Aprar
/
bonitoo
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Array like API

Browse Source
master
EHP 6 years ago
parent e8477fad03
commit 10ca1d2c51
2 changed files with 88 additions and 82 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 102
      inference/src/main/java/cz/aprar/bonitoo/inference/CacheInference.java
  2. 68
      inference/src/test/java/cz/aprar/bonitoo/inference/CacheInferenceTest.java

102
inference/src/main/java/cz/aprar/bonitoo/inference/CacheInference.java
Unescape View File

@ -18,6 +18,7 @@ import java.util.Map;
public class CacheInference {
private static final TTL[] TTL_VALUES = TTL.values();
private static final int NUM_FEATURES = 20;
private final ObjectMapper mapper = new ObjectMapper();
private final Booster booster;
@ -37,59 +38,82 @@ public class CacheInference {
labels = loadLabels(labelIS);
}
public TTL cacheTTL(final FlightData data) throws XGBoostError {
public TTL[] cacheTTL(final FlightData[] data) throws XGBoostError {
return cacheTTL(data, ZonedDateTime.now());
}
/**
* Method for backtesting
*/
public TTL cacheTTL(final FlightData data, final ZonedDateTime now) throws XGBoostError {
final float[] predicts = cacheTTLProbability(data, now);
return TTL_VALUES[argmax(predicts)];
public TTL[] cacheTTL(final FlightData[] data, final ZonedDateTime now) throws XGBoostError {
final float[][] predicts = cacheTTLProbability(data, now);
if (predicts.length == 0) {
return new TTL[] {};
}
final TTL[] ttl = new TTL[predicts.length];
for (int i = 0; i < predicts.length; i++) {
int idx = 0;
float max = Float.MIN_VALUE;
for (int j = 0; j < predicts[i].length; j++) {
if (predicts[i][j] > max) {
max = predicts[i][j];
idx = j;
}
}
ttl[i] = TTL_VALUES[idx];
}
return ttl;
}
public float[] cacheTTLProbability(final FlightData data) throws XGBoostError {
public float[][] cacheTTLProbability(final FlightData[] data) throws XGBoostError {
return cacheTTLProbability(data, ZonedDateTime.now());
}
/**
* Method for backtesting
*/
public float[] cacheTTLProbability(final FlightData data, final ZonedDateTime now) throws XGBoostError {
public float[][] cacheTTLProbability(final FlightData[] data, final ZonedDateTime now) throws XGBoostError {
final DMatrix matrix = createMatrix(data, now);
float[][] predicts = booster.predict(matrix);
return predicts[0];
return booster.predict(matrix);
}
private Booster loadModel(InputStream model) throws XGBoostError, IOException {
return XGBoost.loadModel(model);
}
private DMatrix createMatrix(final FlightData data, final ZonedDateTime now) throws XGBoostError {
final float[] arr = new float[20];
arr[0] = (float) data.getInputPrice();
arr[1] = labels.get("type").getOrDefault(data.getType(), 0);
arr[2] = labels.get("flight.inboundSegments.departure").getOrDefault(data.getInboundDeparture(), 0);
arr[3] = labels.get("flight.inboundSegments.arrival").getOrDefault(data.getInboundArrival(), 0);
arr[4] = labels.get("flight.inboundSegments.origin.airportCode").getOrDefault(data.getInboundOrigin(), 0);
arr[5] = labels.get("flight.inboundSegments.destination.airportCode").getOrDefault(data.getInboundDestination(), 0);
arr[6] = labels.get("flight.inboundSegments.airline.code").getOrDefault(data.getInboundAirlines(), 0);
arr[7] = labels.get("flight.inboundMCX.code").getOrDefault(data.getInboundMCXAirlines(), 0);
arr[8] = labels.get("flight.outboundSegments.departure").getOrDefault(data.getOutboundDeparture(), 0);
