Examples with AmazonPersonalize com.amazonaws.services.personalize.AmazonPersonalize used on opensource projects

Example 11 with AmazonPersonalize

use of com.amazonaws.services.personalize.AmazonPersonalize in project knime-cloud by knime.

the class AbstractAmazonPersonalizeDataUploadNodeModel method createSchema.

private String createSchema(final AmazonPersonalize personalizeClient, final DataTableSpec spec) {
    final StringBuilder schemaNameBuilder = new StringBuilder(getSchemaNamePrefix());
    FieldAssembler<Schema> fieldAssembler = createFieldAssembler(SCHEMA_NAMESPACE);
    for (final String colName : spec.getColumnNames()) {
        if (!colName.startsWith(PREFIX_METADATA_FIELD)) {
            continue;
        }
        final DataColumnSpec colSpec = spec.getColumnSpec(colName);
        final boolean isCategorical;
        final Type type;
        if (colSpec.getType().isCompatible(StringValue.class)) {
            isCategorical = true;
            type = Type.STRING;
        } else if (colSpec.getType().isCompatible(IntValue.class)) {
            isCategorical = false;
            type = Type.INT;
        } else if (colSpec.getType().isCompatible(LongValue.class)) {
            isCategorical = false;
            type = Type.LONG;
        } else {
            isCategorical = false;
            type = Type.DOUBLE;
        }
        schemaNameBuilder.append("-" + type);
        // 'categorical' must be set for metadata
        fieldAssembler = fieldAssembler.name(colName).prop("categorical", isCategorical).type(Schema.create(type)).noDefault();
    }
    final String schemaName = schemaNameBuilder.toString();
    // check if the same schema has been created before
    final List<DatasetSchemaSummary> existingSchemas = AmazonPersonalizeUtils.listAllSchemas(personalizeClient);
    final Optional<DatasetSchemaSummary> schemaSummary = existingSchemas.stream().filter(e -> e.getName().equals(schemaName)).findAny();
    // if so, use this one again
    if (schemaSummary.isPresent()) {
        return schemaSummary.get().getSchemaArn();
    }
    // otherwise create new one
    final Schema schema = fieldAssembler.endRecord();
    final CreateSchemaRequest createSchemaRequest = new CreateSchemaRequest().withName(schemaName).withSchema(schema.toString());
    return personalizeClient.createSchema(createSchemaRequest).getSchemaArn();
}

Example 12 with AmazonPersonalize

use of com.amazonaws.services.personalize.AmazonPersonalize in project knime-cloud by knime.

the class AbstractAmazonPersonalizeDataUploadNodeModel method checkAlreadyExistingDataset.

private void checkAlreadyExistingDataset(final AmazonPersonalize personalizeClient, final String datasetGroupArn, final ExecutionContext exec) throws InterruptedException {
    exec.setMessage("Checking already existing datasets");
    final ListDatasetsResult listDatasets = personalizeClient.listDatasets(new ListDatasetsRequest().withDatasetGroupArn(datasetGroupArn));
    final Optional<DatasetSummary> dataset = listDatasets.getDatasets().stream().filter(e -> e.getDatasetType().equals(m_datasetType)).findFirst();
    if (dataset.isPresent()) {
        if (m_settings.getOverwriteDatasetPolicy().equals(OverwritePolicy.ABORT.toString())) {
            // Abort if dataset already exists
            throw new IllegalStateException("A dataset of type '" + getDatasetType() + "' already exists. Either choose a different dataset group or select to overwrite the existing " + "dataset.");
        } else {
            // Delete the existing dataset
            exec.setMessage("Deleting existing dataset");
            deleteDataset(personalizeClient, datasetGroupArn, dataset.get().getDatasetArn());
        }
    }
    exec.setProgress(1);
}