Accessibility (A11y) in Clean RESTful APIs

Discover how to combine the concepts of Clean Architecture with the Speech-To-Text (STT) concept to promote accessibility for audible educational content and harness the power of Java, Spring Boot, and MongoDB in building a RESTful API focused on accurate audio transcription.

Presentation slides available here: DIO Acceleration: Accessibility in Clean RESTful APIs

Hands-On Overview

The Audio2Text project is an advanced system designed to transcribe audio content into text format, leveraging the power of OpenAI's API for high-quality transcriptions. The project follows Clean Architecture principles to ensure a scalable, maintainable, and well-organized codebase.

Project Structure and References

This project's architecture is inspired by and references several key sources in Clean Architecture and software design. The structure is visualized using a mermaid graph for clarity.

graph RL;
  subgraph "Infrastructure";
    Web(Web & Devices) <--> Con
    UI("User Interface (UI)") <--> Pre
    DB(Databases & External Integrations) <--> Gat

    subgraph "Adapters";
      Con(Controllers) <--> UC
      Pre(Presenters) <--> UC
      Gat(Gateways) -..-> |implements| IGat

      subgraph "Use Cases";
        UC(Use Cases) <--> LA
        UC <--> IGat

        subgraph "Entities";
          LA("Learning Objects (LOs)")
          LA -.- LOA[<em>Metadata includes:\n Transcript, Language,\n Revision, License etc.</em>]
          IGat(Gateway Interfaces)
          IGat -.- IRep(<em>Any external integrations:\n Database, WebClient,\n Producer etc.</em>)
        end
     end
    end
  end

classDef infra fill:#a3c9ff,stroke:#00315c,color:#00315c;
classDef adapters fill:#67dbb1,stroke:#003828,color:#003828;
classDef ucs fill:#ffb1c1,stroke:#5f112b,color:#5f112b;
classDef entities fill:#e2c54b,stroke:#3a3000,color:#3a3000;
classDef entities_secondary fill:#fff0c0,stroke:#3a3000,color:#3a3000;

class Web,Dev,UI,DB,EXT infra;
class Con,Gat,Pre adapters;
class UC ucs;
class LA,IGat entities;
class LOA,IRep entities_secondary;

Loading

Key References and Inspirations:

• Speech2Learning: PhD project by @falvojr, which explores speech-to-text services to improve the accessibility of learning objects.
• The Clean Architecture Blog Post: Robert C. Martin's (Uncle Bob) seminal post on Clean Architecture principles.
• Clean Architecture for Spring Application: A comprehensive guide by LuanPSantos on implementing Clean Architecture in Spring applications.

Development Demands (TODOs)

Our development team has outlined several key TODOs for further implementation and integration:

Environment Setup

TODO 1. Check/Create the mandatory environment variables!

Code built hands-on:

acceleration-a11y-clean-restful-apis/src/main/resources/application.yml

Line 1 in adf2025

# DONE! 1. Check/Create the mandatory environment variables!

Integration with OpenAI API to High-Quality Transcription

OpenAI API Reference: https://platform.openai.com/docs/api-reference/audio/createTranscription

TODO 2. Respecting the TranscriptionService contract, consumes POST "audio/transcription" on OpenAI API!

Code built hands-on:

acceleration-a11y-clean-restful-apis/src/main/java/com/falvojr/audio2text/domain/transcribedaudio/adapter/gateway/TranscriptionServiceImpl.java

Line 46 in 56c773e

// DONE! 2. Respecting the TranscriptionService contract, consumes "audio/transcription" on OpenAI API!

TODO 3. Orchestrate the integration with transcription service.

Code built hands-on:

acceleration-a11y-clean-restful-apis/src/main/java/com/falvojr/audio2text/domain/transcribedaudio/usecase/CreateTranscribedAudioUC.java

Line 38 in a826c1e

// DONE! 3. Orchestrate the integration with transcription service.

Accessibility Showcase

VLibras Widget Reference: https://vlibras.gov.br/doc/widget/installation/webpageintegration.html

TODO 4. Using the JS Fetch API, consumes GET "/transcribed-audios/{id}" locally and show the transcript.

Code built hands-on:

acceleration-a11y-clean-restful-apis/src/main/resources/static/libras.html

Line 24 in fd9f73c

// DONE! 4. Using the JS Fetch API, consume GET "/transcribed-audios/{id}" and display transcript.

TODO 5. Integrate with VLibras: https://vlibras.gov.br/doc/widget

Code built hands-on:

acceleration-a11y-clean-restful-apis/src/main/resources/static/libras.html

Line 43 in fd9f73c

<!-- DONE! 5. Integrate with VLibras: https://vlibras.gov.br/doc/widget -->