METHODS FOR CODE GENERATION FROM SEMANTIC MODELS AND RULES

• US 20120278788A1
• Filed: 04/29/2011
• Published: 11/01/2012
• Est. Priority Date: 04/29/2011
• Status: Abandoned Application

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Crapo, Andrew Walter
• Application Number:
  US13/097,377
• Publication Number:
  US 20120278788A1
• Time in Patent Office: Days
• US Class Current: 717/106
• CPC Class Codes: G06F 8/51 Source to source

Abstract

Computerized methods for translating semantic models and rules into procedural code are described. The methods identify variables present in triples in both the rule conclusion and premise along with properties of the triples that identify a procedural function to incorporate procedural code generated in lieu of the semantic rule. The method further determines if the premise triples comprising variables that are subjects of conclusion triples are “type” statements. Additionally, the domain of the property of the triples in the rule premise and/or conclusion that are not “type” statements but bind the subject variables are also determined. The procedural function is then associated with procedural classes translated from the intersection of semantic classes identified by the premise triples determined to be “type” statements and the classes identified as the domain of the conclusion or premise triples that comprise the subject variables but are not “type” statements. Further, the procedural function is updated in the associated procedural classes with the semantic rule computation determined using premise elements that compute values of object variables in the conclusion triples.

Claims

1. A computerized method for translating semantic models and rules into procedural code in a desired procedural language, comprising:

evaluating a conclusion of a semantic rule to determine information computed by the semantic rule, wherein one or more elements in the semantic rule conclusion comprise one or more triples;

identifying one or more variables in each of the one or more triples for each of the one or more elements in the semantic rule conclusion, wherein the one or more variables are also present in a premise of the semantic rule;

determining a property of each triple in the one or more triples for each of the one or more elements in the semantic rule conclusion for identifying a procedural function where procedural code generated in lieu of the semantic rule will be incorporated;

identifying one or more elements in the semantic rule premise that bind one or more subject variables in the one or more variables, wherein each of the one or more subject variables is a subject of at least one of the one or more triples in the semantic rule conclusion;

determining whether one or more triples in the semantic rule premise comprising one or more of the identified subject variables are “

type”

statements;

determining the domain of the property of each of the one or more triples in the semantic rule premise, the semantic rule conclusion, or a combination thereof, that comprise one or more of the identified subject variables and are not “

type”

statements;

associating the procedural function with one or more procedural classes translated from one or more semantic classes obtained from an intersection of the classes identified by the triples determined to be “

type”

statements and the classes identified as the domain of the triples that comprise the identified subject variables in the semantic rule premise, the semantic rule conclusion, or a combination thereof, and are not “

type”

statements;

determining a computation performed by the semantic rule based on one or more elements in the semantic rule premise that compute a value of one or more object variables in the one or more variables, wherein the each of the one or more object variables is an object of at least one of the one or more triples in the semantic rule conclusion; and

updating the procedural function in associated procedural classes with procedural code generated in lieu of the determined computation to return the computed value.

2. The computerized method of claim 1, further comprising:

translating one or more semantic classes into corresponding classes in a desired procedural language;

translating one or more properties associated with the semantic classes into fields of corresponding procedural classes; and

translating instances associated with the semantic classes into instances of corresponding procedural classes.

3. The computerized method of claim 2, further comprising using the procedural function to process the translated instances of the procedural classes for determining bindings between the one or more variables in the one or more triples in the semantic rule conclusion and the one or more elements in the semantic rule premise.

4. The computerized method of claim 1, wherein the one or more triples define at least a portion of the sematic model and rules in a mathematical directed graph representation, the directed graph comprising semantic class descriptions and the instances associated with the semantic classes as nodes, and the properties associated with the semantic classes as edges of the directed graph representation.

5. The computerized method of claim 1, further comprising translating one or more namespaces corresponding to the one or more semantic classes into correspoding packages in the desired procedural language.

6. A non-transitory computer readable storage medium with an executable program thereon for generating procedural code in a desired computer programming language from semantic models and rules, wherein the program instructs a processing unit to perform the following steps:

evaluating a conclusion of a semantic rule to determine information computed by the semantic rule, wherein one or more elements in the semantic rule conclusion comprise one or more triples;

identifying one or more variables in each of the one or more triples for each of the one or more elements in the semantic rule conclusion, wherein the one or more variables are also present in a premise of the semantic rule;

determining a property of each triple in the one or more triples for each of the one or more elements in the semantic rule conclusion for identifying a procedural function where procedural code generated in lieu of the semantic rule will be incorporated;

identifying one or more elements in the semantic rule premise that bind one or more subject variables in the one or more variables, wherein each of the one or more subject variables is a subject of at least one of the one or more triples in the semantic rule conclusion;

determining whether one or more triples in the semantic rule premise comprising one or more of the identified subject variables are “

type”

statements;

determining the domain of the property of each of the one or more triples in the semantic rule premise, the semantic rule conclusion, or a combination thereof, that comprise one or more of the identified subject variables and are not “

type”

statements;

associating the procedural function with one or more procedural classes translated from one or more semantic classes obtained from an intersection of the classes identified by the triples determined to be “

type”

statements and the classes identified as the domain of the triples that comprise the identified subject variables in the semantic rule premise, the semantic rule conclusion, or a combination thereof, and are not “

type”

statements;

determining a computation performed by the semantic rule based on one or more elements in the semantic rule premise that compute a value of one or more object variables in the one or more variables, wherein the each of the one or more object variables is an object of at least one of the one or more triples in the semantic rule conclusion; and

updating the procedural function in associated procedural classes with procedural code generated in lieu of the determined computation to return the computed value.

