Determining a relationship recommendation for a natural language request

1. A computer-implemented method, comprising:

processing a first natural language (NL) request, wherein the first NL request includes a first artifact and a second artifact;

determining the first artifact and the second artifact are not linked in a first NL language processing system;

identifying a first data set associated with the first artifact and a second data set associated with the second artifact;

generating a first data relationship based on the first data set and the second data set, wherein the first data relationship specifies how the first artifact and the second artifact are to be linked; and

causing the first data relationship to be provided to the user.

2. The computer-implemented method of claim 1 , further comprising:

generating a first data model recommendation by:

determining that a first data model associated with the first data relationship is unavailable in the first NL processing system;

identifying, based on the first artifact, a plurality of candidate data models included in a first data storage system associated with a first Domain Specific Language (DSL) processing system; and

selecting a first candidate data model from the plurality of candidate data models.

3. The computer-implemented method of claim 1 , wherein the first artifact comprises one of a data model, an entity, a named entity, an attribute, a fixed value, a synonym, or a context.

4. The computer-implemented method of claim 1 , wherein generating the first data relationship recommendation comprises:

for each of the first artifact and the second artifact, identifying a corresponding artifact stored in the first NL language processing system to generate a first set of corresponding artifacts;

combining two or more corresponding artifacts of the first set of corresponding artifacts to generate the first data relationship.

5. The computer-implemented method of claim 1 , further comprising:

determining that the user selected the first data relationship; and

storing the first data relationship in the first NL language processing system.

6. The computer-implemented method of claim 1 , further comprising:

storing the first data relationship in the first NL language processing system;

mapping the first data relationship to an intent associated with the first NL request;

extracting a first set of artifacts from the intent; and

generating a first training data set based on the first set of artifacts.

7. The computer-implemented method of claim 1 , further comprising:

storing the first data relationship in the first NL language processing system;

identifying a first set of artifacts based on the first data relationship; and

generating a first training data set based on the first set of artifacts, wherein the first training data set comprises a JavaScript Object Notation (JSON) file storing the first set of artifacts and the first data relationship.

8. The computer-implemented method of claim 1 , further comprising:

storing the first data relationship in the first NL language processing system;

identifying a first set of artifacts based on the first data relationship; and

generating a first training data set based on the first set of artifacts, wherein the first training data set includes a user intent associated with the first da relationship.

9. The computer-implemented method of claim 1 , further comprising:

storing the first data relationship in the first NL language processing system;

identifying a first set of artifacts based on the first data relationship; and

generating a first training data set based on the first set of artifacts by:

selecting a first NL template associated with the first NL processing system, the first NL template including a first set of fields;

for each field included in the first set of fields, extracting, from the first data relationship, a corresponding artifact to generate a second set of artifacts;

generating the first training data set based on a second set of the artifacts and a first DSL template; and

publishing the first training data set.

10. The computer-implemented method of claim 1 , further comprising:

storing the first data relationship in the first NL language processing system;

identifying a first set of artifacts based on the first data relationship; and

generating a first training data set based on the first set of artifacts by:

selecting a first NL template associated with the first NL processing system, wherein the first NL template is one of: an Alexa skill template, a Google Dialog template, a first artificial intelligence modeling language (AIML) template, or a first Artificial Intelligence: RiveScript (AiRS) template, and wherein the first NL template including a first set of fields;

for each field included in the first set of fields, extracting, from the first relationship, a corresponding artifact to generate a second set of artifacts;

generating the first training data set based on a second set of the artifacts and a first DSL template; and

publishing the first training data set.

11. One or more non-transitory computer-readable storage media including instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of:

processing a first natural language (NL) request, wherein the first NL request identifies a first artifact and a second artifact;

determining the first artifact and the second artifact are not linked in a first NL language processing system;

identifying a first data set associated with the first artifact and a second data set associated with the second artifact;

generating a first data relationship based on the first data set and the second data set, wherein the first data relationship specifies how the first artifact and the second artifact are to be linked; and

causing the first data relationship to be provided to a user.

12. The one or more non-transitory computer-readable storage media of claim 11 , further comprising:

generating a first data model recommendation by:

determining that a first data model associated with the first data relationship is unavailable in the first NL processing system;

identifying, based on the first artifact, a plurality of candidate data models included in a first data storage system associated with a first domain-specific language (DSL) language processing system; and

selecting a first candidate data model from the plurality of candidate data models.

13. The one or more non-transitory computer-readable storage media of claim 11 , wherein the first artifact comprises one of a data model, an entity, a named entity, an attribute, a fixed value, a synonym, or a context.

14. The one or more non-transitory computer-readable storage media of claim 11 , wherein generating the first data relationship recommendation comprises:

for each of the first artifact and the second artifact, identifying a corresponding artifact stored in the first NL language processing system to generate a first set of corresponding artifacts;

combining two or more corresponding artifacts of the first set of corresponding artifacts to generate the first data relationship.

15. The one or more non-transitory computer-readable storage media of claim 11 , further comprising:

determining that the user selected the first data relationship; and

storing the first data relationship in the first NL language processing system.

16. The one or more non-transitory computer-readable storage media of claim 11 , further comprising:

storing the first data relationship in the first NL language processing system;

mapping the first data relationship to an intent associated with the first NL request;

extracting a first set of artifacts from the intent; and

generating a first training data set based on the first set of artifacts.

17. The one or more non-transitory computer-readable storage media of claim 11 , further comprising:

storing the first data relationship in the first NL language processing system;

identifying a first set of artifacts based on the first data relationship; and

generating a first training data set based on the first set of artifacts, wherein the first training data set comprises a JavaScript Object Notation (JSON) file storing the first set of artifacts and the first data relationship.

18. The one or more non-transitory computer-readable storage media of claim 11 , further comprising:

storing the first data relationship in the first NL language processing system;

identifying a first set of artifacts based on the first data relationship; and

generating a first training data set based on the first set of artifacts, wherein the first training data set includes a user intent associated with the first da relationship.

19. The one or more non-transitory computer-readable storage media of claim 11 , further comprising:

storing the first data relationship in the first NL language processing system;

identifying a first set of artifacts based on the first data relationship; and

generating a first training data set based on the first set of artifacts by:

selecting a first NL template associated with the first NL processing system, the first NL template including a first set of fields;

for each field included in the first set of fields, extracting, from the first data relationship, a corresponding artifact to generate a second set of artifacts;

generating the first training data set based on a second set of the artifacts and a first DSL template; and

publishing the first training data set.

20. A computing device, comprising:

a memory that includes an application; and

a processor that is coupled to the memory, and when executing the application, performs:

processing a first natural language (NL) request, wherein the first NL request identifies a first artifact and a second artifact;

determining the first artifact and the second artifact are not linked in a first NL language processing system;

identifying a first data set associated with the first artifact and a second data set associated with the second artifact;

generating a first data relationship based on the first data set and the second data set, wherein the first data relationship specifies how the first artifact and the second artifact are to be linked; and

causing the first data relationship to be provided to a user.