QlikElixir.REST.NaturalLanguage (qlik_elixir v0.3.5)

View Source

REST API client for Qlik Cloud Natural Language (Insight Advisor).

Enables natural language queries and AI-powered analytics recommendations for Qlik apps.

Examples

# Ask a question
{:ok, response} = NaturalLanguage.ask("app-123", "What were total sales?", config: config)

# Continue a conversation
{:ok, response} = NaturalLanguage.ask(
  "app-123",
  "Show by region",
  config: config,
  conversation_id: "conv-123"
)

# Get analysis recommendations
{:ok, %{"recommendations" => recs}} = NaturalLanguage.get_recommendations("app-123", config: config)

# Get available fields for NL queries
{:ok, %{"fields" => fields}} = NaturalLanguage.get_fields("app-123", config: config)

Summary

Functions

ask(app_id, question, opts \\ [])

Asks a natural language question about an app's data.

get_model(app_id, opts \\ [])

Gets the natural language model info for an app.

list_analysis_types(app_id, opts \\ [])

Lists available analysis types for an app.

recommend(app_id, request, opts \\ [])

Gets analysis recommendations based on natural language or field selections.

Functions

ask(app_id, question, opts \\ [])

@spec ask(String.t(), String.t(), keyword()) ::
  {:ok, map()} | {:error, QlikElixir.Error.t()}

Asks a natural language question about an app's data.

This is a convenience wrapper around recommend/3 for simple text queries.

Parameters

  • app_id - The app ID to query.
  • question - The natural language question.

Options

  • :config - Required. The configuration struct.
  • :lang - Language code for the question (e.g., "en", "es").

Examples

{:ok, response} = NaturalLanguage.ask("app-123", "What were total sales?", config: config)

get_model(app_id, opts \\ [])

@spec get_model(
  String.t(),
  keyword()
) :: {:ok, map()} | {:error, QlikElixir.Error.t()}

Gets the natural language model info for an app.

Returns model status, supported languages, and vocabulary information.

list_analysis_types(app_id, opts \\ [])

@spec list_analysis_types(
  String.t(),
  keyword()
) :: {:ok, map()} | {:error, QlikElixir.Error.t()}

Lists available analysis types for an app.

Returns information about analysis types (breakdown, trend, comparison, etc.) that can be used with the app's data model.

Options

  • :config - Required. The configuration struct.

recommend(app_id, request, opts \\ [])

@spec recommend(String.t(), map(), keyword()) ::
  {:ok, map()} | {:error, QlikElixir.Error.t()}

Gets analysis recommendations based on natural language or field selections.

Parameters

  • app_id - The app ID to query.
  • request - Map with recommendation request:
    • :text - Natural language question (optional if fields provided).
    • :fields - List of field/master item selections (optional if text provided).
    • :target_analysis - Preferred analysis type (optional).

Options

  • :config - Required. The configuration struct.
  • :lang - Language code (e.g., "en", "es") via Accept-Language header.

Returns

A response containing recommended analyses based on the app's data model.

Supported Analysis Types

"breakdown", "changePoint", "comparison", "contribution", "correlation", "fact", "mutualInfo", "rank", "spike", "trend", "values"

Examples

# Natural language question
{:ok, response} = NaturalLanguage.recommend("app-123", %{text: "show sales by region"}, config: config)

# Field-based recommendation
{:ok, response} = NaturalLanguage.recommend("app-123", %{
  fields: [%{name: "Sales", type: "measure"}, %{name: "Region", type: "dimension"}]
}, config: config)