data-science template

odf init my-project --template data-science

default (Item/Items/SQLite), no app/ changes at all — just a seed CSV and a notebook that loads it into the repository outside the running Project, to show a Repository works standalone in a notebook kernel the same way it does DI-resolved inside one.

Layout

my-project/
├── config.toml
├── app/
│   ├── __init__.py
│   ├── entities.py       # Item
│   ├── repositories.py   # Items(SQLite)
│   └── storages.py       # SQLite
├── data/
│   ├── raw/
│   │   └── items.csv      # source data — read, never edited in place
│   └── processed/         # starts empty
└── notebooks/
    └── example.ipynb

What's added

example.ipynb builds the same SQLite → Items dependency chain the Context would normally wire up for you, but by hand — since a notebook kernel runs outside the resolved object graph:

from opendataframework import Config

from app.entities import Item
from app.repositories import Items
from app.storages import SQLite

config = Config({"component": {"sqlite": {"path": "../app.db"}}})
sqlite = SQLite(config)
items = Items(sqlite)

Then it reads data/raw/items.csv and calls items.save(...) per row — the same ingestion a real Task would do, just exploratory. Anything derived (cleaned, joined, aggregated) belongs in data/processed/, never written back to raw/.

Run it

jupyter notebook notebooks/example.ipynb

Or start the UI as usual (odf run) to browse Items once the notebook has seeded it.