package nvidia.jarvis.nlp

Get desktop application:
View/edit binary Protocol Buffers messages

The Jarvis Core NLP Service provides generic NLP services for custom model use cases. The intent of this service is to allow users to design models for arbitrary use cases that conform simply with input and output types specified in the service. As an explicit example, the ClassifyText function could be used for sentiment classification, domain recognition, language identification, etc.

• rpc ClassifyText (TextClassRequest, TextClassResponse)

  jarvis_nlp_core.proto:28

  ClassifyText takes as input an input/query string and parameters related to the requested model to use to evaluate the text. The service evaluates the text with the requested model, and returns one or more classifications.

  message TextClassRequest

  jarvis_nlp_core.proto:68

  TextClassRequest is the input message to the ClassifyText service.

  • repeated string text = 1

    Each repeated text element is handled independently for handling multiple input strings with a single request

  • uint32 top_n = 2

    Return the top N classification results for each input. 0 or 1 will return top class, otherwise N. Note: Current disabled.

  • optional NLPModelParams model = 3

  message TextClassResponse

  jarvis_nlp_core.proto:99

  TextClassResponse is the return message from the ClassifyText service.

  • repeated ClassificationResult results = 1

• rpc ClassifyTokens (TokenClassRequest, TokenClassResponse)

  jarvis_nlp_core.proto:34

  ClassifyTokens takes as input either a string or list of tokens and parameters related to which model to use. The service evaluates the text with the requested model, performing additional tokenization if necessary, and returns one or more class labels per token.

  message TokenClassRequest

  jarvis_nlp_core.proto:104

  TokenClassRequest is the input message to the ClassifyText service.

  • repeated string text = 1

    Each repeated text element is handled independently for handling multiple input strings with a single request

  • uint32 top_n = 3

    Return the top N classification results for each input. 0 or 1 will return top class, otherwise N. Note: Current disabled.

  • optional NLPModelParams model = 4

• rpc TransformText (TextTransformRequest, TextTransformResponse)

  jarvis_nlp_core.proto:40

  TransformText takes an input/query string and parameters related to the requested model and returns another string. The behavior of the function is defined entirely by the underlying model and may be used for tasks like translation, adding punctuation, augment the input directly, etc.

Jarvis NLP Services implement task-specific APIs for popular NLP tasks including intent recognition (as well as slot filling), and entity extraction.

• rpc AnalyzeEntities (AnalyzeEntitiesRequest, TokenClassResponse)

  jarvis_nlp.proto:23

  AnalyzeEntities accepts an input string and returns all named entities within the text, as well as a category and likelihood.

  message AnalyzeEntitiesRequest

  jarvis_nlp.proto:103

  AnalyzeEntitiesRequest is the input message for the AnalyzeEntities service

  • string query = 1

    The string to analyze for intent and slots

  • optional AnalyzeEntitiesOptions options = 2

    Optional configuration for the request, including providing context from previous turns and hardcoding a domain/language

• rpc AnalyzeIntent (AnalyzeIntentRequest, AnalyzeIntentResponse)

  jarvis_nlp.proto:31

  AnalyzeIntent accepts an input string and returns the most likely intent as well as slots relevant to that intent. The model requires that a valid "domain" be passed in, and optionally supports including a previous intent classification result to provide context for the model.

  message AnalyzeIntentRequest

  jarvis_nlp.proto:66

  AnalyzeIntentRequest is the input message for the AnalyzeIntent service

  • string query = 1

    The string to analyze for intent and slots

  • optional AnalyzeIntentOptions options = 2

    Optional configuration for the request, including providing context from previous turns and hardcoding a domain/language

  message AnalyzeIntentResponse

  jarvis_nlp.proto:76

  AnalyzeIntentResponse is returned by the AnalyzeIntent service, and includes information related to the query's intent, (optionally) slot data, and its domain.

  • optional Classification intent = 1

    Intent classification result, including the label and score

  • repeated TokenClassValue slots = 2

    List of tokens explicitly marked as filling a slot relevant to the intent, where the tokens may not exactly match the input (based on the recombined values after tokenization)

  • string domain_str = 3

    Returns the inferred domain for the query if not hardcoded in the request. In the case where the domain was hardcoded in AnalyzeIntentRequest, the returned domain is an exact match to the request. In the case where no domain matches the query, intent and slots will be unset. DEPRECATED, use Classification domain field.

