bsccm-i2i

Minimal bootstrap package for BSCCM image-to-image experiments.

Dependency model:

• pyproject.toml defines dependency intent.
• requirements*.lock pins exact versions and hashes for reproducible installs.

Install (Reproducible, Recommended)

python3 -m venv .venv
. .venv/bin/activate
python -m pip install pip-tools
pip-sync requirements-dev.lock
bsccm-i2i --help

Use this path for normal setup. Do not run pip-compile unless you are updating dependencies. If requirements-dev.lock is not present yet, use the editable install path below.

Install (Editable Dev, Fast Iteration)

python3 -m venv .venv
. .venv/bin/activate
python -m pip install -e ".[dev]"
# Verifies installed console entrypoint wiring
bsccm-i2i --help
# Verifies direct module execution/import path
python -m bsccm_i2i.cli.main --help

Build, Lint, and Test

Run from the repository root:

cd /Users/jignasupathak/Documents/job-search-2026-lol/bsccm-i2i
. .venv/bin/activate

Build a distributable package:

python -m pip install build
python -m build

Lint checks:

ruff check src tests

Lint rules are configured in pyproject.toml under [tool.ruff.lint] and currently include:

• E, F (pycodestyle/pyflakes basics)
• I (import sorting)
• UP (pyupgrade)
• B (bugbear)

If lint fails, first try auto-fixes:

ruff check src tests --fix

Run tests:

pytest -q

Configuration (.env)

For dataset auto-download via bsccm.download_dataset, Dryad may require a token.

1. Create a local env file:

cp .env.example .env

2. Set your token in .env:

BSCCM_DRYAD_TOKEN=your_token_here

bsccm-i2i loads .env automatically at CLI startup.

Split-first workflow (required before train):

bsccm-i2i split
# Copy SPLIT_ID from command output

Example smoke run using auto-download:

bsccm-i2i train experiment=baseline_unet split.id=<SPLIT_ID> trainer=smoke data.num_workers=0

Enable datamodule progress logs during setup/loading:

bsccm-i2i train experiment=baseline_unet split.id=<SPLIT_ID> trainer=smoke logging.data_progress=true data.num_workers=0

Example smoke run with an existing local dataset path (no download):

bsccm-i2i train experiment=baseline_unet split.id=<SPLIT_ID> trainer=smoke data.root_dir=/ABS/PATH/TO/BSCCM-tiny data.num_workers=0

Dependency Maintenance (Maintainers Only)

python -m pip install pip-tools
pip-compile pyproject.toml --generate-hashes -o requirements.lock

Run these only when dependencies change in pyproject.toml, then commit updated lock files.