstractCmmlInfo() {
return abstractCmmlInfo;
}

public void setAbstractCmmlInfo(CMMLInfo abstractCmmlInfo) {
this.abstractCmmlInfo = abstractCmmlInfo;
}

public void setDataCmmlInfo(CMMLInfo dataCmmlInfo) {
this.dataCmmlInfo = dataCmmlInfo;
}

public CMMLInfo getDataCmmlInfo() {
return dataCmmlInfo;
}
}
Loading

0 comments on commit e040f6a

Please sign in to comment.