gFeed/DataMinerAlgorithmsCrawler/src/main/java/org/gcube/data/publishing/gCatFeeder/collectors/dm/DMAlgorithmsInfoCollector.java

277 lines
8.9 KiB
Java

package org.gcube.data.publishing.gCatFeeder.collectors.dm;
import java.time.LocalDateTime;
import java.time.Month;
import java.time.ZoneId;
import java.time.chrono.IsoChronology;
import java.time.format.DateTimeFormatter;
import java.time.format.DateTimeFormatterBuilder;
import java.time.format.ResolverStyle;
import java.time.temporal.ChronoField;
import java.util.*;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import org.gcube.data.analysis.dataminermanagercl.server.DataMinerService;
import org.gcube.data.analysis.dataminermanagercl.server.dmservice.SClient;
import org.gcube.data.analysis.dataminermanagercl.shared.process.Operator;
import org.gcube.data.analysis.dataminermanagercl.shared.process.OperatorCategory;
import org.gcube.data.analysis.dataminermanagercl.shared.process.OperatorsClassification;
import org.gcube.data.publishing.gCatFeeder.collectors.dm.model.InternalAlgorithmDescriptor;
import org.gcube.data.publishing.gCatFeeder.collectors.dm.model.Parameter;
import org.gcube.data.publishing.gCatFeeder.collectors.dm.model.UserIdentity;
import org.gcube.data.publishing.gCatFeeder.model.EnvironmentConfiguration;
import org.gcube.data.publishing.gCatFeeder.utils.ISUtils;
import org.gcube.data.publishing.gCatfeeder.collectors.DataCollector;
import org.gcube.data.publishing.gCatfeeder.collectors.model.faults.CollectorFault;
import lombok.extern.slf4j.Slf4j;
import org.gcube.portlets.user.uriresolvermanager.exception.IllegalArgumentException;
import org.gcube.portlets.user.uriresolvermanager.exception.UriResolverMapException;
@Slf4j
public class DMAlgorithmsInfoCollector implements DataCollector<InternalAlgorithmDescriptor> {
/*
* i.e. "Basic statistic max min average {Published by Giancarlo Panichi (giancarlo.panichi) on 2018/07/20 10:24 GMT}"
*/
private static final Pattern descriptionPattern = Pattern.compile("\\{Published by (.*)\\((.*(\\.)?.*\\)) on (.*)\\}$");
static final DateTimeFormatter versionDateParser= DateTimeFormatter.ofPattern("uuuu/MM/dd HH:mm 'GMT'[Z]")
.withResolverStyle(ResolverStyle.STRICT)
.withLocale(Locale.getDefault())
.withZone(ZoneId.systemDefault());
/* static final DateTimeFormatter versionDateParser=new DateTimeFormatterBuilder()
.parseCaseInsensitive().parseLenient()
.appendValue(ChronoField.YEAR,4)
.appendLiteral('/')
.appendValue(ChronoField.MONTH_OF_YEAR,2)
.appendLiteral('/')
.appendValue(ChronoField.DAY_OF_MONTH,2)
.appendLiteral(' ')
.appendValue(ChronoField.HOUR_OF_DAY,2)
.appendLiteral(':')
.appendValue(ChronoField.MINUTE_OF_HOUR,2)
.appendLiteral(' ')
.appendOffset("+HHMM", "GMT")
.toFormatter().withChronology(IsoChronology.INSTANCE)
.withResolverStyle(ResolverStyle.SMART); */
private Map<String,String> env=null;
public void setEnvironmentConfiguration(EnvironmentConfiguration envConfig) {
if(envConfig!=null) {
log.debug("Current Environment Configuration is : "+envConfig.getCurrentConfiguration());
this.env=envConfig.getCurrentConfiguration();
}else {
env=Collections.emptyMap();
log.warn("NO ENVIRONMENT CONFIGURATION FOUND");
}
}
@Override
public Set<InternalAlgorithmDescriptor> collect() throws CollectorFault {
try {
log.trace("Collecting information for Dataminer algorithms");
HashSet<InternalAlgorithmDescriptor> toReturn=new HashSet();
SClient client=new DataMinerService().getClient();
String wpsbaseUrl=getWPSBasePath();
List<OperatorsClassification> opClassifications=client.getOperatorsClassifications();
log.debug("Found {} classifications.",opClassifications.size());
for(OperatorsClassification opClass: opClassifications) {
// Load info
String opClassName =opClass.getName();
List<Operator> ops=opClass.getOperators();
log.debug("Found {} operators under classification {} ",ops.size(),opClassName);
for(Operator op : ops) {
InternalAlgorithmDescriptor desc=new InternalAlgorithmDescriptor();
desc.setClassName(opClassName);
// OperatorCategory Info
OperatorCategory cat = op.getCategory();
String categoryBriefDescription = cat.getBriefDescription();
String categoryDescription= cat.getDescription();
String categoryID=cat.getId();
