Scipp

26.3.1 · active · verified Fri Apr 17

Scipp is a Python library for multi-dimensional data arrays with labeled dimensions, designed for scientific data analysis, especially in neutron and muon scattering. It provides unit-aware data structures and operations, enabling robust handling of physical quantities. The current version is 26.3.1, and it maintains a rapid release cadence with monthly major updates.

Common errors

```
ModuleNotFoundError: No module named 'scipp'
```
cause The scipp package is not installed in the current Python environment.

fix Run `pip install scipp` to install the library.
```
scipp.core.UnitError: Units do not match
```
cause Attempting an operation (e.g., addition, subtraction) on Scipp variables or data arrays that have incompatible physical units.

fix Ensure all operands have compatible units. Use `my_var.to(target_unit)` to convert units or `my_var.without_units()` to remove units before the operation.
```
TypeError: 'Variable' object is not subscriptable
```
cause Trying to access elements of a `scipp.Variable` object using array indexing (e.g., `my_variable[0]`) instead of its underlying NumPy array.

fix Access the underlying NumPy array first using `.values` attribute: `my_variable.values[0]`.
```
scipp.core.DimensionError: Cannot perform operation ... dimensions ... do not align
```
cause Operations between data arrays with non-matching dimensions, dimension labels, or dimension order without explicit alignment.

fix Ensure dimensions align or explicitly broadcast/reorder them. For element-wise operations, dimensions and their order must match. Use `my_da.transpose()` or ensure common dimensions are present.

Warnings

breaking Scipp dropped support for Python 3.10. Users must upgrade to Python 3.11 or newer.
Fix: Upgrade your Python environment to 3.11 or a later version.
breaking The custom HTML representation for Scipp objects in Jupyter notebooks (`sc` notebook HTML repr) was removed. Objects now rely on their standard `__repr__` method.
Fix: No direct fix; observe the new default text-based representation. If custom visualization is needed, consider external plotting libraries or manual rendering.
gotcha Scipp is strictly unit-aware. Operations on `Variable` or `DataArray` objects with incompatible units will raise a `UnitError`.
Fix: Ensure units are compatible for operations, convert units using `.to()` method, or explicitly strip units with `var.without_units()` if desired.
gotcha Accessing the underlying NumPy array data requires using the `.values` attribute. Directly indexing a `Variable` object will result in a `TypeError`.
Fix: To get the NumPy array from a `Variable` or `DataArray`'s data, use `my_var.values` or `my_data_array.values`.

Install

pip install scipp Install with pip

Imports

scipp
```
import scipp as sc
```
Variable
```
from scipp import Variable
```
DataArray
```
from scipp import DataArray
```
Dataset
```
from scipp import Dataset
```

Quickstart

This quickstart demonstrates how to create a unit-aware `Variable` and then encapsulate it within a `DataArray` along with its coordinates. It then performs a basic operation (sum) that respects Scipp's unit-aware nature.

import scipp as sc
import numpy as np

# Create a variable with units
x = sc.linspace(dim='x', start=0.1, stop=0.9, num=10, unit='m')
y = sc.sin(x)

# Create a DataArray, including coordinates and data
data_array = sc.DataArray(data=y, coords={'x': x})

print("Original DataArray:\n", data_array)

# Perform a unit-aware operation (e.g., sum over 'x' dimension)
sum_result = data_array.sum('x')
print("\nSum along 'x' dimension:\n", sum_result)

view raw JSON →