Let me write your networking code
summary
Yelp automates client/server parity by generating Swagger/OpenAPI specs into Kotlin data classes and Retrofit interfaces via a Gradle plugin that runs in CI and publishes Android libraries to an internal Nexus. The pipeline handles ~400 mobile API endpoints and ~100 business endpoints, produces nullable vs required types, enums and operation-id headers, and works in production, but introduces surprises: breaking changes, swapped parameter bugs, ridiculously long generated class names, and missing-response-header access problems.
AI-written summary · about
chapters (11)
- 0:00Introduction and API basics
- 2:58Understanding Swagger specs
- 5:55Defining endpoints and models
- 10:39Swagger UI and editor tools
- 12:25Implementing network requests
- 15:01Introduction to Swagger Codegen
- 18:08Developing the Gradle plugin
- 20:24Code generation walkthrough
- 23:20CI integration and build tasks
- 25:52Scaling and managing infrastructure
- 30:13Challenges and future outlook
related talks
19mWriting an Open API Parser
41mReviewing Kotlin
37mTesting API Resiliency using Kotlin: Leveraging FP at Scale
42mEvolving your Kotlin API painlessly for clients
54m10X Android Development
19mAutomate what you repeat: From boilerplate to automation
05mHow to end manual REST-API client coding
32mHow we drive Developer Productivity at Yelp