qbiocode.embeddings package#

Embeddings Module for QBioCode#

This module provides feature embedding and encoding methods for quantum machine learning. It includes functions for computing various embeddings and quantum feature maps.

Available Functions#

get_embeddings: Reduce a train/test pair with any named method – classical (pca, nmf, lle, isomap, spectral, umap, none) or QuVINE graph embeddings (quvine_rwr, quvine_fused, node2vec, …)
pqk: Projected Quantum Kernel embedding
is_transductive: Whether a method sees test features at embed time
check_embedding_name: Validate a method name up front, before doing any work

Available Constants#

SKLEARN_METHODS: the classical method names, always available
QUVINE_HEADLINE_METHODS: the QuVINE names worth trying first
QUVINE_METHODS: every QuVINE name. Listed even without the [quvine] extra – resolving a name is stdlib-only, so discovery works and only running a method raises. Empty only if the QuVINE subpackage itself cannot be imported.

Available Classes#

ConvAutoencoder: Convolutional autoencoder for dimensionality reduction. Imported lazily – it is the only part of this package that needs torch.

Usage#

>>> from qbiocode.embeddings import get_embeddings, pqk
>>> # Reduce a train/test pair -- the scaler-style split is the point: the
>>> # transform is fitted on train only for every inductive method.
>>> X_train_emb, X_test_emb = get_embeddings("pca", X_train, X_test, n_components=2)
>>> # A QuVINE graph embedding, same call shape (needs pip install "qbiocode[quvine]")
>>> X_train_emb, X_test_emb = get_embeddings("quvine_rwr", X_train, X_test, n_components=8)
>>> # Projected Quantum Kernel embedding
>>> X_pqk = pqk(X, n_components=4)

Submodules:

Summary#

__all__ Classes:

ConvAutoencoder

Base class for all neural network modules.

__all__ Functions:

`check_embedding_name`	Normalize and validate an embedding name, raising if it is unknown.
`get_embeddings`	Apply an embedding to the training and test datasets.
`is_transductive`	Return True if `embedding` is fit on the combined train and test rows.
`pqk`	This function generates quantum circuits, computes projections of the data onto these circuits.

__all__ Data:

QUVINE_HEADLINE_METHODS
QUVINE_METHODS
SKLEARN_METHODS