ResolveDataset

The ResolveDataset class handles data loading and preprocessing for RESOLVE.

Class Definition

class ResolveDataset:
    """
    Dataset container for RESOLVE models.

    Handles loading, validation, and preprocessing of plot-level
    and species composition data.
    """

Constructor

ResolveDataset(header, species, roles, targets)

Create a dataset from pandas DataFrames.

Parameters:

Parameter	Type	Description
header	pd.DataFrame	Plot-level data (one row per plot)
species	pd.DataFrame	Species occurrences (one row per species-plot)
roles	dict	Column name mappings
targets	dict	Target variable configurations

Example:

dataset = ResolveDataset(
    header=header_df,
    species=species_df,
    roles=roles,
    targets=targets,
)

Class Methods

from_csv(header, species, roles, targets)

Load dataset from CSV files.

Parameters:

Parameter	Type	Description
header	str \| Path	Path to header CSV file
species	str \| Path	Path to species CSV file
roles	dict	Column name mappings
targets	dict	Target variable configurations

Returns: ResolveDataset

Example:

dataset = ResolveDataset.from_csv(
    header="data/plots.csv",
    species="data/species.csv",
    roles=roles,
    targets=targets,
)

Properties

schema

Returns the ResolveSchema describing the dataset structure.

schema = dataset.schema
print(f"Plots: {schema.n_plots}")
print(f"Species: {schema.n_species}")

plot_ids

Returns array of plot identifiers.

plot_ids = dataset.plot_ids

Methods

get_coordinates()

Returns coordinates array or None if not available.

coords = dataset.get_coordinates()
# Shape: (n_plots, 2) or None

get_covariates()

Returns covariates array or None if not available.

covariates = dataset.get_covariates()
# Shape: (n_plots, n_covariates) or None

get_targets()

Returns dictionary of target arrays.

targets = dataset.get_targets()
# {"area": array(...), "habitat": array(...)}

---

ResolveSchema

The ResolveSchema dataclass describes dataset structure.

Attributes

Attribute	Type	Description
n_plots	int	Number of plots
n_species	int	Number of unique species
n_genera	int	Number of unique genera
n_families	int	Number of unique families
covariate_names	list[str]	Names of covariate columns
targets	dict	Target configurations
has_coordinates	bool	Whether coordinates are available
has_taxonomy	bool	Whether taxonomy is available