7. The non-transitory computer readable storage medium of 6, wherein the program instructs the processing unit to perform the following steps:

translating one or more semantic classes into classes in a desired procedural language;

translating one or more properties corresponding to the semantic classes into fields of corresponding procedural classes; and

translating instances corresponding to the semantic classes into instances of corresponding procedural classes.

8. The non-transitory computer readable storage medium of 6, wherein the program instructs the processing unit to use the procedural function to process the translated instances of the procedural classes for determining bindings between the one or more variables in the triple in the semantic rule conclusion and the one or more elements in the semantic rule premise.

9. The non-transitory computer readable storage medium of 6 wherein the one or more triples define at least a portion of the sematic model and rules in a mathematical directed graph representation, the directed graph comprising semantic class descriptions and the instances associated with the semantic classes as nodes, and the properties associated with the semantic classes as edges of the directed graph representation.

10. The non-transitory computer readable storage medium of 6, wherein the program instructs the processing unit to translate one or more namespaces corresponding to the one or more semantic classes into corresponding packages in the desired procedural language.

11. A system for generating procedural code in a desired computer programming language from semantic models and rules, comprising:

a repository comprising one or more semantic classes, one or more properties corresponding to the semantic classes, one or more instances of the semantic classes, one or more rules corresponding to the semantic classes, or combinations thereof;

a processing unit coupled to the repository, wherein the processing unit;

translates one or more semantic classes into classes in a desired procedural language;

translates one or more properties corresponding to the semantic classes into fields of corresponding procedural classes;

translates instances corresponding to the semantic classes into instances of corresponding procedural classes;

evaluates a conclusion of a semantic rule to determine information computed by the semantic rule, wherein one or more elements in the semantic rule conclusion comprise one or more triples;

identifies one or more variables in each of the one or more triples for each of the one or more elements in the semantic rule conclusion, wherein the one or more variables are also present in a premise of the semantic rule,

determines a property of each triple in the one or more triples for each of the one or more elements in the semantic rule conclusion for identifying a procedural function where procedural code generated in lieu of the semantic rule will be incorporated;

identifies one or more elements in the semantic rule premise that bind one or more subject variables in the one or more variables, wherein each of the one or more subject variables is a subject of at least one of the one or more triples in the semantic rule conclusion;

determines whether one or more triples in the semantic rule premise comprising one or more of the identified subject variables are “

type”

statements;

determines the domain of the property of each of the one or more triples in the semantic rule premise, the semantic rule conclusion, or a combination thereof, that comprise one or more of the identified subject variables and are not “

type”

statements;

associates the procedural function with one or more procedural classes translated from one or more semantic classes obtained from an intersection of the classes identified by the triples determined to be “

type”

statements and the classes identified as the domain of the triples that comprise the identified subject variables in the semantic rule premise, the semantic rule conclusion, or a combination thereof, and are not “

type”

statements;

determines a computation performed by the semantic rule based on one or more elements in the semantic rule premise that compute a value of one or more object variables in the one or more variables, wherein the each of the one or more object variables is an object of at least one of the one or more triples in the semantic rule conclusion; and

updates the procedural function in associated procedural classes with procedural code generated in lieu of the determined computation to return the computed value.

12. The system of claim 11, further comprising one or more input devices, one or more output devices, or combinations thereof, to allow a user to configure translation of the semantic classes, the properties, the instances, the semantic rules, or combinations thereof, to the desired procedural language.

13. The system of claim 11, wherein the processing unit generates an executable application corresponding to the semantic classes, the properties, the instances, the semantic rules, or combinations thereof, in the desired procedural language.

14. A computerized method for translating semantic models and rules into procedural code in a desired procedural language, the method comprising:

translating one or more semantic classes into classes in a desired procedural language;

translating one or more properties corresponding to the semantic classes into fields of corresponding procedural classes;

translating instances corresponding to the semantic classes into instances of corresponding procedural classes;

evaluating a conclusion of a semantic rule to determine information computed by the semantic rule, wherein one or more elements in the semantic rule conclusion comprise one or more triples;

identifying one or more variables in each of the one or more triples for each of the one or more elements in the semantic rule conclusion, wherein the one or more variables are also present in a premise of the semantic rule;

determining a property of each triple in the one or more triples for each of the one or more elements in the semantic rule conclusion for identifying a procedural function where procedural code generated in lieu of the semantic rule will be incorporated;

identifying one or more elements in the semantic rule premise that bind one or more subject variables in the one or more variables, wherein each of the one or more subject variables is a subject of at least one of the one or more triples in the semantic rule conclusion;

determining whether one or more triples in the semantic rule premise comprising one or more of the identified subject variables are “

type”

statements;

determining the domain of the property of each of the one or more triples in the semantic rule premise, the semantic rule conclusion, or a combination thereof, that comprise one or more of the identified subject variables and are not “

type”

statements;

associating the procedural function with one or more procedural classes translated from one or more semantic classes obtained from an intersection of the classes identified by the triples determined to be “

type”

statements and the classes identified as the domain of the triples that comprise the identified subject variables in the semantic rule premise, the semantic rule conclusion, or a combination thereof, and are not “

type”

statements;

determining a computation performed by the semantic rule based on one or more elements in the semantic rule premise that compute a value of one or more object variables in the one or more variables, wherein the each of the one or more object variables is an object of at least one of the one or more triples in the semantic rule conclusion; and

updating the procedural function in associated procedural classes with procedural code generated in lieu of the determined computation to return the computed value.

15. The computerized method of claim 14, further comprising generating an executable application corresponding to the semantic classes, the properties, the instances, the semantic rules, or combinations thereof, in the desired procedural language.