Skip to content

Frequently Asked Questions (FAQ)

1. Why are there two different ElasticSearch clients, one in the scraper and one in the UI?

The scraper and the UI each interact with ElasticSearch, but they serve different roles:

  • Scraper ES Client (Python):
  • Inserts new documents into the image_metadata index
  • Writes confidence scores, descriptions, and image URLs
  • Creates the index if it doesn't already exist

  • Used during offline batch processing

  • UI ES Client (Node.js):

  • Reads documents already indexed by the scraper
  • Applies filters (manufacturer, SKU prefix, confidence ranges)
  • Updates review status fields (pending-approve, pending-reject, approved, rejected)
  • Never downloads or processes images

Because these operate in different languages (Python vs. Node) and have different responsibilities (write-heavy vs. read/update-heavy), each uses its own official Elastic client library.

2. What does the Prisma Adapter do in NextAuth, and why do we need it?

NextAuth uses adapters to store user account and session data in a database.
The Prisma Adapter tells NextAuth to store this data using Prisma models instead of an in-memory store or JSON files.

The adapter handles:

  • Creating users
  • Linking OAuth accounts (if used)
  • Storing hashed passwords (if using credentials provider)
  • Saving and restoring user sessions
  • Managing verification tokens (if email login is used)
  • Mapping NextAuth’s schema onto schema.prisma

Without the Prisma Adapter:

  • User accounts would not persist
  • Sessions would disappear after server restarts
  • Admin/reviewer roles could not be stored

The Prisma Adapter ensures authentication integrates cleanly with Postgres/SQLite.

3. What is the difference between prisma generate and prisma migrate dev?

These commands serve totally different purposes:

npx prisma generate

  • Regenerates the Prisma Client (@prisma/client)
  • Required whenever the schema changes, even without a DB migration
  • Does NOT modify your database
  • Used by the UI to access the database with updated TypeScript types

You run it after: - Pulling from Git - Updating schema.prisma - Installing dependencies on a new machine

npx prisma migrate dev

  • Applies schema changes to the actual database
  • Creates a new migration folder under /prisma/migrations
  • Updates the database to match the schema
  • Only used in development (not production)

You run it after: - Adding/removing fields in the Prisma schema
- Creating new tables
- Changing data types

If you skip migrations, the database schema and Prisma Client will become inconsistent.