Analysis tools for SLEAP-based plant root phenotyping.

📖 Documentation | 📊 HackMD Trait Reference | 📝 Paper

pip install sleap-roots

uv is a fast Python package manager (10-100x faster than conda):

# Install uv (one-time setup)
curl -LsSf https://astral.sh/uv/install.sh | sh

# Clone and setup
git clone https://github.com/talmolab/sleap-roots && cd sleap-roots
uv sync

# Run commands
uv run pytest tests/
uv run black sleap_roots tests

Why uv? Environment setup in ~30 seconds (vs 10-15 minutes with conda) with reproducible builds via lockfile.

conda env create -f environment.yml
conda activate sleap-roots

Trait pipelines supported:

• DicotPipeline – Primary + lateral roots (e.g. soy, canola)
• YoungerMonocotPipeline – Primary + crown roots (e.g. early rice)
• OlderMonocotPipeline – Crown roots only (e.g. later rice)
• PrimaryRootPipeline – Primary root only
• LateralRootPipeline – Lateral roots only
• MultipleDicotPipeline – Multi-plant dicot setup (batch from a single image)

1. Compute traits for a single plant

import sleap_roots as sr

series = sr.Series.load(
    series_name="919QDUH",
    h5_path="tests/data/canola_7do/919QDUH.h5",
    primary_path="tests/data/canola_7do/919QDUH.primary.slp",
    lateral_path="tests/data/canola_7do/919QDUH.lateral.slp"
)

pipeline = sr.DicotPipeline()
traits = pipeline.compute_plant_traits(series, write_csv=True)

2. Compute traits for a batch

paths = sr.find_all_slp_paths("tests/data/soy_6do")
plants = sr.load_series_from_slps(paths, h5s=True)

pipeline = sr.DicotPipeline()
batch_df = pipeline.compute_batch_traits(plants, write_csv=True)

3. Use a single trait utility

from sleap_roots.lengths import get_root_lengths

pts = series.get_primary_points(frame_idx=0)
lengths = get_root_lengths(pts)

Validate SLEAP predictions before computing traits with an interactive HTML viewer:

sleap-roots viewer predictions/ --output viewer.html

Features keyboard navigation, confidence scores, and multiple output modes. See the Prediction Viewer Guide for details.

Turn sleap-roots-predict per-scan output into a sleap-roots-contracts ResultEnvelope for Bloom write-back:

python -m trait_extractor <input_dir> <output_dir>

The trait_extractor/ service package lives at the repo root and is intentionally not shipped in the sleap-roots wheel. It is also published as a container image, ghcr.io/talmolab/sleap-roots-trait-extractor. See the Trait-Extractor Service guide for docker run usage, the tag scheme, and the ScanMetadata sidecar schema.

Explore tutorials under notebooks/:

cd notebooks
jupyter lab

You can use the test data in tests/data or replace it with your own.

# 1. Clone the repository
git clone https://github.com/talmolab/sleap-roots && cd sleap-roots

# 2. Install dependencies (creates .venv automatically)
uv sync

# 3. Run tests
uv run pytest tests/

# 4. Format code
uv run black sleap_roots tests

# 5. Check docstrings
uv run pydocstyle sleap_roots/

Adding dependencies:

uv add numpy  # Add runtime dependency
uv add --group dev pytest  # Add dev dependency

# 1. Clone the repository
git clone https://github.com/talmolab/sleap-roots && cd sleap-roots

# 2. Create environment
conda env create -f environment.yml
conda activate sleap-roots

# 3. Run tests
pytest tests/

See the latest trait documentation here:

👉 HackMD: sleap-roots Trait Docs

Created by the Talmo Lab and Busch Lab at the Salk Institute, as part of the Harnessing Plants Initiative.

• Elizabeth Berrigan
• Lin Wang
• Andrew O'Connor
• Talmo Pereira

E.M. Berrigan et al., "Fast and Efficient Root Phenotyping via Pose Estimation", Plant Phenomics.
DOI: 10.34133/plantphenomics.0175