@uploadflow/mongoose

Zero-setup Mongoose AttachmentStore for the uploadflow file-upload pipeline, plus a lookupStages() aggregation $lookup helper (no N+1). Use it so @LinkedUpload / @WithAttachments (NestJS) or linked() (Express) persist attachments to MongoDB without you writing any DB code.

npm i @uploadflow/mongoose @uploadflow/core

Peer dep: mongoose (^7 or ^8).

Setup

The store registers its own schema/collection — no model to define.

import { MongooseAttachmentStore } from '@uploadflow/mongoose';

const store = MongooseAttachmentStore.forRoot({
  connection,               // a mongoose Connection
  collection: 'attachments' // optional (default 'attachments')
});

NestJS (inject the connection, e.g. via getConnectionToken()):

UploadModule.forRootAsync({
  inject: [getConnectionToken()],
  useFactory: (connection) => ({
    storage: { /* ... */ },
    attachmentStore: MongooseAttachmentStore.forRoot({ connection }),
  }),
});

Stored schema

Each attachment document: recordId, module, key, url, size, mime, originalName, createdBy, createdAt. recordId + module are indexed.

Reading in an aggregation (no N+1)

Drop the join into your existing pipeline:

const docs = await MediaModel.aggregate([
  { $match: { companyId } },
  ...MongooseAttachmentStore.lookupStages({ module: 'MEDIA', as: 'attachments' }),
  { $sort: { createdAt: -1 } },
]);

lookupStages({ localField?, foreignField?, module?, as?, collection? }) returns the $lookup stages; it casts the local record id to a string to match the stored recordId. Also available as an instance method: store.lookupStages({ ... }) (uses the store's collection).

Under urlMode: 'signed', stamp fresh signed URLs on the joined array afterwards:

await storage.resolveDownloadUrls(docs, { path: 'attachments[].key' });

Full docs: https://github.com/OWNER/uploadflow#readme

MIT