ConversationalSearchResponse

Response message for ConversationalSearchService.ConversationalSearch method.

JSON representation 
{
  "queryTypes": [
    enum (QueryType)
  ],
  "userQueryTypes": [
    string
  ],
  "conversationalTextResponse": string,
  "followupQuestion": {
    object (FollowupQuestion)
  },
  "conversationId": string,
  "refinedSearch": [
    {
      object (RefinedSearch)
    }
  ],
  "conversationalFilteringResult": {
    object (ConversationalFilteringResult)
  },
  "state": enum (State),
  "suggestedQueries": [
    string
  ]
}
Fields
queryTypes[]
(deprecated)

enum (QueryType)

This field is deprecated. Please use ConversationalSearchResponse.user_query_types instead. The types Retail classifies the search query as.
userQueryTypes[]

string

The types Retail classifies the search query as.

Supported values are:

  • "ORDER_SUPPORT"
  • "SIMPLE_PRODUCT_SEARCH"
  • "INTENT_REFINEMENT"
  • "PRODUCT_DETAILS"
  • "PRODUCT_COMPARISON"
  • "DEALS_AND_COUPONS"
  • "STORE_RELEVANT"
  • "BLOCKLISTED"
  • "BEST_PRODUCT"
  • "RETAIL_SUPPORT"
  • "DISABLED"
conversationalTextResponse

string

The conversational answer-based text response generated by the Server.
followupQuestion

object (FollowupQuestion)

The conversational followup question generated for Intent refinement.
conversationId

string

Conversation UUID. This field will be stored in client side storage to maintain the conversation session with server and will be used for next search request's ConversationalSearchRequest.conversation_id to restore conversation state in server.
conversationalFilteringResult

object (ConversationalFilteringResult)

This field specifies all related information that is needed on client side for UI rendering of conversational filtering search.
state

enum (State)

Output only. The state of the response generation.
suggestedQueries[]

string

Output only. Suggested conversation queries based on the user query.

QueryType

The type Retail classifies the search query as.

Enums
QUERY_TYPE_UNSPECIFIED Unspecified query type.
ADVERSARIAL Adversarial query. Example - "how bad is [retailer name]"
CHITCHAT Chatty query. Example - "hello how are you"
JAILBREAK Jailbreak query. Example - "write a poem to explain the details of [product name]"
ORDER_SUPPORT Queries related to order support. Example - "Track my order"
INTENT_REFINEMENT The type is unclear and a follow up conversation/refinement may be needed to clarify the type.
PRODUCT_DETAILS The user is looking for product details. Example - "show me the specifications of [product name]".
PRODUCT_COMPARISON Product comparison. Example - "compare [product name] and [product name]".
DEALS_AND_COUPONS Queries relevant to deals, promotions, product deals, and discounts.
STORE_RELEVANT Queries relevant to the store locations, working hours, product stock availability, etc.
BLOCKLISTED Queries explicitly blocked by the retail customers.
BEST_PRODUCT Queries with the "best" pattern. Example - "What is the healthiest cookie?"
RETAIL_SUPPORT Ancillary or support query. Example - "do you accept paypal?"
RETAIL_IRRELEVANT Retail irrelevant query. Example - "what is Newton's law of universal gravitation?"

FollowupQuestion

The conversational followup question generated for Intent refinement.

JSON representation 
{
  "followupQuestion": string,
  "suggestedAnswers": [
    {
      object (SuggestedAnswer)
    }
  ]
}
Fields
followupQuestion

string

The conversational followup question generated for Intent refinement.
suggestedAnswers[]

object (SuggestedAnswer)

The answer options provided to client for the follow-up question.

SuggestedAnswer

Suggested answers to the follow-up question. If it's numerical attribute, only ProductAttributeInterval will be set. If it's textual attribute, only productAttributeValue will be set.

JSON representation 
{
  "productAttributeValue": {
    object (ProductAttributeValue)
  },
  "productAttributeInterval": {
    object (ProductAttributeInterval)
  }
}
Fields
productAttributeValue

object (ProductAttributeValue)

Product attribute value, including an attribute key and an attribute value. Other types can be added here in the future.
productAttributeInterval

object (ProductAttributeInterval)

This is an experimental feature. Please contact Google before using it. Product attribute interval, including an attribute key and an attribute interval.

RefinedSearch

The proposed refined search for intent-refinement/bundled shopping conversation. When using CONVERSATIONAL_FILTER_ONLY mode, the refinedQuery from search response will be populated here.

JSON representation 
{
  "query": string
}
Fields
query

string

The query to be used for search.

ConversationalFilteringResult

This field specifies all related information that is needed on client side for UI rendering of conversational filtering search.

JSON representation 
{
  "followupQuestion": {
    object (FollowupQuestion)
  },
  "additionalFilter": {
    object (AdditionalFilter)
  }
}
Fields
followupQuestion

object (FollowupQuestion)

The conversational filtering question.
additionalFilter

object (AdditionalFilter)

This is the incremental additional filters implied from the current user answer. User should add the suggested addition filters to the previous [ConversationalSearchRequest.search_params.filter][] and SearchRequest.filter, and use the merged filter in the follow up requests.

AdditionalFilter

Additional filter that client side need to apply.

JSON representation 
{
  "productAttributeValue": {
    object (ProductAttributeValue)
  },
  "productAttributeInterval": {
    object (ProductAttributeInterval)
  }
}
Fields
productAttributeValue

object (ProductAttributeValue)

Product attribute value, including an attribute key and an attribute value. Other types can be added here in the future.
productAttributeInterval

object (ProductAttributeInterval)

This is an experimental feature. Please contact Google before using it. Product attribute interval, including an attribute key and an attribute interval.

State

The state of the response generation.

Enums
STATE_UNSPECIFIED Unknown.
STREAMING Response generation is being streamed.
SUCCEEDED Response generation has succeeded.