qtensor

qtensor is a modern NumPy-based reconstruction of QTENSORSPACE, a named-party tensor library the author originally wrote in Cython/Numeric around 2002 for the quantum-communication protocols in Aschauer, Quantum communication in noisy environments, PhD thesis, LMU Munich, 2005. See CITATION.cff for how to cite this package and its predecessor.

The package implements named-party tensor algebra for small quantum-information experiments, including:

• Party and Qubit
• Qtensor, including dagger(), normalize(), trace_value(), eigenvectors(), and is_eigenvector()
• Ket, Bra, P, density_operator, operator, and diagonal_vector
• convenience constructors such as ghz_state, w_state, smolin_state, duer_state, and graph_state
• correlation-analysis helpers such as sigma_y, pauli, expectation, correlation_tensor, flatten_correlation_tensor, and combined_shadow_map
• Named-party tensor multiplication without manually inserting identity operators
• Built-in gates such as H, CNOT, sigma_x, sigma_z, and sigma_0
• Partial trace, basis formatting, diagonal-density evolution, and spectral helpers such as singular_values, nuclear_norm, and frobenius_norm

The installed console entry point is qtensor, which currently prints the package name.

Running examples

From the repository root:

uv run python examples/bell_state.py
uv run python examples/teleportation.py
uv run python examples/diagonal_density.py
uv run python examples/cluster_state.py
uv run python examples/structured_states.py
uv run python examples/correlation_analysis.py

The example coverage is:

• bell_state.py: Bell-state generation from |1,0> with H and CNOT
• teleportation.py: pure-state teleportation and success probability
• diagonal_density.py: diagonal-density evolution via abs_squared(CNOT(...))
• cluster_state.py: cluster-state correlation signatures and reduced-state eigenvectors
• structured_states.py: GHZ, W, Smolin, Duer, and graph-state constructors
• correlation_analysis.py: Pauli expectations, correlation tensors, flattening, and combined shadow maps

Building documentation

MkDocs is configured as a docs dependency group, not as a build-system dependency. Build dependencies belong under [build-system] and are only for packaging the Python distribution itself.

Use:

uv run --group docs mkdocs serve
uv run --group docs mkdocs build --strict

The site configuration lives in mkdocs.yml and the documentation pages live in docs/.

Development notes

This codebase was written almost entirely by an AI coding agent (Claude Code, Anthropic), directed and reviewed by the author: design decisions, physics/API correctness, and acceptance testing (lint, type checks, unit tests) were the author's responsibility throughout. Commits with agent contributions carry a Co-Authored-By trailer identifying the model. See AGENTS.md for the working conventions the agent followed.

Because a work with little or no human authorship has uncertain copyright status, the package is released under CC0 rather than a traditional copyright license — see LICENSE.