String categoryName=cat.getName();
desc.setCategoryBriefDescription(categoryBriefDescription);
desc.setCategoryID(categoryID);
desc.setCategoryName(categoryName);
desc.setCategoryDescription(categoryDescription);
// Operator info
String opBriefDescription=op.getBriefDescription();
String opDescription=op.getDescription();
String opID=op.getId();
String operatorName=op.getName();
desc.setBriefDescription(opBriefDescription);
desc.setDescription(opDescription);
desc.setId(opID);
desc.setName(operatorName);
desc.setAuthor(parseUser(getAuthor(opDescription)));
desc.setMaintainer(parseUser(getAuthor(opDescription)));
desc.setVersion(parseDescriptionForDate(opDescription));
// Try to use version as creation time
LocalDateTime toSetCreationDate=null;
try{
toSetCreationDate=LocalDateTime.parse(desc.getVersion(),versionDateParser);
}catch(Throwable t){
log.debug("Version {} is not a date. Using 1900-01-01 as Creation date..",desc.getVersion());
toSetCreationDate = LocalDateTime.of(1900, Month.JANUARY,1,0,0);
}
desc.setCreationDate(toSetCreationDate);
// Parameters info
for(org.gcube.data.analysis.dataminermanagercl.shared.parameters.Parameter param:client.getInputParameters(op)) {
String paramDescription=param.getDescription();
String paramName=param.getName();
String paramType=param.getTypology().toString();
String paramValue=param.getValue();
desc.getInputParameters().add(
new Parameter(paramName, paramType, paramDescription, paramValue));
}
for(org.gcube.data.analysis.dataminermanagercl.shared.parameters.Parameter param:client.getOutputParameters(op)) {
String paramDescription=param.getDescription();
String paramName=param.getName();
String paramType=param.getTypology().toString();
String paramValue=param.getValue();
desc.getOutputParameters().add(
new Parameter(paramName, paramType, paramDescription, paramValue));
}
desc.getTags().add(categoryName);
String guiBasePath=getGUIBasePath();
if(guiBasePath!=null)
desc.setGuiLink(getGUIBasePath()+"?"+
DataMinerCollectorProperties.getProperty(DataMinerCollectorProperties.GUI_PARAM_NAME)+"="+opID);
desc.setGatewayName(env.get(Constants.GATEWAY_NAME));
if(wpsbaseUrl!=null) {
desc.setWpsLink(wpsbaseUrl+"?Request=DescribeProcess&Version=1.0.0&Service=WPS"+"&Identifier="+opID);
}
desc.setPrivateFlag(Boolean.parseBoolean(env.get(Constants.PRIVATE)));
toReturn.add(desc);
}
// Create bean
}
return toReturn;
}catch(Exception e) {
throw new CollectorFault("Unable to retrieve information",e);
}
}
private String getGUIBasePath() {
return env.get(Constants.GUI_BASE_URL);
}
private String getAuthor(String algorithmDescription) {
String toReturn=parseDescriptionForUser(algorithmDescription);
if(toReturn==null)
toReturn=env.get(Constants.DEFAULT_AUTHOR);
if(toReturn==null)
toReturn=DataMinerCollectorProperties.getProperty(DataMinerCollectorProperties.DEFAULT_AUTHOR);
return toReturn;
}
static final UserIdentity parseUser(String userString) {
try{
String splitter=null;
if(userString.contains(" "))
splitter=" ";
else if (userString.contains(".")) splitter="\\.";
String[] splitted=userString.split(splitter);
return new UserIdentity(splitted[0], splitted[1], null, null);
}catch(NullPointerException e){
System.err.println("Error with userString "+userString);
e.printStackTrace(System.err);
throw e;
}
}
/*
* e.g. "Basic statistic max min average {Published by Giancarlo Panichi (giancarlo.panichi) on 2018/07/20 10:24 GMT}"
*/
static final String parseDescriptionForUser(String description) {
Matcher m=descriptionPattern.matcher(description);
if(m.find())
return m.group(1); // group 0 == {...}, group 1 == Giancarlo Panichi, group 2 == giancarlo.panichi
else return null;
}
static final String parseDescriptionForDate(String description) {
Matcher m=descriptionPattern.matcher(description);
if(m.find())
return m.group(4); // group 0 == {...}, group 1 == Giancarlo Panichi, group 2 == giancarlo.panichi, group 4 == 2018/07/20 10:24 GMT
else return "n.d.";
}
static final String getWPSBasePath() {
try{
return ISUtils.queryForServiceEndpointsByName("DataAnalysis", "DataMiner").get(0).profile().
accessPoints().iterator().next().address();
}catch(Throwable t) {
log.warn("Unable to find DM proxy. No WPS URL will be provided",t);
return null;
}
}
}