Shapir: Standardizing and Democratizing Access to Web APIs

Document Title

Shapir: Standardizing and Democratizing Access to Web APIs

Document Information

• Subject Area: Standardization and Democratization of Web API Access
• Keywords: Web APIs, API Description Languages, Schema.org, Semantic Web, API Integration, Programming Interfaces, Data Interoperability, Mavo, ShapirJS, WoOPI

Research Background and Problem Statement

• Problems or Challenges:
  1. The syntax and data models of Web APIs across different websites are not standardized, requiring users to manually write complex HTTP requests, handle pagination, authentication, etc.
  2. Developers spend significant time reading documentation and learning each API, posing high difficulty for non-professional users.
  3. The lack of a unified standardized model for APIs limits cross-website data integration and reuse.
• Significance:
  • Standardizing and simplifying API access can greatly reduce development workload, enable more users to reuse data across multiple websites, and improve development efficiency.
• Research Motivation and Related Work:
  • Based on Schema.org and existing API description methods (e.g., OpenAPI, Swagger), the authors aim to develop a unified API standard and tools to overcome limitations in describing data models and relationships in current methods.

Solution

• Method or Solution:
  1. Propose the "Web of Objects Programming Interface (WoOPI)":
     • Provide a standardized API description language built on Schema.org, capable of describing objects, attributes, methods, and their relationships managed by APIs.
  2. Develop "ShapirJS":
     • A tool to wrap Web API data into locally operable JavaScript objects, simplifying API calls.
  3. Provide "ShapirUI":
     • A visual tool that allows users to generate WoOPI descriptions for APIs through a graphical interface.
  4. Integration with "Mavo":
     • Enables non-developers to create interactive, API-based data management applications by writing HTML.
• Innovations:
  • Map API data models based on Schema.org standards to achieve unified operations across multiple APIs.
  • Provide a graphical tool (ShapirUI) for creating WoOPI descriptions without programming, lowering technical barriers.
  • Seamlessly integrate with Mavo, allowing non-developers to operate API data using only HTML.
• Implementation Steps and Techniques:
  1. WoOPI generates standardized object-method models by analyzing existing API practices (e.g., CRUD methods).
  2. ShapirJS dynamically generates JavaScript API libraries based on WoOPI models, automatically handling pagination, authentication, and asynchronous operations.
  3. ShapirUI combines semantic analysis and intelligent matching to assist users in mapping site APIs to Schema.org types.
  4. Integration with Mavo supports simple HTML-based API operations through standard attributes and methods.

Research Outcomes

• Specific Outcomes:
  1. Designed the WoOPI description language, capable of describing approximately 90% of existing Web APIs.
  2. Developed the ShapirJS library and ShapirUI tool, successfully integrating them with Mavo.
  3. ShapirJS supports unified management of cross-site data using standard object types (e.g., VideoObject) while automatically completing necessary API calls.
• Advantages:
  1. ShapirJS accelerates multi-API tasks by approximately 5.6 times compared to Swagger.
  2. Non-programmers using ShapirJS and Mavo can complete data management applications within minutes, lowering technical barriers.
  3. Modular design of the toolchain supports future replacement and expansion.
• Experiments and Evaluation:
  • Developers using ShapirJS for API tasks completed them significantly faster than traditional methods (e.g., Swagger).
  • Non-programmers took only 4 minutes to build cross-API applications using Mavo and ShapirUI.
  • User studies showed high usability, enabling users without technical backgrounds to quickly participate.
• Limitations and Future Directions:
  1. Schema.org's attribute representation is overly broad (inconsistencies in singular/plural forms and types), requiring stronger validation mechanisms in the future.
  2. Currently, only API data reading capabilities are implemented; future work should expand to support write operations.
  3. Mapping website APIs to Schema.org standard types is challenging; exploring automation or community collaboration could improve consistency.
  4. Limitations in cross-website data operations (e.g., inability to store data from one site on another).

Through this research, Shapir and its ecosystem bring transformative simplification to Web API data access and integration, while promoting the creation and practice of standardized interoperable tools.