Auto-documentation for application program interfaces based on network requests and responses

• US 10,225,330 B2
• Filed: 05/08/2018
• Issued: 03/05/2019
• Est. Priority Date: 07/28/2017
• Status: Active Grant

• Current Assignee: Kong Inc.
• Original Assignee: Kong Inc.
• Inventors: Palladino, Marco, Marietti, Augusto
• Primary Examiner(s): Joo, Joshua
• Application Number:
  US15/974,532
• Publication Number:
  US 20190037005A1
• Time in Patent Office: 301 Days
• Field of Search: None
• CPC Class Codes: G06F 8/73 Program documentation
• CPC Class Codes: H04L 67/01 Protocols
• CPC Class Codes: H04L 67/02 based on web technology, e.g. hypertext transfer protocol [HTTP]
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46
• CPC Class Codes: H04L 67/1004 Server selection for load balancing
• CPC Class Codes: H04L 67/563 Data redirection of data network streams
• CPC Class Codes: H04L 67/568 Storing data temporarily at an intermediate stage, e.g. caching

Abstract

Disclosed embodiments are directed at systems, methods, and architecture for providing auto-documentation to APIs. The auto documentation plugin is architecturally placed between an API and a client thereof and parses API requests and responses in order to generate auto-documentation. In some embodiments, the auto-documentation plugin is used to update preexisting documentation after updates. In some embodiments, the auto-documentation plugin accesses an online documentation repository. In some embodiments, the auto-documentation plugin makes use of a machine learning model to determine how and which portions of an existing documentation file to update.

Claims

1. A method comprising:

receiving, at a network node, one or more network packets containing an API request or API response associated with a first API;

proxying, by the network node, the API request or response;

parsing, by the network node, the API request or response;

determining, by the network node, a first behavioral indicator based on content of the API request or response as observed from said parsing;

determining, by the network node, that the first behavioral indicator is inconsistent with documentation for the first API based on a variance from a machine learning model; and

automatically updating, by the network node, the documentation for the first API based on the behavioral indicator.

2. The method of claim 1, wherein the behavioral indicator is any of:

a title of a parameter;

a value provided for the parameter;

a sequence of requests and responses;

a host or endpoint identified in the API request or response;

or

a protocol or language used.

3. The method of claim 1, wherein the determination based on the machine learning model is further based on an observational history of a plurality of APIs.

4. The method of claim 1, wherein the determination based on the machine learning model is further based on an observational history of a plurality of APIs having a sequence of API requests and responses.

5. The method of claim 1, further comprising:

updating the machine learning model based on the content of the API request or response as observed from said parsing.

6. The method of claim 1, wherein said determining inconsistency between the behavioral indicator and the documentation further comprises:

identifying that no documentation file exists for the first API; and

generating a documentation file for the first API.

7. The method of claim 1, wherein said determining inconsistency between the behavioral indicator and the documentation is further based on a semantic analysis of the documentation and the content of the API request or response as observed from said parsing.

8. A method comprising:

receiving, at a gateway, an API request and an API response, wherein the API response corresponds to the API request, the API request and the API response associated with a first API;

determining, by the gateway, parameters of the API request and values of the API response that correspond to the parameters;

determining, by auto-documentation code, that any of the parameters or values are inconsistent with documentation for the first API based on a variance from a machine learning model; and

automatically updating by the auto-documentation code, the documentation for the first API based on any of the parameters or values are inconsistent with documentation for the first API.

9. The method of claim 8, further comprising:

determining, by auto-documentation code, that any factors of;

a sequence of requests and responses including the API request and the API response;

a host or endpoint identified in the API request and the API response;

or

a protocol or language used by the API request and an API response;

are inconsistent with documentation for the first API; and

automatically updating by the auto-documentation code, the documentation for the first API based on any of the factors that are inconsistent with documentation for the first API.

10. The method of claim 8, wherein the determination based on the machine learning model is further based on an observational history of a plurality of APIs.

11. The method of claim 8, wherein the determination based on the machine learning model is further based on an observational history of a plurality of APIs having a sequence of API requests and responses.

12. The method of claim 8, further comprising:

updating the machine learning model based on the any of the parameters or values as observed.

13. The method of claim 8, wherein said determining inconsistency between any of the parameters or values and the documentation further comprises:

identifying that no documentation file exists for the first API; and

generating a documentation file for the first API based on any of the parameters or values.

14. The method of claim 8, wherein said determining inconsistency between the any of the parameters or values and the documentation is further based on a semantic analysis of the documentation and any of the parameters or values as observed.

15. A system comprising:

a network node within a network configured to receive one or more network packets containing an API request or API response associated with a first API, the network node further configured to proxy the API request or API response; and

memory containing auto-documentation code that when executed is configured to parse the API request or response and determine a first behavioral indicator based on content of the API request or response as parsed, and wherein the executed auto-documentation code is further configured to determine whether the parsed behavioral indicator is inconsistent with documentation for the first API based on a variance from a machine learning model within the memory, and when inconsistency is determined, automatically update the documentation for the first API based on the behavioral indicator.

16. The system of claim 15, wherein the behavioral indicator is any of:

a title of a parameter;

a value provided for the parameter;

a sequence of requests and responses;

a host or endpoint identified in the API request or response;

or

a protocol or language used.

17. The system of claim 15, wherein the determination based on the machine learning model is further based on an observational history of a plurality of APIs.

18. The system of claim 15, wherein the determination based on the machine learning model is further based on an observational history of a plurality of APIs having a sequence of API requests and responses.

19. The system of claim 15, wherein the executed auto-documentation code is further configured to update the machine learning model based on the content of the API request or response as observed.

20. The system of claim 15, wherein the determination that the behavioral indicator and the documentation are inconsistent includes:

an identification that no documentation file exists for the first API, and wherein the auto-documentation code is further configured to generate a documentation file for the first API.

21. The system of claim 15, wherein the auto-documentation code determines the behavioral indicator and the documentation are inconsistent further based on a semantic analysis of the documentation and the content of the API request or response as observed.