soilspecdata 0.0.3

Download and load soil spectral data

SoilSpecData

A Python package for handling soil spectroscopy data, with a focus on the Open Soil Spectral Library (OSSL).

Installation

pip install soilspecdata

If you want to install the development version, run in the project root:

pip install -e .[dev]

Features

• Easy loading and handling of OSSL dataset
• Support for both VISNIR (Visible Near-Infrared) and MIR (Mid-Infrared) spectral data
• Flexible wavelength range filtering
• Convenient access to soil properties and metadata
• Automatic caching of downloaded data
• Get aligned spectra and target variable(s)
• Further datasets to come …

Quick Start

# Import the package
from soilspecdata.datasets.ossl import get_ossl

Load the OSSL dataset:

ossl = get_ossl()

• Get MIR spectra (600-4000 cm⁻¹):

mir_data = ossl.get_mir(require_valid=True)

• Get VISNIR spectra with custom wavelength range:

visnir_data = ossl.get_visnir(wmin=500, wmax=1000, require_valid=True)

• Get soil properties (e.g., CEC):

properties = ossl.get_properties(['cec_usda.a723_cmolc.kg'], require_complete=True)

For more details on the OSSL dataset and its variables, see the OSSL documentation.

• Get metadata (e.g., geographical coordinates):

metadata = ossl.get_properties(['longitude.point_wgs84_dd', 'latitude.point_wgs84_dd'], require_complete=False)

• Or to get directly aligned spectra and target variable(s):

X, y, ids = ossl.get_aligned_data(
    spectra_data=mir_data,
    target_cols='cec_usda.a723_cmolc.kg'
)

X.shape, y.shape, ids.shape

((57062, 1701), (57062, 1), (57062,))

• Plot the first 20 MIR spectra:

from matplotlib import pyplot as plt

plt.figure(figsize=(12, 3))
plt.plot(mir_data.wavenumbers, mir_data.spectra[:20,:].T, alpha=0.3, color='steelblue', lw=1)
plt.gca().invert_xaxis()
plt.grid(True, linestyle='--', alpha=0.7)

plt.xlabel('Wavenumber (cm⁻¹)')
plt.ylabel('Absorbance');

Data Structure

The package returns spectra data in a structured format containing:

• Wavenumbers
• Spectra measurements
• Measurement type (reflectance/absorbance)
• Sample IDs

Properties and metadata are returned as pandas DataFrames indexed by sample ID.

Cache Management

By default, the OSSL dataset is cached in ~/.soilspecdata/. To force a fresh download:

ossl = get_ossl(force_download=True)

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

Apache2

Citation

TBC