Implementing the Dhikr Resource

After setting up the artists resource, we shifted focus to a new and more complex one: Dhikr, which is heavily dependent on audio files.

Schema Setup #

We started by defining the Mongoose schema for dhikr documents in MongoDB:

const mongoose = require("mongoose");

const dhikrSchema = new mongoose.Schema(
  {
    title: { type: String, required: true },
    description: { type: String },
    artist: { type: String },

    listens: { type: Number, default: 0 },
    likes: { type: Number, default: 0 },
    likedBy: { type: [String] },
    tags: { type: [String] },
    duration: { type: Number },

    status: {
      type: String,
      enum: ["processing","failed","pending","verified"],
      default: "processing",
    },

    uploadedType: {
      type: String,
      enum: ["link","audio","video"],
    },

    link: { type: String },
  },
  { timestamps: true }
);

module.exports = mongoose.model("Dhikr", dhikrSchema);

This schema captures not only metadata (title, description, artist, tags, etc.) but also the file type, status, and interaction metrics (likes, listens).

Controller + Service Setup #

We created a controller file and a service file for Dhikr.
All business logic lives in the service layer, while the controller just coordinates requests and responses. This code splitting makes the project easier to maintain and extend.

createDhikr Endpoint #

Our first endpoint was createDhikr. Here’s how we designed it:

• It accepts multipart/form-data with both metadata and an audio file.
• Because audio files can be large, we used Node.js worker threads to offload file processing to a separate worker.
• We accounted for three upload types:
  • Audio → directly uploaded to Cloudflare R2
  • Video → processed with ffmpeg to extract audio
  • Link → processed with yt-dlp to download audio from YouTube (or other supported platforms)

Once processed, the resulting audio file is uploaded to Cloudflare R2 for storage.

Swapping Audio Files #

We also added an endpoint to swap out old audio files:

• The old file is deleted from Cloudflare R2
• A worker thread uploads the new file
• The MongoDB document is updated with the new file reference

Current Status #

At this stage, the Dhikr resource supports:

• Schema for saving dhikr documents in MongoDB
• A createDhikr endpoint that handles audio, video, and links
• Worker-based processing for large file uploads
• R2 storage integration for finalized audio files
• An endpoint to replace existing audio files safely

This is a big milestone for the Sufii API, as Dhikr will be one of the core resources in the ecosystem. 🎧