Rasa::HTTP::ModelApi

All URIs are relative to http://localhost:5005

Method	HTTP request	Description
parse_model_message	POST /model/parse	Parse a message using the Rasa model
predict_model_action	POST /model/predict	Predict an action on a temporary state
replace_model	PUT /model	Replace the currently loaded model
test_model_intent	POST /model/test/intents	Perform an intent evaluation
test_model_stories	POST /model/test/stories	Evaluate stories
train_model	POST /model/train	Train a Rasa model
unload_model	DELETE /model	Unload the trained model

parse_model_message

parse_model_message(parse_model_message_request, opts)

Parse a message using the Rasa model

Predicts the intent and entities of the message posted to this endpoint. No messages will be stored to a conversation and no action will be run. This will just retrieve the NLU parse results.

Examples

require 'time'
require 'rasa-http-api'
# setup authorization
Rasa::HTTP.configure do |config|
  # Configure Bearer authorization (JWT): JWT
  config.access_token = 'YOUR_BEARER_TOKEN'

  # Configure API key authorization: TokenAuth
  config.api_key['TokenAuth'] = 'YOUR API KEY'
  # Uncomment the following line to set a prefix for the API key, e.g. 'Bearer' (defaults to nil)
  # config.api_key_prefix['TokenAuth'] = 'Bearer'
end

api_instance = Rasa::HTTP::ModelApi.new
parse_model_message_request = Rasa::HTTP::ParseModelMessageRequest.new # ParseModelMessageRequest | 
opts = {
  emulation_mode: 'WIT' # String | Specify the emulation mode to use. Emulation mode transforms the response JSON to the format expected by the service specified as the emulation_mode. Requests must still be sent in the regular Rasa format.
}

begin
  # Parse a message using the Rasa model
  result = api_instance.parse_model_message(parse_model_message_request, opts)
  p result
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->parse_model_message: #{e}"
end

Using the parse_model_message_with_http_info variant

This returns an Array which contains the response data, status code and headers.

<Array(, Integer, Hash)> parse_model_message_with_http_info(parse_model_message_request, opts)

begin
  # Parse a message using the Rasa model
  data, status_code, headers = api_instance.parse_model_message_with_http_info(parse_model_message_request, opts)
  p status_code # => 2xx
  p headers # => { ... }
  p data # => <ParseResult>
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->parse_model_message_with_http_info: #{e}"
end

Parameters

Name	Type	Description	Notes
parse_model_message_request	ParseModelMessageRequest
emulation_mode	String	Specify the emulation mode to use. Emulation mode transforms the response JSON to the format expected by the service specified as the emulation_mode. Requests must still be sent in the regular Rasa format.	[optional]

Return type

ParseResult

Authorization

JWT, TokenAuth

HTTP request headers

Content-Type: application/json
Accept: application/json

predict_model_action

predict_model_action(event, opts)

Predict an action on a temporary state

Predicts the next action on the tracker state as it is posted to this endpoint. Rasa will create a temporary tracker from the provided events and will use it to predict an action. No messages will be sent and no action will be run.

Examples

require 'time'
require 'rasa-http-api'
# setup authorization
Rasa::HTTP.configure do |config|
  # Configure Bearer authorization (JWT): JWT
  config.access_token = 'YOUR_BEARER_TOKEN'

  # Configure API key authorization: TokenAuth
  config.api_key['TokenAuth'] = 'YOUR API KEY'
  # Uncomment the following line to set a prefix for the API key, e.g. 'Bearer' (defaults to nil)
  # config.api_key_prefix['TokenAuth'] = 'Bearer'
end

api_instance = Rasa::HTTP::ModelApi.new
event = [Rasa::HTTP::Event.new({event: 'active_loop', name: 'name_example', value: 3.56, entities: [Rasa::HTTP::EntitiesAddedEventAllOfEntities.new({entity: 'entity_example', value: 3.56})]})] # Array<Event> | 
opts = {
  include_events: 'ALL' # String | Specify which events of the tracker the response should contain. * `ALL` - every logged event. * `APPLIED` - only events that contribute to the trackers state. This excludes reverted utterances and actions that got undone. * `AFTER_RESTART` -  all events since the last `restarted` event.   This includes utterances that got reverted and actions that got undone.  * `NONE` - no events.
}

begin
  # Predict an action on a temporary state
  result = api_instance.predict_model_action(event, opts)
  p result
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->predict_model_action: #{e}"
end

Using the predict_model_action_with_http_info variant

This returns an Array which contains the response data, status code and headers.

