private void testPredictDiabetes(boolean skipDecoding)
throws MLHttpClientException, JSONException {
String payload = "[[1,89,66,23,94,28.1,0.167,21],[2,197,70,45,543,30.5,0.158,53]]";
String url =
skipDecoding
? "/api/models/" + modelId + "/predict?skipDecoding=true"
: "/api/models/" + modelId + "/predict";
response = mlHttpclient.doHttpPost(url, payload);
assertEquals(
"Unexpected response received",
Response.Status.OK.getStatusCode(),
response.getStatusLine().getStatusCode());
String reply = mlHttpclient.getResponseAsString(response);
JSONArray predictions = new JSONArray(reply);
assertEquals(
"Expected 2 predictions but received only " + predictions.length(),
2,
predictions.length());
if (skipDecoding) {
assertEquals(
"Expected a double value but found " + predictions.get(0),
true,
predictions.get(0) instanceof Double);
assertEquals(
"Expected a double value but found " + predictions.get(1),
true,
predictions.get(1) instanceof Double);
}
}
/**
* A test case for building a model with the given learning algorithm
*
* @param algorithmName Name of the learning algorithm
* @param algorithmType Type of the learning algorithm
* @throws MLHttpClientException
* @throws IOException
* @throws JSONException
* @throws InterruptedException
*/
private void buildModelWithLearningAlgorithm(String algorithmName, String algorithmType)
throws MLHttpClientException, IOException, JSONException, InterruptedException {
modelName =
MLTestUtils.createModelWithConfigurations(
algorithmName,
algorithmType,
MLIntegrationTestConstants.RESPONSE_ATTRIBUTE_DIABETES,
MLIntegrationTestConstants.TRAIN_DATA_FRACTION,
projectId,
versionSetId,
mlHttpclient);
modelId = mlHttpclient.getModelId(modelName);
response = mlHttpclient.doHttpPost("/api/models/" + modelId);
assertEquals(
"Unexpected response received",
Response.Status.OK.getStatusCode(),
response.getStatusLine().getStatusCode());
response.close();
// Waiting for model building to end
boolean status =
MLTestUtils.checkModelStatusCompleted(
modelName, mlHttpclient, MLIntegrationTestConstants.THREAD_SLEEP_TIME_LARGE, 1000);
// Checks whether model building completed successfully
assertEquals("Model building did not complete successfully", true, status);
}
/**
* A test case for predicting for a given set of data points from a file.
*
* @throws MLHttpClientException
* @throws JSONException
*/
private void testPredictDiabetesFromFile() throws MLHttpClientException, JSONException {
response =
mlHttpclient.predictFromCSV(modelId, MLIntegrationTestConstants.DIABETES_DATASET_TEST);
assertEquals(
"Unexpected response received",
Response.Status.OK.getStatusCode(),
response.getStatusLine().getStatusCode());
String reply = mlHttpclient.getResponseAsString(response);
JSONArray predictions = new JSONArray(reply);
assertEquals(7, predictions.length());
}
/**
* A test case for predicting for a given set of data points
*
* @throws MLHttpClientException
* @throws JSONException
*/
private void testPredictGammaTelescope() throws MLHttpClientException, JSONException {
String payload =
"[[18.8562,16.46,2.4385,0.5282,0.2933,25.1269,-6.5401,-16.9327,11.461,162.848],"
+ "[191.8036,49.7183,3.0006,0.2093,0.1225,146.2148,143.6098,31.6216,44.3492,245.4199]]";
response = mlHttpclient.doHttpPost("/api/models/" + modelId + "/predict", payload);
assertEquals(
"Unexpected response received",
Response.Status.OK.getStatusCode(),
response.getStatusLine().getStatusCode());
String reply = mlHttpclient.getResponseAsString(response);
JSONArray predictions = new JSONArray(reply);
assertEquals(2, predictions.length());
}
/**
* A test case for predicting with a dataset incompatible with the trained dataset in terms of
* number of features
*
* @throws MLHttpClientException
* @throws JSONException
*/
private void testPredictDiabetesInvalidNumberOfFeatures()
throws MLHttpClientException, JSONException {
String payload = "[[1,89,66,23,94,28.1,0.167],[2,197,70,45,543,30.5,0.158]]";
response = mlHttpclient.doHttpPost("/api/models/" + modelId + "/predict", payload);
assertEquals(
"Unexpected response received",
Response.Status.INTERNAL_SERVER_ERROR.getStatusCode(),
response.getStatusLine().getStatusCode());
}