silx

3.0.0 verified Fri May 01 auth: no python

silx is a collection of Python packages to support the development of data assessment, reduction and analysis applications at synchrotron radiation facilities. It provides tools for reading HDF5/NeXus files, histogramming, fitting, 1D/2D visualization, and image processing. Current version is 3.0.0, with releases approximately every 6–12 months.

pip install silx

Common errors

error ModuleNotFoundError: No module named 'silx.gui' ↓

cause Qt bindings not installed.

fix

Install a Qt binding: pip install silx[gui] or pip install PyQt5.

error ImportError: cannot import name 'fabioh5' from 'silx.io' ↓

cause The module was renamed from fabioh5 to fabio_h5 in silx 1.0.

fix

Change import to: from silx.io import fabio_h5.

error AttributeError: module 'silx' has no attribute 'plot' ↓

cause The top-level silx.plot module was removed in silx 1.0.

fix

Use from silx.gui import plot instead.

error OSError: [Errno 101] Network is unreachable: 'https://...' ↓

cause silx tries to download a test file or load an online resource without internet.

fix

Disable online access or set environment variable SILX_DISABLE_NET=1.

Warnings

breaking In silx 1.0.0, the old top-level modules (silx.plot, silx.io.fabioh5, silx.math) were reorganized. Code using old import paths will break. ↓

fix Update imports: silx.gui.plot instead of silx.plot; silx.io.fabio_h5 instead of silx.io.fabioh5; silx.math.histogram moved to silx.sx.histogram.

breaking silx 2.0.0 dropped Python 2 support and removed deprecated APIs like silx.math.fit and silx.io.specfile. ↓

fix Use silx.sx.fit or external fitting libraries. Use silx.io.fabio_h5 or h5py directly for SPEC files.

gotcha silx.gui.plot requires a Qt backend. If no Qt bindings are installed (PyQt5, PySide2, etc.), importing silx.gui will raise an ImportError. ↓

fix Install a Qt binding: pip install silx[gui] or pip install PyQt5 (or PySide6).

deprecated Functions in silx.math (like silx.math.histogram) are deprecated in favor of silx.sx functions. ↓

fix Replace calls: silx.math.histogram → silx.sx.histogram.

gotcha silx.io.fabio_h5.open() returns a FabioFrame, not a numpy array. Access the data via .data attribute. ↓

fix Use fr = fabio_h5.open('file'); data = fr.data.

Imports

silx
```
import silx
```
utils
wrong
```
from silx.utils import *
```
correct
```
from silx import utils
```
Avoid star imports; use explicit symbols.
io
wrong
```
from silx.io import fabioh5
```
correct
```
from silx.io import fabio_h5
```
In older versions (pre-2.0) the module was named fabioh5; renamed to fabio_h5.
plot
wrong
```
from silx.plot import ...
```
correct
```
from silx.gui import plot
```
The plotting module moved to silx.gui.plot in version 1.0.0.
OpenCL
wrong
```
from silx import ocl
```
correct
```
from silx.opencl import ocl
```
OpenCL support is under silx.opencl, not top-level.

Quickstart

Basic imports and usage of silx: reading files, histogram, and plotting (GUI).

import silx
import numpy as np
from silx import sx
from silx.io import fabio_h5
from silx.gui import plot

# Read a HDF5/NeXus image
import os
filepath = os.environ.get('SILX_TEST_FILE', 'example.h5')
if os.path.exists(filepath):
    data = fabio_h5.open(filepath)
    print(type(data))

# Simple histogram
x = np.random.normal(size=1000)
hist, edges = sx.histogram(x, bins=10)
print('histogram:', hist)

# Create a 1D plot window (GUI only if interactive)
# plot1D = plot.Plot1D()
# plot1D.addCurve(x, y)
# plot1D.show()