The back-end API for ITEMS marketplace app.

Quick Notes

Useful Commands

Stack

• Python FastAPI
• PostrgreSQL
• Poetry
• Alembic (SQLAlchemy)
• Docker (or Kuberneter)
• pydantic
• S3 storage

First time setup

...among many things

• Install nbstripout for jupyter notebook commit output stripping.
• Copy .env.example to .env and fill out values.

Run

Build and Run:

make run

Visit: http://localhost:8000/docs

Database: localhost:5432

Recommendation Feed (Food TikTok)

The meal recommendation system combines user swipes with high-rated reviews to build a personalized queue.
Key pieces you might need:

• Seed local data (10 users × 5 reviews each + likes):

  poetry run python admin/seed_recommendation_data.py

• Notebook demo: admin/recommendation_test.ipynb (run all cells to see top picks for a sample user).

• API: GET /users/me/feed?limit=20 returns the pre-ranked queue for the authenticated user.

• Queue regeneration: if a user’s user_feed_item table is empty, the backend automatically recomputes their feed using the latest reviews/ratings.

Use this section as the canonical reference when hooking up the frontend swipe/scroll experience.

Tests

Unit tests

Run all tests while the api is running:

make test

Poetry guide

Add package:

poetry add <name>

Install from pyproject.toml or poetry.lock:

poetry install

Code quality

A collection of general rules and guidelines for code style and linting.