<Array(, Integer, Hash)> predict_model_action_with_http_info(event, opts)

begin
  # Predict an action on a temporary state
  data, status_code, headers = api_instance.predict_model_action_with_http_info(event, opts)
  p status_code # => 2xx
  p headers # => { ... }
  p data # => <PredictResult>
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->predict_model_action_with_http_info: #{e}"
end

Parameters

Name	Type	Description	Notes
event	Array<Event>
include_events	String	Specify which events of the tracker the response should contain. * `ALL` - every logged event. * `APPLIED` - only events that contribute to the trackers state. This excludes reverted utterances and actions that got undone. * `AFTER_RESTART` - all events since the last `restarted` event. This includes utterances that got reverted and actions that got undone. * `NONE` - no events.	[optional][default to 'AFTER_RESTART']

Return type

PredictResult

Authorization

JWT, TokenAuth

HTTP request headers

Content-Type: application/json
Accept: application/json

replace_model

replace_model(model_request)

Replace the currently loaded model

Updates the currently loaded model. First, tries to load the model from the local (note: local to Rasa server) storage system. Secondly, tries to load the model from the provided model server configuration. Last, tries to load the model from the provided remote storage.

Examples

require 'time'
require 'rasa-http-api'
# setup authorization
Rasa::HTTP.configure do |config|
  # Configure Bearer authorization (JWT): JWT
  config.access_token = 'YOUR_BEARER_TOKEN'

  # Configure API key authorization: TokenAuth
  config.api_key['TokenAuth'] = 'YOUR API KEY'
  # Uncomment the following line to set a prefix for the API key, e.g. 'Bearer' (defaults to nil)
  # config.api_key_prefix['TokenAuth'] = 'Bearer'
end

api_instance = Rasa::HTTP::ModelApi.new
model_request = Rasa::HTTP::ModelRequest.new # ModelRequest | 

begin
  # Replace the currently loaded model
  api_instance.replace_model(model_request)
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->replace_model: #{e}"
end

Using the replace_model_with_http_info variant

This returns an Array which contains the response data (nil in this case), status code and headers.

<Array(nil, Integer, Hash)> replace_model_with_http_info(model_request)

begin
  # Replace the currently loaded model
  data, status_code, headers = api_instance.replace_model_with_http_info(model_request)
  p status_code # => 2xx
  p headers # => { ... }
  p data # => nil
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->replace_model_with_http_info: #{e}"
end

Parameters

Name	Type	Description	Notes
model_request	ModelRequest

Return type

nil (empty response body)

Authorization

JWT, TokenAuth

HTTP request headers

Content-Type: application/json
Accept: application/json

test_model_intent

test_model_intent(body, opts)

Perform an intent evaluation

Evaluates NLU model against a model or using cross-validation.

Examples

require 'time'
require 'rasa-http-api'
# setup authorization
Rasa::HTTP.configure do |config|
  # Configure Bearer authorization (JWT): JWT
  config.access_token = 'YOUR_BEARER_TOKEN'

  # Configure API key authorization: TokenAuth
  config.api_key['TokenAuth'] = 'YOUR API KEY'
  # Uncomment the following line to set a prefix for the API key, e.g. 'Bearer' (defaults to nil)
  # config.api_key_prefix['TokenAuth'] = 'Bearer'
end

api_instance = Rasa::HTTP::ModelApi.new
body = 'body_example' # String | 
opts = {
  model: 'rasa-model.tar.gz', # String | Model that should be used for evaluation. If the parameter is set, the model will be fetched with the currently loaded configuration setup. However, the currently loaded model will not be updated. The state of the server will not change. If the parameter is not set, the currently loaded model will be used for the evaluation.
  callback_url: 'https://example.com/rasa_evaluations', # String | If specified the call will return immediately with an empty response and status code 204. The actual result or any errors will be sent to the given callback URL as the body of a post request.
  cross_validation_folds: 56 # Integer | 
}

begin
  # Perform an intent evaluation
  result = api_instance.test_model_intent(body, opts)
  p result
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->test_model_intent: #{e}"
end

Using the test_model_intent_with_http_info variant

This returns an Array which contains the response data, status code and headers.

