Datasets

Datasets are the boundary between science code and layout code.

Core object

WellDataset is the canonical normalized container.

The public API is designed so computed results can be added back into the dataset cleanly instead of forcing users through a file-loader path.

Current dataset operations

• add_curve(...)
• add_array(...)
• add_raster(...)
• add_series(...)
• add_dataframe(...)
• sort_index(...)
• convert_index_unit(...)
• reindex_to(...)
• merge(...)
• rename_channel(...)

Rules for computed channels

A computed channel should provide:

• a numeric reference axis
• an explicit axis unit
• compatible value shape
• a value unit when known

The renderer should not need to infer whether the basis is depth or time, feet or meters, or whether samples align with another channel.

Merge policy

Dataset merges support explicit collision handling:

• error
• replace
• rename
• skip

That makes it practical to keep raw channels and processed channels together without silently overwriting one another.

Example files

• examples/api_dataset_ingest_demo.py
• examples/api_dataset_alignment_demo.py
• examples/api_dataset_merge_demo.py