  • optional Classification domain = 4

    Returns the inferred domain for the query if not hardcoded in the request. In the case where the domain was hardcoded in AnalyzeIntentRequest, the returned domain is an exact match to the request. In the case where no domain matches the query, intent and slots will be unset.

• rpc PunctuateText (TextTransformRequest, TextTransformResponse)

  jarvis_nlp.proto:35

  PunctuateText takes text with no- or limited- punctuation and returns the same text with corrected punctuation and capitalization.

• rpc NaturalQuery (NaturalQueryRequest, NaturalQueryResponse)

  jarvis_nlp.proto:39

  NaturalQuery is a search function that enables querying one or more documents or contexts with a query that is written in natural language.

  message NaturalQueryRequest

  jarvis_nlp.proto:111

  • string query = 1

    The natural language query

  • uint32 top_n = 2

    Maximum number of answers to return for the query. Defaults to 1 if not set.

  • string context = 3

    Context to search with the above query

  message NaturalQueryResponse

  jarvis_nlp.proto:129

  • repeated NaturalQueryResult results = 1

AnalyzeEntitiesOptions is an optional configuration message to be sent as part of an AnalyzeEntitiesRequest with query metadata

Used in: AnalyzeEntitiesRequest

• string lang = 4

  Optional language field. Assumed to be "en-US" if not specified.

AnalyzeIntentContext is reserved for future use when we may send context back in a a variety of different formats (including raw neural network hidden states)

Reserved for future use

Used in: AnalyzeIntentOptions

(message has no fields)

AnalyzeIntentOptions is an optional configuration message to be sent as part of an AnalyzeIntentRequest with query metadata

Used in: AnalyzeIntentRequest

• oneof context

  Optionally provide context from previous interactions to bias the model's prediction

  • string previous_intent = 1

  • AnalyzeIntentContext vectors = 2

• string domain = 3

  Optional domain field. Domain must be supported otherwise an error will be returned. If left blank, a domain detector will be run first and then the query routed to the appropriate intent classifier (if it exists)

• string lang = 4

  Optional language field. Assumed to be "en-US" if not specified.

Classification messages return a class name and corresponding score

Used in: AnalyzeIntentResponse, ClassificationResult, TokenClassValue

• string class_name = 1

• float score = 2

ClassificationResults contain zero or more Classification messages If the number of Classifications is > 1, top_n > 1 must have been specified.

Used in: TextClassResponse

• repeated Classification labels = 1

NLPModelParams is a metadata message that is included in every request message used by the Core NLP Service and is used to specify model characteristics/requirements

Used in: TextClassRequest, TextTransformRequest, TokenClassRequest

• string model_name = 1

  Requested model to use. If unavailable, the request will return an error

Used in: NaturalQueryResponse

• string answer = 1

  text which answers the query

• float score = 2

  Score representing confidence in result

Span of a particular result

Used in: TokenClassValue

• uint32 start = 1

• uint32 end = 2

TextTransformRequest is a request type intended for services like TransformText which take an arbitrary text input

Used as request type in: JarvisCoreNLP.TransformText, JarvisNLP.PunctuateText

• repeated string text = 1

  Each repeated text element is handled independently for handling multiple input strings with a single request

• uint32 top_n = 2

• optional NLPModelParams model = 3

TextTransformResponse is returned by the TransformText method. Responses are returned in the same order as they were requested.

Used as response type in: JarvisCoreNLP.TransformText, JarvisNLP.PunctuateText

• repeated string text = 1

TokenClassResponse returns a single TokenClassSequence per input request

Used as response type in: JarvisCoreNLP.ClassifyTokens, JarvisNLP.AnalyzeEntities

• repeated TokenClassSequence results = 1

TokenClassSequence is used for returning a sequence of TokenClassValue objects in the original order of input tokens

Used in: TokenClassResponse

• repeated TokenClassValue results = 1

TokenClassValue is used to correlate an input token with its classification results

Used in: AnalyzeIntentResponse, TokenClassSequence

• string token = 1

• repeated Classification label = 2

• repeated Span span = 3