Generating application programming interface knowledge graphs

A non-transitory computer-readable medium having encoded therein programming code executable by a processor to perform operations comprising:

• obtaining first application programming interface (API) documentation associated with a first API;
• obtaining second API documentation associated with a second API;
• mining the first API documentation to extract a first plurality of semantic triples;
• mining the second API documentation to extract a second plurality of semantic triples;
• selecting a first subset of semantic triples from the first plurality of semantic triples;
• selecting a second subset of semantic triples from the second plurality of semantic triples;
• generating a first ontology based on the first API documentation and the first subset of semantic triples, the first ontology including one or more first concepts associated with the first API, one or more first attributes associated with the one or more first concepts, and one or more first taxonomical relationships between the one or more first concepts;
• generating a second ontology based on the second API documentation and the second subset of semantic triples, the second ontology including one or more second concepts associated with the second API, one or more second attributes associated with the one or more second concepts, and one or more second taxonomical relationships between the one or more second concepts;
• correlating the first ontology with the second ontology based on the one or more first concepts of the first ontology and the one or more second concepts of the second ontology;
• based on the first ontology, the second ontology, and the correlating the first ontology with the second ontology, generating a knowledge graph of the first API and the second API;
• obtaining a natural language user query provided by a user;
• converting the natural language user query to a semantic user query using natural language understanding;
• searching the knowledge graph using the semantic user query to identify an API call as a response to the natural language user query;
• executing the API call using the knowledge graph and the natural language user query to provide a result for the natural language user query;
• and
• providing the result to the user.

A system comprising:

• one or more processors;
• and
• one or more computer-readable media configured to store instructions that in response to being executed by the one or more processors cause the system to perform operations, the operations comprising:
• obtaining first application programming interface (API) documentation associated with a first API;
• obtaining second API documentation associated with a second API;
• mining the first API documentation to extract a first plurality of semantic triples;
• mining the second API documentation to extract a second plurality of semantic triples;
• selecting a first subset of semantic triples from the first plurality of semantic triples;
• selecting a second subset of semantic triples from the second plurality of semantic triples;
• generating a first ontology based on the first API documentation and the first subset of semantic triples, the first ontology including one or more first concepts associated with the first API, one or more first attributes associated with the one or more first concepts, and one or more first taxonomical relationships between the one or more first concepts;
• generating a second ontology based on the second API documentation and the second subset of semantic triples, the second ontology including one or more second concepts associated with the second API, one or more second attributes associated with the one or more second concepts, and one or more second taxonomical relationships between the one or more second concepts;
• correlating the first ontology with the second ontology based on the one or more first concepts of the first ontology and the one or more second concepts of the second ontology;
• based on the first ontology, the second ontology, and the correlating the first ontology with the second ontology, generating a knowledge graph of the first API and the second API;
• obtaining a natural language user query provided by a user;
• converting the natural language user query to a semantic user query using natural language understanding;
• searching the knowledge graph using the semantic user query to identify an API call as a response to the natural language user query;
• executing the API call using the knowledge graph and the natural language user query to provide a result for the natural language user query;
• and
• providing the result to the user.