Configuring the Natural Language Engine service

nle.home

NLE base installation directory.

Default is the value of the NLE_HOME environment variable (see Data pack locations).

nle.dataHome

Base directory for the data files downloaded and/or created by NLE.

Note: Nuance recommends that you change the value to a path outside the installation directory so that NLE does not create any files in the installation directory.

nle.host

IP address or hostname of the machine on which the NLE instance is running. This property is used by NLE when providing model URLs to NTpE.

The value localhost can be used when NTpE and NLE are co-located on the same machine (and Nuance Resource Manager is disabled via rm.enabled).

nle.port

Port on which NLE listens for incoming requests, either HTTP or WebSocket.

Indicates Java 8 is installed on the Natural Language Engine service host. Only relevant when using Management Station.

Note: If using Management Station and running NLE on a host with Java 17, you must set usingJava8 to false in the Natural Language Engine service properties.

license.path

Port and IP address of the Nuance License Manager, using the format:

port@ip_address

rm.enabled

Allows the Natural Language Engine to register its capabilities with the Resource Manager.

When the NTpE property resourceManager : enabled is also enabled, NLE publishes its capabilities to the Resource Manager, and requests from the Resource Manager a suitable NTpE instance when it needs to perform tokenization for a given session.

rm.uris

Hostname (or IP address) and port (usually 9001) to connect to the Nuance Resource Manager listener.

Specifies the time to wait, in milliseconds, before the Natural Language Engine attempts to reconnect to Resource Manager.

rm.hostNameVerificationEnabled

The service requests SSL validation with a CA-signed certificate when registering with Nuance Resource Manager servers when using HTTPS.

ntpe.uri

Hostname (or IP address) and port of the NTpE instance that NLE will use to perform tokenization. The default port is 9092.

This value is used when ntpe.resourceManagerEnabled is false.

ntpe.resourceManagerEnabled

When enabled, NLE will locate a suitable NTpE resource via the Nuance Resource Manager, in which case the property above is ignored.

ntpe.hostNameVerificationEnabled

Specifies whether to use a self-signed certificate for SSL validation with NTpE or alternatively, to perform a complete validation with a CA-signed SSL certifcate.

resource.download.connectTimeout

Timeout, in milliseconds, when connecting to the remote server when downloading resources.

Default is 10 seconds.

resource.download.socketTimeout

Timeout, in milliseconds, when performing a resource download request from the remote server.

Default is 10 seconds.

resource.download.httpClientCache.enabled

Specifies whether the HTTP client cache for resources is used.

Note: Do not set this property to false, because it causes the NLE service to fail.

resource.download.httpClientCache.maxEntries

Maximum number of serialized cache entries that will be stored on disk in the HTTP client cache (by default ${nle.dataHome}/httpModelCache).

Note: These caches are small so it is acceptable to have them on disk.

Specifies whether NLE verifies the resource HTTPS server certificate when downloading resources.

resource.diskCache.maximumSizeMetric

resource.diskCache.maximumFileCount

Maximum number of resource files that can be stored in the disk cache.

This parameter is ignored when resource.diskCache.maximumSizeMetric (above) is set to Size, in which case parameter below applies.

resource.diskCache.maximumSizeInMegabytes

Maximum size of all resource files stored in the disk cache, in megabytes (MB).

This parameter is ignored when resource.diskCache.maximumSizeMetric is set to Count, in which case parameter above applies.

model.cache.maximumSize

Maximum number of semantic models that can be stored in the cache.

model.cache.expireAfterAccess

Specifies how long a model is kept in the cache after being accessed. After this interval (expressed in seconds), models are removed the next time the cache is cleaned.

Default is 5 days.

model.cache.maintenancePeriodInMinutes

Delay, in minutes, between cache maintenance operations. This indicates how often the disk cache is cleaned.

Default is 1 hour.

wordSet.cache.maximumSize

wordSet.cache.expireAfterAccess

Specifies how long a wordset is kept in the cache after being accessed. After this interval (expressed in seconds), wordsets are removed the next time the cache is cleaned.

Default: 5 days

wordSet.cache.maintenancePeriodInMinutes

Delay, in minutes, between cache maintenance operations. This indicates how often the disk cache is cleaned.

Default: 1 hour

watcher.enabled

Specifies whether the watcher service is enabled.

Note: Do not change unless instructed by Nuance.

watcher.alarms.enabled

Specifies whether NLE sends alarms to the watcher service.

Note: Do not change unless instructed by Nuance.

trust.maxClockDriftInMilliseconds

Specifies the drift tolerance (in milliseconds) between client and server hosts for security verification purposes. Based on difference in UUID timestamps, with a valid range of 1-120 seconds and default value of 10 seconds.

server.ssl.enabled

Specifies whether SSL (TLS) is enabled.

https.keyStore

Path to the keystore file, which holds a valid SSL certificate.

Note: Nuance highly recommends that you generate a new keystore and certificate.

https.keyStorePassword

Encrypted keystore password used to protect the keystore file.