neurovlm.core.BrainSearchResult

class neurovlm.core.BrainSearchResult(scores, metadata, latents, query_embeddings, retrieval_space, generated_flatmaps=None, masker=None, decoder=None, dataset_image_getter=None)

Container for brain retrieval or generation outputs.

Parameters:

• scores (Tensor | None)

• metadata (DataFrame | None)

• latents (Tensor | None)

• query_embeddings (Tensor)

• retrieval_space (Literal['mse', 'infonce'])

• generated_flatmaps (Tensor | None)

• masker (Any)

• decoder (Any)

• dataset_image_getter (Callable[[str, int], Nifti1Image] | None)

__init__(scores, metadata, latents, query_embeddings, retrieval_space, generated_flatmaps=None, masker=None, decoder=None, dataset_image_getter=None)

Parameters:

• scores (Tensor | None)

• metadata (DataFrame | None)

• latents (Tensor | None)

• query_embeddings (Tensor)

• retrieval_space (Literal['mse', 'infonce'])

• generated_flatmaps (Tensor | None)

• masker (Any)

• decoder (Any)

• dataset_image_getter (Callable[[str, int], Nifti1Image] | None)

Return type:

None

Methods

__init__(scores, metadata, latents, ...[, ...])
dataset_niftis_for(table)	Load dataset-native NIfTI images aligned with rows from top_k output.
niftis_for(table)	Decode NIfTI images aligned with rows from top_k output.
plot([index, query_index, threshold, ...])	Plot one brain image from this result.
plot_topk_item(row[, threshold, ...])	Plot one row selected from results.top_k(...).iloc[...].
plot_topk_row(table[, index, source, ...])	Plot one row from a top_k table using dataset-native or decoded image.
print([k, query_index])	Print top-k brain results.
to_nifti([index])	Convert brain outputs to NIfTI images.
top_k([k, query_index])	Return top-k brain matches.
top_k_niftis([k, query_index])	Return top_k table and its row-aligned NIfTI images.

Attributes

dataset_image_getter
decoder
df	Return a dataframe view of ranked brain retrieval results.
generated_flatmaps
images	Return generated NIfTI images.
masker
scores
metadata
latents
query_embeddings
retrieval_space