Configuring the Natural Language Processing service
The NLP service communicates with the Nuance Resource Manager to allocate suitable Krypton recognition engine and Natural Language Engine resources based on the capabilities requested by the voice application.
Management Station displays commonly-used properties on the Settings tab, and the following tables shows some of these properties. (Management Station does not display less-common parameters, but you can add them to the display at any time. For a complete list, see Speech product parameters.)
Commonly-used service properties
Use the Management Station Settings tab to modify or add these service properties:
|
Service property |
Description |
Data type |
Default |
|---|---|---|---|
|
Listening port (WebSocket) used by the Natural Language Processing service. |
INT |
8443 |
|
| server.ssl.enabled |
Specifies whether SSL (TLS) is enabled. Note: Nuance recommends that you use different ports for secure and unsecured connections; therefore, if you change this property from the default, consider modifying |
BOOLEAN | true |
| httpClient.requestTimeout |
Maximum seconds that the NLP service waits for a response from the resource manager when allocating a resource. |
INT | 5000 (milliseconds) |
| licenseManager.address |
Port and IP address of the Nuance License Manager, using the format: port@ip_address |
STRING | 27000@localhost |
|
Number of licenses to acquire at startup. |
INT |
1 |
|
| resourceManager.urls |
Hostname or IP address and port (usually 9001) of the Nuance Resource Manager. |
STRING |
[https://hostname:port] |
| showEntityConfidence |
Displays the entity level confidence, value, literal, and origin in the legacy HTTP and Websocket responses available from NLE recognition results. |
BOOLEAN | false |
| suppressSensitiveData | Suppresses sensitive data when writing to NLP service diagnostic logs. | BOOLEAN | true |
| resultNbestExtraKeys |
Adds the Krypton recognition engine confidence score to the XML result. |
STRING | (empty) |
HTTP and HTTPS client settings
Use the Management Station Settings tab to add or modify add these service properties.
The httpClient properties control how the NLP service behaves as a client. (For example, when the service downloads resources.) As a client, NLPS can send both HTTP and HTTPS requests. Some of these properties are specifically for HTTPS connections to servers while others are generic and apply to both HTTP and HTTPS.
|
Service property |
Description |
Data type |
Default |
|---|---|---|---|
|
Allows the NLP service to treat 3xx responses as redirects. (Applies to both HTTP and HTTPS.) |
BOOLEAN | true | |
|
Specifies the maximum number of HTTP or HTTPS redirects to follow, when httpClient.followRedirects is enabled. |
INT | 10 | |
|
The service requests SSL validation with a CA-signed certificate when registering with Nuance Resource Manager servers when using HTTPS. |
BOOLEAN | false | |
|
Path to the service's keystore file. |
STRING (a filepath) | ||
|
A keystore password for the NLP service file corresponding to httpClient.keystore. |
String (clear text, not encrypted) |
||
| httpClient.metadata.trustAll | Enables NLP service to trust all certificates received from application servers reached via an HTTPS connection. | BOOLEAN | true |
|
The path to the truststore file for the NLP service. |
STRING (a filepath) | ||
|
The encrypted passphrase for the service's truststore file. |
String (clear text, not encrypted) |
||
| httpClient.executorMaxThreads |
Maximum number of threads created by individual Dragon Voice services. |
INT | 400 |
| httpClient.maxConnectionsPerDestination |
Maximum number of network connections accepted by individual Dragon Voice services. |
INT | 400 |
| webSocket.policy.maxTextMessageBufferSize |
Maximum size (bytes) of text messages exchanged between NLP service and other Dragon Voice services. |
INT | 32768 |
| webSocket.policy.maxTextMessageSize |
Maximum size (bytes) of text messages exchanged between NLP service and other Dragon Voice services. |
INT | 262144 |