arr[9] = labels.get("flight.outboundSegments.arrival").getOrDefault(data.getOutboundArrival(), 0);
arr[10] = labels.get("flight.outboundSegments.origin.airportCode").getOrDefault(data.getOutboundOrigin(), 0);
arr[11] = labels.get("flight.outboundSegments.destination.airportCode").getOrDefault(data.getOutboundDestination(), 0);
arr[12] = labels.get("flight.outboundSegments.airline.code").getOrDefault(data.getOutboundAirlines(), 0);
arr[13] = labels.get("flight.outboundMCX.code").getOrDefault(data.getOutboundMCXAirlines(), 0);
arr[14] = data.getNights();
arr[15] = computePrebooking(data.getOutDepartureDate(), now);
arr[16] = now.getDayOfMonth();
arr[17] = now.getDayOfWeek().getValue() - 1;
arr[18] = data.getOutDepartureDate().getDayOfMonth();
arr[19] = data.getOutDepartureDate().getDayOfWeek().getValue() - 1;
return new DMatrix(arr, 1, arr.length);
private DMatrix createMatrix(final FlightData[] data, final ZonedDateTime now) throws XGBoostError {
final float[] arr = new float[data.length * NUM_FEATURES];
for (int i = 0; i < data.length; i++) {
final FlightData current = data[i];
final int start = NUM_FEATURES * i;
arr[start] = (float) current.getInputPrice();
arr[start + 1] = labels.get("type").getOrDefault(current.getType(), 0);
arr[start + 2] = labels.get("flight.inboundSegments.departure").getOrDefault(current.getInboundDeparture(), 0);
arr[start + 3] = labels.get("flight.inboundSegments.arrival").getOrDefault(current.getInboundArrival(), 0);
arr[start + 4] = labels.get("flight.inboundSegments.origin.airportCode").getOrDefault(current.getInboundOrigin(), 0);
arr[start + 5] = labels.get("flight.inboundSegments.destination.airportCode").getOrDefault(current.getInboundDestination(), 0);
arr[start + 6] = labels.get("flight.inboundSegments.airline.code").getOrDefault(current.getInboundAirlines(), 0);
arr[start + 7] = labels.get("flight.inboundMCX.code").getOrDefault(current.getInboundMCXAirlines(), 0);
arr[start + 8] = labels.get("flight.outboundSegments.departure").getOrDefault(current.getOutboundDeparture(), 0);
arr[start + 9] = labels.get("flight.outboundSegments.arrival").getOrDefault(current.getOutboundArrival(), 0);
arr[start + 10] = labels.get("flight.outboundSegments.origin.airportCode").getOrDefault(current.getOutboundOrigin(), 0);
arr[start + 11] = labels.get("flight.outboundSegments.destination.airportCode").getOrDefault(current.getOutboundDestination(), 0);
arr[start + 12] = labels.get("flight.outboundSegments.airline.code").getOrDefault(current.getOutboundAirlines(), 0);
arr[start + 13] = labels.get("flight.outboundMCX.code").getOrDefault(current.getOutboundMCXAirlines(), 0);
arr[start + 14] = current.getNights();
arr[start + 15] = computePrebooking(current.getOutDepartureDate(), now);
arr[start + 16] = now.getDayOfMonth();
arr[start + 17] = now.getDayOfWeek().getValue() - 1;
arr[start + 18] = current.getOutDepartureDate().getDayOfMonth();
arr[start + 19] = current.getOutDepartureDate().getDayOfWeek().getValue() - 1;
}
return new DMatrix(arr, data.length, NUM_FEATURES);
}
private float computePrebooking(final ZonedDateTime outdeparture, final ZonedDateTime now) {
@ -101,18 +125,4 @@ public class CacheInference {
};
return mapper.readValue(labels, typeRef);
}
private int argmax(final float[] data) {
int idx = 0;
float max = Float.MIN_VALUE;
for (int i = 0; i < data.length; i++) {
if (data[i] > max) {
max = data[i];
idx = i;
}
}
return idx;
}
}

