BornAgain ≻ git-develop ≻ Documentation ≻ Reference ≻ Data import

Overview

BornAgain reads no data files itself: experimental data are loaded on the Python side and kept as NumPy arrays. A fitting or plotting script then compares these arrays with simulation.simulate().intensities().

The tools:

  • numpy for ASCII tables and matrices;
  • Fabio for detector image formats (pip install fabio);
  • h5py for arbitrary HDF5 containers and nexusformat for NeXus-aware access (pip install h5py or pip install nexusformat);
  • ba.read_columns, ba.read_motofit, and ba.read_bersans for column tables with scan-axis preparation and instrument formats that Fabio does not cover; all three accept gz/bz2-compressed files.

The supported data kinds, each with its format description and import recipe:

Any other format can be read with any Python tool that suits it: BornAgain only ever works with plain NumPy arrays. If a column is to serve as a scan axis, remember that BornAgain requires strictly ascending values there.

Using Loaded Data in Fits

How to compare the loaded arrays with simulated intensities in a residual function, including the treatment of masked detector pixels, is described in Python tutorial > Fitting and in the fit reference.

Examples