cyanea-py

Overview

cyanea-py wraps the full Cyanea Rust ecosystem in a Python package via PyO3. Every function available in the Rust API is callable from Python with native-feeling ergonomics — no subprocess calls, no FFI boilerplate. Data passes between Python and Rust with zero-copy where possible, and NumPy arrays and Polars DataFrames are supported natively.

Key Concepts

Zero-Copy Data Exchange

PyO3’s buffer protocol integration means that NumPy arrays can be passed directly to Rust functions without copying the underlying memory. This is critical for large matrices (expression data, distance matrices, alignments) where copying would dominate runtime.

NumPy and Polars Interop

Functions that accept or return tabular data work with both NumPy arrays and Polars DataFrames. For example, cyanea.describe() accepts a list, a NumPy array, or a Polars Series. Functions returning tabular results (VCF records, BED intervals) return Polars DataFrames by default.

Maturin Build

cyanea-py is built with Maturin, the standard Rust→Python build tool. It produces wheels for Linux (x86_64, aarch64), macOS (x86_64, arm64), and Windows (x86_64).

Installation

pip install cyanea

Or from source:

pip install maturin
cd cyanea-py
maturin develop --release

Usage Examples

import cyanea
import numpy as np

# Sequence analysis
mrna = cyanea.transcribe("ATGGCTAGCAAAGAC")
protein = cyanea.translate("ATGGCTAGCAAAGAC")
gc = cyanea.gc_content("ATGGCTAGCAAAGAC")

# File parsing
records = cyanea.parse_fasta(open("genome.fa").read())
variants = cyanea.parse_vcf(open("calls.vcf").read())

# Statistics
data = np.random.randn(1000)
summary = cyanea.describe(data)
pvals = [0.001, 0.04, 0.03, 0.5]
adjusted = cyanea.benjamini_hochberg(pvals)

# Machine learning
embedding = cyanea.pca(data_matrix, n_features=100, n_components=2)
clusters = cyanea.kmeans(data_matrix, n_features=100, k=5)

# Phylogenetics
tree = cyanea.build_nj(labels, distance_matrix)

Performance

Use Cases

• Jupyter notebooks — Interactive analysis with Rust performance.
• Pipeline scripts — Drop-in replacement for BioPython with 10–30x speedups.
• Data science — Combine with pandas, scikit-learn, and seaborn workflows.