[ALS-5755] Switch time series processor to ISO timestamps

- Make service that does this
- Isolate time series logic a bit more
- Tests

processing/src/main/java/edu/harvard/hms/dbmi/avillach/hpds/processing/AbstractProcessor.java

@@ -672,7 +672,7 @@ protected PhenoCube getCube(String path) {
 	 * Useful for federated pic-sure's where there are fewer
 	 * guarantees about concept paths.
 	 */
-	protected Optional<PhenoCube<?>> nullableGetCube(String path) {
+	public Optional<PhenoCube<?>> nullableGetCube(String path) {
 		try {
 			return Optional.ofNullable(store.get(path));
 		} catch (InvalidCacheLoadException | ExecutionException e) {

processing/src/main/java/edu/harvard/hms/dbmi/avillach/hpds/processing/timeseries/TimeSeriesConversionService.java

@@ -0,0 +1,15 @@
+package edu.harvard.hms.dbmi.avillach.hpds.processing.timeseries;
+
+import org.springframework.stereotype.Service;
+
+import java.time.Instant;
+import java.time.format.DateTimeFormatter;
+
+@Service
+public class TimeSeriesConversionService {
+
+    public String toISOString(Long unixTimeStamp) {
+        Instant instant = Instant.ofEpochMilli(unixTimeStamp);
+        return DateTimeFormatter.ISO_INSTANT.format(instant);
+    }
+}

processing/src/main/java/edu/harvard/hms/dbmi/avillach/hpds/processing/TimeseriesProcessor.java → processing/src/main/java/edu/harvard/hms/dbmi/avillach/hpds/processing/timeseries/TimeseriesProcessor.java

@@ -1,9 +1,13 @@
-package edu.harvard.hms.dbmi.avillach.hpds.processing;
+package edu.harvard.hms.dbmi.avillach.hpds.processing.timeseries;
 
 import java.io.FileNotFoundException;
 import java.io.IOException;
 import java.util.*;
 
+import edu.harvard.hms.dbmi.avillach.hpds.processing.AbstractProcessor;
+import edu.harvard.hms.dbmi.avillach.hpds.processing.AsyncResult;
+import edu.harvard.hms.dbmi.avillach.hpds.processing.HpdsProcessor;
+import edu.harvard.hms.dbmi.avillach.hpds.processing.QueryProcessor;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
@@ -36,14 +40,16 @@ public class TimeseriesProcessor implements HpdsProcessor {
 	private Logger log = LoggerFactory.getLogger(QueryProcessor.class);
 
 	private AbstractProcessor abstractProcessor;
+	private final TimeSeriesConversionService conversionService;
 
 	private final String ID_CUBE_NAME;
 	private final int ID_BATCH_SIZE;
 	private final int CACHE_SIZE;
 
 	@Autowired
-	public TimeseriesProcessor(AbstractProcessor abstractProcessor) {
+	public TimeseriesProcessor(AbstractProcessor abstractProcessor, TimeSeriesConversionService conversionService) {
 		this.abstractProcessor = abstractProcessor;
+		this.conversionService = conversionService;
 		// todo: handle these via spring annotations
 		CACHE_SIZE = Integer.parseInt(System.getProperty("CACHE_SIZE", "100"));
 		ID_BATCH_SIZE = Integer.parseInt(System.getProperty("ID_BATCH_SIZE", "0"));
@@ -115,14 +121,24 @@ private void addDataForConcepts(Collection<String> pathList, Set<String> exporte
 				if (cube.isStringType()) {
 					KeyAndValue<String> keyAndValue = (KeyAndValue) kvObj;
 					// "PATIENT_NUM","CONCEPT_PATH","NVAL_NUM","TVAL_CHAR","TIMESTAMP"
-					String[] entryData = { keyAndValue.getKey().toString(), conceptPath, "", keyAndValue.getValue(),
-							keyAndValue.getTimestamp().toString() };
+					String[] entryData = {
+						keyAndValue.getKey().toString(),
+						conceptPath,
+						"",
+						keyAndValue.getValue(),
+						conversionService.toISOString(keyAndValue.getTimestamp())
+					};
 					dataEntries.add(entryData);
 				} else { // numeric
 					KeyAndValue<Double> keyAndValue = (KeyAndValue) kvObj;
 					// "PATIENT_NUM","CONCEPT_PATH","NVAL_NUM","TVAL_CHAR","TIMESTAMP"
-					String[] entryData = { keyAndValue.getKey().toString(), conceptPath,
-							keyAndValue.getValue().toString(), "", keyAndValue.getTimestamp().toString() };
+					String[] entryData = {
+						keyAndValue.getKey().toString(),
+						conceptPath,
+						keyAndValue.getValue().toString(),
+						"",
+						conversionService.toISOString(keyAndValue.getTimestamp())
+					};
 					dataEntries.add(entryData);
 				}
 				//batch exports so we don't take double memory (valuesForKeys + dataEntries could be a lot of data points)

processing/src/test/java/edu/harvard/hms/dbmi/avillach/hpds/processing/timeseries/TimeSeriesConversionServiceTest.java

@@ -0,0 +1,21 @@
+package edu.harvard.hms.dbmi.avillach.hpds.processing.timeseries;
+
+import org.junit.Test;
+
+import java.util.TimeZone;
+
+import static org.junit.Assert.assertEquals;
+
+
+public class TimeSeriesConversionServiceTest {
+
+    TimeSeriesConversionService subject = new TimeSeriesConversionService();
+
+    @Test
+    public void shouldConvertToIsoString() {
+        String actual = subject.toISOString(0L);
+        String expected = "1970-01-01T00:00:00Z";
+
+        assertEquals(expected, actual);
+    }
+}

service/src/main/java/edu/harvard/hms/dbmi/avillach/hpds/service/QueryService.java

@@ -7,6 +7,7 @@
 import java.util.function.Predicate;
 import java.util.stream.Collectors;
 
+import edu.harvard.hms.dbmi.avillach.hpds.processing.timeseries.TimeseriesProcessor;
 import edu.harvard.hms.dbmi.avillach.hpds.service.util.QueryDecorator;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;