Automated API Naming via Intent Analysis and Entity Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Managing a multitude of APIs in complex IT systems is challenging due to vague, ambiguous, or misleading API names, which hinder effective communication and lead to redundant efforts and reduced API re-usability.
Innovation Solution
Automatically generating standardized API names that are intuitive and accurately convey API functionality, using techniques such as named entity recognition and machine learning-based intent analysis, and enabling automatic publishing, validation, and bulk repair of API names.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If manual API naming is used, then developers can create custom names, but the names become vague, ambiguous, or misleading and inconsistent across APIs
Solution Approach 1:
The system enables self-service by automatically generating API names using named entity recognition and intent analysis on API specification documents. The automated naming service extracts entities and determines intent from specification text, assembling standardized names without requiring manual developer intervention, thus maintaining consistency while easing API creation.
Solution Approach 2:
The system changes the parameters of the naming process by transforming unstructured specification text into structured named entities and intent classifications. This parameter transformation enables consistent naming by converting variable developer input into standardized name components following the pattern {intent} {named entity} API.
2Manufacturing precision
If standardized automated naming is implemented, then naming consistency and clarity improve, but the complexity of the naming system increases
Solution Approach 1:
The patent introduces an intermediary automated naming service that mediates between API specifications and final names. This service uses named entity recognition and intent analysis as intermediate processing steps, transforming raw specification text into structured components before assembling the final standardized name, thereby managing complexity through modular intermediate stages.
Solution Approach 2:
The naming system is segmented into distinct functional components: named entity recognition module, intent analysis module, and name assembly module. Each component handles a specific aspect of the naming process, allowing the system to manage complexity through modular design while maintaining overall consistency through standardized assembly rules.
3Stability of the object's composition
If legacy API names are not repaired, then existing systems remain unchanged, but user experience deteriorates and search effectiveness decreases
Solution Approach 1:
The system performs preliminary action by proactively identifying and repairing legacy API names before they cause significant user experience degradation. The automated naming service analyzes existing API names, detects inconsistencies or ambiguities, and repairs them using the same named entity recognition and intent analysis processes, preventing future search effectiveness issues while maintaining system stability through gradual updates.
Data Source
AI summary
Techniques are disclosed for application programming interface management in information processing systems. For example, a method comprises obtaining one or more data sets descriptive of at least one application programming interface, and automatically generating a name for the at least one application programming interface, wherein the name comprises a naming format applied across a plurality of application programming interfaces searchable by names.