DataBrowser#

Interactive Jupyter widget for browsing and loading datasets from Hugging Face Hub.

Notebook | Link |

|----------|——| | Demo | notebooks/browser.ipynb | | Open in Colab | |

class quantem.data.DataBrowser(**kwargs: Any)[source]#

Bases: AnyWidget

Interactive browser for quantem.data datasets on HF Hub.

Displays a filterable table of available datasets. Click a dataset to see its metadata, then click Load to download it into memory.

Parameters:: technique (str, optional) – Initial technique filter (e.g. "4dstem", "hrtem").

Examples

>>> from quantem.data import DataBrowser
>>> browser = DataBrowser()
>>> browser  # displays widget in notebook
>>> # After selecting and loading a dataset in the UI:
>>> browser.data.shape
(256, 256)

catalog_json#: A trait for unicode strings.

selected_name#: A trait for unicode strings.

selected_info_json#: A trait for unicode strings.

loading#: A boolean (True, False) trait.

loaded_name#: A trait for unicode strings.

__init__(technique: str | None = None, **kwargs)[source]#: Public constructor

technique_filter#: A trait for unicode strings.

refresh()[source]#: Re-fetch the dataset catalog from HF Hub.

property data: ndarray | None#: The loaded dataset as a NumPy array, or None if nothing is loaded.

property metadata: dict | None#: Metadata dict for the loaded dataset, or None.

property techniques: list[str]#: Valid technique names.

summary()[source]#: Print a human-readable summary of the browser state.

__repr__() → str[source]#: Return a simple repr to avoid expensive ipywidgets trait serialization.