com.apama.pa.pmml
Event ServiceParams


Sevice Params Interface

Pass the Service configuration parameters to the service handler factory
Member Summary
 action<string >setConfigDirectory

Set the path of the Config Directory
 action<string >setWorkDirectory

Set the path of the Work Directory
 action<string >setPMMLFileDirectory

Set the path to directory containing the PMML file
 action<string >setPMMLFileName

PMML File to be loaded by the scoring engine
 action<boolean >recordLatency

Add latency instrumentation to events being consumed/sent by the plugin
 action<boolean >logLatency

Log the total time being spent by the plugin to apply a model for an Input event
 action<string >setModelName

Set the name of the default model to be used
 action<integer >maxBatchSize

Maximum number of input requests that can be batched together
 action<string >subscribeToChannel

Plugin can optionally subscribe to a correlator channel and listen for Input requests
 action<string >sendToChannel

Plugin can optionally send responses directly to a channel
 action<string >addConfiguration

Add additional Java Spring Bean config initialization files
 action<string >addResource

Add an external resource like a custom function or a lookup table
 action<sequence<string > >addResources

Add multiple resources to the Predictive Analytics Plugin
 action<integer >setAssetStorePollingInterval

Plugin checks the Asset store for any changes to PMML models or resources at regular intervals
 action<boolean >cleanAssetStoreOnShutdown

Cleans the predictive analytics plugin asset store during shut down.
 action<integer >setThreadCount

Set the number of threads to use by the Plugin
 action<integer >setQueueSize

Set the Queue Size to be used for holding Input Requests Queue's the input requests if the Predictive Analytics Engine is busy scoring requests. This can happen for complex models. If the Input queue size is reached, further requests with be rejected with 'ADAPA_Error' being populated in the Output response payload
 action<integer >setThrottleThreshold

Set the threshold when throttling of scoring requests should be started. This parameter determines the number of scoring requests that will be active in the predictive analytics engine at any given point of time. This is done to reduce the overall heap usage by the application and preventing the process going out of memory
 action<string, string >addParam

Add a miscellaneous configuration parameter
 action<string, string > returns stringgetStringParam

Get the value of a configuration parameter which is already set
 action<string, boolean > returns booleangetBooleanParam

Get the value of a configuration parameter which is already set
 action<string, integer > returns integergetIntegerParam

Get the value of a configuration parameter which is already set
 action<string, float > returns floatgetFloatParam

Get the value of a configuration parameter which is already set
 action<string, decimal > returns decimalgetDecimalParam

Get the value of a configuration parameter which is already set
 action< > returns dictionary<string, string >getServiceParams

Get a copy of all the service parameters currently configured
 
Member Detail

addConfiguration

action<string > addConfiguration
Add additional Java Spring Bean config initialization files

The configuration file can override the default properties used by the Predictive Analytics Plugin like the Asset Store. It can also provide details about custom resources that can be loaded

See Predictive Analytics Samples for more details

addParam

action<string, string > addParam
Add a miscellaneous configuration parameter

addResource

action<string > addResource
Add an external resource like a custom function or a lookup table

These resources will be loaded before initialization of any PMML models

A resource file can be a pre-built 'jar' package or a csv file or an excel file

Please refer to Predictive Analytics Engine Javadoc for more details on loading custom resources

addResources

action<sequence<string > > addResources
Add multiple resources to the Predictive Analytics Plugin

cleanAssetStoreOnShutdown

action<boolean > cleanAssetStoreOnShutdown
Cleans the predictive analytics plugin asset store during shut down.

Defaults to true

getBooleanParam

action<string, boolean > returns boolean getBooleanParam
Get the value of a configuration parameter which is already set

getBooleanParam(string paramName, boolean defaultValue) returns boolean

getDecimalParam

action<string, decimal > returns decimal getDecimalParam
Get the value of a configuration parameter which is already set

getDecimalParam(string paramName, decimal defaultValue) returns decimal

getFloatParam

action<string, float > returns float getFloatParam
Get the value of a configuration parameter which is already set

getFloatParam(string paramName, float defaultValue) returns float

getIntegerParam

action<string, integer > returns integer getIntegerParam
Get the value of a configuration parameter which is already set

getIntegerParam(string paramName, integer defaultValue) returns integer

getServiceParams

action< > returns dictionary<string, string > getServiceParams
Get a copy of all the service parameters currently configured

getServiceParams() returns dictionary

getStringParam

action<string, string > returns string getStringParam
Get the value of a configuration parameter which is already set

getStringParam(string paramName, string defaultValue) returns string

logLatency

action<boolean > logLatency
Log the total time being spent by the plugin to apply a model for an Input event

maxBatchSize

action<integer > maxBatchSize
Maximum number of input requests that can be batched together

The plugin will still consume the batch even when the number of input requests is less than the maxBatchSize

Defaults to 1000

recordLatency

action<boolean > recordLatency
Add latency instrumentation to events being consumed/sent by the plugin

sendToChannel

action<string > sendToChannel
Plugin can optionally send responses directly to a channel

Users can configure the channel on which they would like to receive the responses and subscribe using monitor.subscribe()

setAssetStorePollingInterval

action<integer > setAssetStorePollingInterval
Plugin checks the Asset store for any changes to PMML models or resources at regular intervals

Default value is 0 i.e. plugin will not check for updates.

setConfigDirectory

action<string > setConfigDirectory
Set the path of the Config Directory

Should point to the config folder of the plugin installation directory.

This is assumed to be APAMA_HOME/adapters/config by default

setModelName

action<string > setModelName
Set the name of the default model to be used

Providing a default model name is optional

setPMMLFileDirectory

action<string > setPMMLFileDirectory
Set the path to directory containing the PMML file

setPMMLFileName

action<string > setPMMLFileName
PMML File to be loaded by the scoring engine

setQueueSize

action<integer > setQueueSize
Set the Queue Size to be used for holding Input Requests Queue's the input requests if the Predictive Analytics Engine is busy scoring requests. This can happen for complex models. If the Input queue size is reached, further requests with be rejected with 'ADAPA_Error' being populated in the Output response payload

Defaults to 1,000,000 Input requests

setThreadCount

action<integer > setThreadCount
Set the number of threads to use by the Plugin

Defaults to number of cores avaiable on the host system

setThrottleThreshold

action<integer > setThrottleThreshold
Set the threshold when throttling of scoring requests should be started. This parameter determines the number of scoring requests that will be active in the predictive analytics engine at any given point of time. This is done to reduce the overall heap usage by the application and preventing the process going out of memory

Defaults to 10,000 requests being scored by the Predictive Analytics Engine

setWorkDirectory

action<string > setWorkDirectory
Set the path of the Work Directory

Used to store temporary files generated by the scoring engine

subscribeToChannel

action<string > subscribeToChannel
Plugin can optionally subscribe to a correlator channel and listen for Input requests

Responses can be sent back to the context on which the plugin is initialised or can be sent on a channel configured using sendToChannel