<Array(, Integer, Hash)> test_model_intent_with_http_info(body, opts)

begin
  # Perform an intent evaluation
  data, status_code, headers = api_instance.test_model_intent_with_http_info(body, opts)
  p status_code # => 2xx
  p headers # => { ... }
  p data # => <NLUEvaluationResult>
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->test_model_intent_with_http_info: #{e}"
end

Parameters

Name	Type	Description	Notes
body	String
model	String	Model that should be used for evaluation. If the parameter is set, the model will be fetched with the currently loaded configuration setup. However, the currently loaded model will not be updated. The state of the server will not change. If the parameter is not set, the currently loaded model will be used for the evaluation.	[optional]
callback_url	String	If specified the call will return immediately with an empty response and status code 204. The actual result or any errors will be sent to the given callback URL as the body of a post request.	[optional][default to 'None']
cross_validation_folds	Integer		[optional]

Return type

NLUEvaluationResult

Authorization

JWT, TokenAuth

HTTP request headers

Content-Type: application/x-yaml, application/json
Accept: application/json

test_model_stories

test_model_stories(body, opts)

Evaluate stories

Evaluates one or multiple stories against the currently loaded Rasa model.

Examples

require 'time'
require 'rasa-http-api'
# setup authorization
Rasa::HTTP.configure do |config|
  # Configure Bearer authorization (JWT): JWT
  config.access_token = 'YOUR_BEARER_TOKEN'

  # Configure API key authorization: TokenAuth
  config.api_key['TokenAuth'] = 'YOUR API KEY'
  # Uncomment the following line to set a prefix for the API key, e.g. 'Bearer' (defaults to nil)
  # config.api_key_prefix['TokenAuth'] = 'Bearer'
end

api_instance = Rasa::HTTP::ModelApi.new
body = 'body_example' # String | 
opts = {
  e2e: false # Boolean | Perform an end-to-end evaluation on the posted stories.
}

begin
  # Evaluate stories
  result = api_instance.test_model_stories(body, opts)
  p result
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->test_model_stories: #{e}"
end

Using the test_model_stories_with_http_info variant

This returns an Array which contains the response data, status code and headers.

<Array(, Integer, Hash)> test_model_stories_with_http_info(body, opts)

begin
  # Evaluate stories
  data, status_code, headers = api_instance.test_model_stories_with_http_info(body, opts)
  p status_code # => 2xx
  p headers # => { ... }
  p data # => <EvaluationStoriesResult>
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->test_model_stories_with_http_info: #{e}"
end

Parameters

Name	Type	Description	Notes
body	String
e2e	Boolean	Perform an end-to-end evaluation on the posted stories.	[optional][default to false]

Return type

EvaluationStoriesResult

Authorization

JWT, TokenAuth

HTTP request headers

Content-Type: text/yml
Accept: application/json

train_model

File train_model(yaml_training_request, opts)

Train a Rasa model

Trains a new Rasa model. Depending on the data given only a dialogue model, only a NLU model, or a model combining a trained dialogue model with an NLU model will be trained. The new model is not loaded by default.

Examples

require 'time'
require 'rasa-http-api'
# setup authorization
Rasa::HTTP.configure do |config|
  # Configure Bearer authorization (JWT): JWT
  config.access_token = 'YOUR_BEARER_TOKEN'

