Extracting program features for assisting software development

• US 10,725,748 B2
• Filed: 11/19/2018
• Issued: 07/28/2020
• Est. Priority Date: 11/19/2018
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Narayanan, Srivatsn, Shanmugam, Kesavan, Wilson-Thomas, Mark A., Lim, Vivian Julia, Keech, Jonathan Daniel, Fu, Shengyu
• Primary Examiner(s): Khatri, Anil
• Application Number:
  US16/195,318
• Publication Number:
  US 20200159505A1
• Time in Patent Office: 617 Days
• Field of Search: 717104-110, 717116-120
• CPC Class Codes: G06F 8/20 Software design
• CPC Class Codes: G06F 8/31 Programming languages or programming paradigms
• CPC Class Codes: G06F 8/35 model driven
• CPC Class Codes: G06F 8/36 Software reuse
• CPC Class Codes: G06F 8/436 Semantic checking
• CPC Class Codes: G06F 8/74 Reverse engineering; Extracting design information from source code
• CPC Class Codes: G06F 8/75 Structural analysis for program understanding
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/025 Extracting rules from data

Abstract

Improving the results and process of machine learning service in computer program development. A client'"'"'s codebase is accessed. A set of features are extracted from the client'"'"'s codebase. One or more features from the set of features are then selected. Thereafter, at least one of the selected features is sent to a machine learning service that uses the received feature(s) to build custom model(s) for the client'"'"'s computer system.

Claims

1. A computer system comprising:

one or more processors; and

one or more computer-readable media having stored thereon computer-executable instructions that are structured such that, when executed by the one or more processors, cause the computer system to perform the following;

communicate via a computer network with a client system that has access to a client codebase;

cause the client system to extract a set of features from the client codebase, extracting the set of features comprising;

compiling the client codebase;

building a semantic model based on the compiled client codebase; and

analyzing the semantic model to obtain the set of features;

select one or more features from the set of features based on one or more rules defined by a client associated with the client codebase;

receive at least one of the selected features from the client system via the computer network;

based on the received features, build one or more custom machine learning models that are tailored to the client codebase, such that the one or more custom machine learning models are built without having to have direct access to the client codebase; and

send the one or more custom machine learning models to the client system via the computer network.

2. The computer system of claim 1, wherein the semantic model is built using at least some metadata of the client codebase.

3. The computer system of claim 1, the selection of one or more features comprising:

accessing a set of applicable rules defined by the client associated with the client codebase; and

filtering the set of features based on the set of applicable rules defined by the client to extract the one or more features.

4. The computer system of claim 1, the selection of one or more features comprising:

accessing a set of applicable rules defined by at least one of the machine learning models provided by a machine learning service; and

filtering the set of features based on the set of applicable rules to extract the one or more features.

5. The computer system of claim 1, wherein the at least one of the one or more custom machine learning models is sent from a machine learning service to the client computer system.

6. The computer system of claim 1, wherein the at least one of the one or more custom machine learning models is stored at a machine learning service and made accessible to the client.

7. The computer system of claim 1, the computer system further caused to perform the following:

receive a custom machine learning model built by the machine learning service in response to sending the at least one of the selected features to the machine learning service; and

apply the received custom machine learning model to a client programming environment of the client computer system.

8. The computer system of claim 7, the computer system further caused to:

in response to the applying the received custom machine learning model to the client programming environment, generate a computer-readable prediction of a programming utility of the client programming environment.

9. The computer system of claim 8, the computer system further caused to:

generate a computer-readable suggestion of a number of programming options to the client'"'"'s system based on the prediction of a client'"'"'s programming needs.

10. The computer system of claim 8, the computer system further caused to:

automatically complete a portion of source code based on the prediction of the programming utility.

11. The computer system of claim 7, the computer system further caused to:

in response to the applying the received custom machine learning models to the client'"'"'s programming environment, detect one or more possible code misuses within the client programming environment.

12. The computer system of claim 11, the computer system further caused to:

in response to the detection of possible code misuse, the computer system recommend a correction to the client to correct at least one of the detected one or more possible code misuses.

13. The computer system of claim 11, the computer system further caused to:

in response to the detection of possible code misuse, automatically correct at least one of the detected one or more possible code misuse.

14. The computer system of claim 1, the computer system further caused to:

re-featurize the client codebase in response to an event.

15. The computer system of claim 14, wherein the event is a change of the client codebase.

16. The computer system of claim 14, wherein the event is a change of at least one of the one or more custom machine learning models.

17. A method for extracting data features from a client codebase for creation of custom models in a service, the method comprising:

communicating via a computer network with a client system that has access to a client codebase;

causing the client system to extract a set of features from the client codebase, extracting the set of features comprising;

compiling the client codebase;

building a semantic model based on the compiled client codebase; and

analyzing the semantic model to obtain the set of features;

selecting one or more features from the set of features based on one or more rules defined by a client associated with the client codebase;

receiving at least one of the extracted features from the client system via the computer network;

based on the received features, building one or more custom machine learning models that are tailored to the client codebase, such that the one or more custom machine learning models are built without having to have direct access to the client codebase; and

sending the one or more custom machine learning models to the client system via the computer network.

18. The method of claim 17, wherein the semantic model is built using at least some metadata of the client codebase.

19. A hardware storage medium having stored thereon computer executable instructions that are structured to be executable by one or more processors of a computer system to thereby cause the computer system to:

communicate via a computer network with a client system that has access to a client codebase;

cause the client system to extract a set of features from the client codebase, extracting the set of features comprising;

compiling the client codebase;

building a semantic model based on the compiled client codebase; and

analyzing the semantic model to obtain the set of features;

select one or more features from the set of features based on one or more rules defined by a client associated with the client codebase;

receive at least one of the extracted features from the client system via the computer network;

based on the received features, build one or more custom machine learning models that are tailored to the client codebase, such that the one or more custom machine learning models are built without having to have direct access to the client codebase; and

send the one or more custom machine learning models to the client system via the computer network.

20. The hardware storage medium of claim 19, the selection of one or more features comprising:

accessing a set of applicable rules defined by the client associated with the client codebase; and

filtering the set of features based on the set of applicable rules defined by the client to extract the one or more features.