68
inference/src/test/java/cz/aprar/bonitoo/inference/CacheInferenceTest.java
Unescape View File

@ -29,34 +29,27 @@ public class CacheInferenceTest {
classLoader.getResourceAsStream("model/encoder.json"));
}
@Test(dataProvider = "inferenceData")
void testInference(final FlightData input, final TTL expected) throws XGBoostError {
final TTL result = ci.cacheTTL(input, TEST_DATE);
assertEquals(result, expected);
}
@DataProvider(name = "inferenceData")
public Object[][] inferenceData() {
return new Object[][]{
{new FlightData(
"WS",
toList("2020-05-09T23:59:00", "2020-05-10T10:30:00"),
toList("2020-05-10T08:55:00", "2020-05-10T11:30:00"),
toList("MCO", "FRA"),
toList("FRA", "PRG"),
toList("LH", "LH"),
"LH",
toList("2020-05-01T09:50:00", "2020-05-01T11:55:00"),
toList("2020-05-01T11:00:00", "2020-05-01T21:55:00"),
toList("PRG", "FRA"),
toList("FRA", "MCO"),
toList("LH", "LH"),
"LH",
39766.0,
7,
ZonedDateTime.of(2020, 5, 1, 9, 50, 0, 0, UTC)
), TTL.D3},
{new FlightData(
@Test
void testInference() throws XGBoostError {
final FlightData[] input = {new FlightData(
"WS",
toList("2020-05-09T23:59:00", "2020-05-10T10:30:00"),
toList("2020-05-10T08:55:00", "2020-05-10T11:30:00"),
toList("MCO", "FRA"),
toList("FRA", "PRG"),
toList("LH", "LH"),
"LH",
toList("2020-05-01T09:50:00", "2020-05-01T11:55:00"),
toList("2020-05-01T11:00:00", "2020-05-01T21:55:00"),
toList("PRG", "FRA"),
toList("FRA", "MCO"),
toList("LH", "LH"),
"LH",
39766.0,
7,
ZonedDateTime.of(2020, 5, 1, 9, 50, 0, 0, UTC)
),
new FlightData(
"PYTON",
"",
"",
@ -73,8 +66,8 @@ public class CacheInferenceTest {
336.258,
0,
ZonedDateTime.of(2019, 12, 18, 5, 45, 0, 0, UTC)
), TTL.D14},
{new FlightData(
),
new FlightData(
"AVIA",
toList("2020-02-07T02:25:00", "2020-02-07T14:50:00"),
toList("2020-02-07T13:10:00", "2020-02-07T16:55:00"),
@ -91,8 +84,8 @@ public class CacheInferenceTest {
5971.77978,
4,
ZonedDateTime.of(2020, 1, 28, 10, 35, 0, 0, UTC)
), TTL.D2},
{new FlightData(
),
new FlightData(
"HH",
toList("2019-11-01T16:30:00", "2019-11-01T23:35:00"),
toList("2019-11-01T21:12:00", "2019-11-02T07:45:00"),
@ -109,8 +102,8 @@ public class CacheInferenceTest {
17723.0,
12,
ZonedDateTime.of(2019, 10, 18, 8, 10, 0, 0, UTC)
), TTL.D2},
{new FlightData(
),
new FlightData(
"unknown",
toList(ZonedDateTime.now().format(DateTimeFormatter.ISO_LOCAL_DATE_TIME)),
toList(ZonedDateTime.now().format(DateTimeFormatter.ISO_LOCAL_DATE_TIME)),
@ -127,8 +120,11 @@ public class CacheInferenceTest {
0.0,
0,
ZonedDateTime.now()
), TTL.D1}
};
)};
final TTL[] expected = {TTL.D3, TTL.D14, TTL.D2, TTL.D2, TTL.D1};
final TTL[] result = ci.cacheTTL(input, TEST_DATE);
assertEquals(result, expected);
}
private String toList(final String... data) {

Write Preview
Loading…
Cancel
Save