evo-objects

GitHub source

The evo-objects package provides both a low-level API client and typed Python classes for working with geoscience objects in Evo.

Typed Objects

The typed objects module provides intuitive Python classes for working with Evo geoscience objects. Instead of dealing with raw API responses, you work with PointSet, Regular3DGrid, Variogram, and other domain-specific types that provide:

• Simple property access (e.g., pointset.num_points, grid.bounding_box)
• to_dataframe() for getting data as pandas DataFrames
• Rich HTML display in Jupyter notebooks (via %load_ext evo.widgets)
• Clickable links to Evo Portal and Viewer

Loading objects

Three convenience functions let you load any typed object by reference, path, or UUID:

from evo.objects.typed import object_from_path, object_from_uuid, object_from_reference

# By file path in the workspace
obj = await object_from_path(manager, "my-folder/assay-data")

# By UUID
obj = await object_from_uuid(manager, "b208a6c9-6881-4b97-b02d-acb5d81299bb")

# By full object reference URL
obj = await object_from_reference(manager, reference_url)

The correct typed class (PointSet, Regular3DGrid, etc.) is selected automatically based on the object's schema.

Class hierarchy

All typed geoscience objects inherit from BaseObject, which provides common properties (name, description, tags) and operations (update, refresh, validate). Objects with spatial data extend BaseSpatialObject, adding bounding box and coordinate reference system support.

Objects — each has a corresponding Data class used for creation:

Object	Data class	Description
PointSet	PointSetData	Point cloud with XYZ locations and attributes
Regular3DGrid	Regular3DGridData	Uniform-cell 3D grid with cell and vertex data
RegularMasked3DGrid	RegularMasked3DGridData	Regular grid with masked (inactive) cells
Tensor3DGrid	Tensor3DGridData	Grid with variable cell sizes per axis
Variogram	VariogramData	Variogram model for geostatistical analysis
BlockModel	—	Proxy to Block Model Service (see below)

Supporting types — see common types (BoundingBox, Rotation, Point3, Size3d, etc.) and attributes (Attributes, Attribute).

BlockModel (via evo-blockmodels)

The BlockModel type is a geoscience object that acts as a proxy to the Block Model Service. When evo-blockmodels is installed (pip install evo-objects[blockmodels]), the full range of block model operations is available directly on the BlockModel object — no need to use the low-level BlockModelAPIClient.

from evo.objects.typed import BlockModel, RegularBlockModelData, Point3, Size3d, Size3i

data = RegularBlockModelData(
    name="My Block Model",
    origin=Point3(x=0, y=0, z=0),
    n_blocks=Size3i(nx=10, ny=10, nz=5),
    block_size=Size3d(dx=2.5, dy=5.0, dz=5.0),
    cell_data=my_dataframe,
)
bm = await BlockModel.create_regular(manager, data)

# Load an existing block model
bm = await object_from_path(manager, "my-folder/block-model")

# Get data as a DataFrame
df = await bm.to_dataframe()

# Add a new attribute
await bm.add_attribute(data_df, "new_attribute", unit="g/t")

# Create and run a report
report = await bm.create_report(spec)
result = await report.run(manager)
df = result.to_dataframe()

After a compute task (e.g., kriging) adds attributes on the server, call refresh() to update the local object:

bm = await bm.refresh()
bm.attributes  # Now shows newly added attributes

For the full evo-blockmodels typed API (RegularBlockModel, Reports, Units), see the evo-blockmodels documentation.