aind-smartspim-classification

Code for classifying the cell candidate outputs from aind-smartspim-segmentation within the SmartSPIM pipeline. It uses the aind-large-scale-prediction package to efficiently process large amounts of image data.

This repository takes as input the cell proposals that will be classified by the CellFinder model. The output is a CSV with the following columns:

• Cell Counts: Number of positive cells.
• Cell Likelihood Mean: Mean of the probabilities of a cell being a cell.
• Cell Likelihood STD: Standard deviation of the probability of a cell being a cell.
• Noncell Counts: Number of negative cells.
• Noncell Likelihood Mean: Mean of the probabilities of negative cells.
• Noncell Likelihood STD: Standard deviation of the probabilities of negative cells.

---

Prerequisites

• Python 3.9
• A CUDA-capable GPU (the pipeline is designed to run on GPU; see environment/Dockerfile for the full hardware/software stack)
• Conda — the production environment uses the cell_class conda environment defined in environment/environment.yml

For local development outside of Code Ocean, a CUDA GPU is not strictly required to run tests, but is needed to run the full pipeline.

---

Installation

From the repository root, install the package and its dependencies:

pip install -e .

To install development tools (linters, test runner, coverage):

pip install -e ".[dev]"

Note: The Python package (aind_smartspim_classification) lives inside code/. The pyproject.toml at the root configures setuptools to find it there automatically.

---

Running in Code Ocean

The capsule entry point is code/run. It activates the cell_class conda environment, sets KERAS_BACKEND=torch, and executes code/run_capsule.py. Inputs are mounted at /data and outputs are written to /results.

For SLURM-based execution, use code/run_slurm instead.

---

Contributing

Linters and testing

After installing with pip install -e ".[dev]", the following tools are available:

• Run the test suite with coverage:

pytest code/tests/ --cov=aind_smartspim_classification --cov-report=term-missing

• Check documentation coverage:

interrogate .

• Check code style:

flake8 code/

• Auto-format code:

black code/

• Sort imports:

isort code/

Pull requests

For internal members, please create a branch. For external members, please fork the repository and open a pull request from the fork. We primarily use Angular style for commit messages:

<type>(<scope>): <short summary>

where scope (optional) describes the packages affected and type (mandatory) is one of:

• build: Changes that affect build tools or external dependencies (example scopes: pyproject.toml)
• ci: Changes to CI configuration files and scripts
• docs: Documentation only changes
• feat: A new feature
• fix: A bugfix
• perf: A code change that improves performance
• refactor: A code change that neither fixes a bug nor adds a feature
• test: Adding missing tests or correcting existing tests

Documentation

To generate RST source files for Sphinx documentation, run from the repo root:

sphinx-apidoc -o doc_template/source/ code/

Then build HTML:

sphinx-build -b html doc_template/source/ doc_template/build/html

More info on Sphinx installation can be found here.