  # Configure API key authorization: TokenAuth
  config.api_key['TokenAuth'] = 'YOUR API KEY'
  # Uncomment the following line to set a prefix for the API key, e.g. 'Bearer' (defaults to nil)
  # config.api_key_prefix['TokenAuth'] = 'Bearer'
end

api_instance = Rasa::HTTP::ModelApi.new
yaml_training_request = Rasa::HTTP::YAMLTrainingRequest.new # YAMLTrainingRequest | The training data should be in YAML format.
opts = {
  save_to_default_model_directory: true, # Boolean | 
  force_training: true, # Boolean | 
  augmentation: 'augmentation_example', # String | 
  num_threads: 'num_threads_example', # String | 
  callback_url: 'https://example.com/rasa_evaluations' # String | If specified the call will return immediately with an empty response and status code 204. The actual result or any errors will be sent to the given callback URL as the body of a post request.
}

begin
  # Train a Rasa model
  result = api_instance.train_model(yaml_training_request, opts)
  p result
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->train_model: #{e}"
end

Using the train_model_with_http_info variant

This returns an Array which contains the response data, status code and headers.

<Array(File, Integer, Hash)> train_model_with_http_info(yaml_training_request, opts)

begin
  # Train a Rasa model
  data, status_code, headers = api_instance.train_model_with_http_info(yaml_training_request, opts)
  p status_code # => 2xx
  p headers # => { ... }
  p data # => File
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->train_model_with_http_info: #{e}"
end

Parameters

Name	Type	Description	Notes
yaml_training_request	YAMLTrainingRequest	The training data should be in YAML format.
save_to_default_model_directory	Boolean		[optional][default to true]
force_training	Boolean		[optional][default to false]
augmentation	String		[optional][default to '50']
num_threads	String		[optional][default to '1']
callback_url	String	If specified the call will return immediately with an empty response and status code 204. The actual result or any errors will be sent to the given callback URL as the body of a post request.	[optional][default to 'None']

Return type

File

Authorization

JWT, TokenAuth

HTTP request headers

Content-Type: application/yaml
Accept: application/octet-stream, application/json

unload_model

unload_model

Unload the trained model

Unloads the currently loaded trained model from the server.

Examples

require 'time'
require 'rasa-http-api'
# setup authorization
Rasa::HTTP.configure do |config|
  # Configure Bearer authorization (JWT): JWT
  config.access_token = 'YOUR_BEARER_TOKEN'

  # Configure API key authorization: TokenAuth
  config.api_key['TokenAuth'] = 'YOUR API KEY'
  # Uncomment the following line to set a prefix for the API key, e.g. 'Bearer' (defaults to nil)
  # config.api_key_prefix['TokenAuth'] = 'Bearer'
end

api_instance = Rasa::HTTP::ModelApi.new

begin
  # Unload the trained model
  api_instance.unload_model
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->unload_model: #{e}"
end

Using the unload_model_with_http_info variant

This returns an Array which contains the response data (nil in this case), status code and headers.

<Array(nil, Integer, Hash)> unload_model_with_http_info

begin
  # Unload the trained model
  data, status_code, headers = api_instance.unload_model_with_http_info
  p status_code # => 2xx
  p headers # => { ... }
  p data # => nil
rescue Rasa::HTTP::ApiError => e
  puts "Error when calling ModelApi->unload_model_with_http_info: #{e}"
end

Parameters

This endpoint does not need any parameter.

Return type

nil (empty response body)

Authorization

JWT, TokenAuth

HTTP request headers

Content-Type: Not defined
Accept: application/json

Files

ModelApi.md

Latest commit

History

ModelApi.md

File metadata and controls

Rasa::HTTP::ModelApi

parse_model_message

Examples

Using the parse_model_message_with_http_info variant

Parameters

Return type

Authorization

HTTP request headers

predict_model_action

Examples

Using the predict_model_action_with_http_info variant

Parameters

Return type

Authorization

HTTP request headers

replace_model

Examples

Using the replace_model_with_http_info variant

Parameters

Return type

Authorization

HTTP request headers

test_model_intent

Examples

Using the test_model_intent_with_http_info variant

Parameters

Return type

Authorization

HTTP request headers

test_model_stories

Examples

Using the test_model_stories_with_http_info variant

Parameters

Return type

Authorization

HTTP request headers

train_model

Examples

Using the train_model_with_http_info variant

Parameters

Return type

Authorization

HTTP request headers

unload_model

Examples

Using the unload_model_with_http_info variant

Parameters

Return type

Authorization

HTTP request headers