xarray-einstats

0.10.0 · active · verified Fri Apr 10

xarray-einstats is a Python library that provides a high-level API to combine the symbolic array manipulation of `einstats` with `xarray`'s labeled dimensions for statistical operations and linear algebra. It aims to simplify common data analysis tasks on `xarray.DataArray` and `xarray.Dataset` objects, enabling operations like reduction, broadcasting, and reshaping with `einops`-like patterns. The current version is 0.10.0, and it generally follows an irregular release cadence, with updates typically driven by new features or important bug fixes.

Warnings

breaking As of version 0.10.0, `xarray-einstats` requires Python 3.12 or newer. This is a significant bump from previous versions (e.g., 0.9.0+ required Python 3.10+).
Fix: Upgrade your Python environment to 3.12 or newer. If you need to use an older Python version (e.g., 3.10 or 3.11), you must pin `xarray-einstats` to a version less than 0.10.0 (e.g., `xarray-einstats<0.10`).
breaking Version 0.10.0 introduced minimum version requirements for key dependencies: `xarray>=2024.2.0`, `scipy>=1.11.0`, `numpy>=1.25.0`, and `einstats>=0.7.0`. Older versions of these libraries are no longer supported.
Fix: Ensure all dependent libraries are updated to their respective minimum versions before upgrading `xarray-einstats` to 0.10.0 or higher. Use `pip install xarray-einstats --upgrade` to let pip handle dependencies or manually update them.
gotcha The `dims` argument in `xarray-einstats` functions (e.g., `mean`, `sum`, `quantile`) accepts various formats: a string, a list of strings, or a dictionary for `einops`-like patterns. Misunderstanding the dictionary format, especially for reshaping or combining dimensions, is a common source of errors.
Fix: Refer to the `einstats` and `xarray-einstats` documentation for detailed examples of `dims` argument usage. Start with simple string or list of strings for basic operations before moving to more complex dictionary patterns for `einops`-like reshaping.

Install

pip install xarray-einstats Install stable version

Imports

stats
```
import xarray_einstats.stats as xestats
```
The primary module for statistical functions.
mean
```
from xarray_einstats.stats import mean
```
A direct import for a specific statistical function.

Quickstart

This example demonstrates creating an xarray DataArray and then using `xarray-einstats.stats.mean` to calculate the mean along a specific dimension. It also shows a basic `einops`-like pattern for combining and summing dimensions using a dictionary in the `dims` argument.

import xarray as xr
import numpy as np
import xarray_einstats.stats as xestats

# Create a dummy xarray DataArray
data = xr.DataArray(
    np.random.normal(size=(2, 3, 4)),
    coords={"a": [0, 1], "b": [0, 1, 2], "c": [0, 1, 2, 3]},
    dims=["a", "b", "c"],
)

print("Original data:\n", data)

# Calculate the mean along dimension 'b'
mean_b = xestats.mean(data, dims="b")
print("\nMean along 'b' dimension:\n", mean_b)

# Use einops-like patterns for more complex operations
# Reshape 'b c' into a new dimension '(b c)' and then sum
sum_bc_flattened = xestats.sum(data, dims={"b c": "_"}) # '_' signifies a new dimension combining 'b' and 'c'
print("\nSum after flattening 'b' and 'c' dimensions:\n", sum_bc_flattened)

view raw JSON →