Batch integration, with pyFAI as the authority
2D/3D/4D HDF5 with lazy per-frame loading; 1D, 2D and radial methods in 2θ, q or d. The application only prepares the frame — binning is pyFAI's, so the numbers match a direct pyFAI call.
Desktop data processing for high-throughput scattering and diffraction at HEPS. SAXS/WAXS: it reads 2D/3D/4D HDF5 detector data, runs batch pyFAI azimuthal integration (1D / 2D / radial; 2θ, q and d), then overlays, background-subtracts, averages, merges and Guinier-analyses the curves. Powder diffraction: reference phases from CIF with ring prediction, peak detection and pseudo-Voigt fitting, search-match, unit-cell refinement, and in-situ Series with frame-by-frame peak tracking. Provenance is written into every export and session.

Daisy Batch Integrator is the offline data-processing desktop application for high-throughput scattering and diffraction experiments at HEPS (PyQt5 GUI). Integration is delegated to pyFAI: the application only prepares the frame (flips, mask, non-finite pixels, polarization validation) and hands it to pyFAI once, so the numbers are identical to calling pyFAI directly.
[frame,y,x] and 4D [frame,channel,y,x] HDF5, with lazy per-frame loading for large files; load or edit a PONI, masks (.npy, .edf, images), detector orientation 0–4, and call pyFAI-calib for interactive calibration..dat/.h5 output..dat, .h5, TIFF, PNG/PDF/SVG plots; provenance (geometry, integration parameters, fit settings, software version) written into the files and the session; sessions can be saved and restored.ctx exposes the current pattern and series, with results applicable back to the UI.File and frames → PONI / mask / calibration → batch integration → curve visualization and processing → (optional) diffraction analysis: reference phase → peak detection and fitting → search-match → unit-cell refinement → export and session.
Portable, no-install packages are provided for Windows 10/11 and Linux; running from source needs Python ≥ 3.11, PyQt5, pyFAI ≥ 2025.12.1, h5py, numpy, matplotlib and scipy. MIT licence. Boundaries: reference rings and sticks are a positional model (no structure factors or intensities), search-match structure-factor intensities are a scoring aid only, and unit-cell refinement is lattice least squares rather than Rietveld — the full boundary table is in the user guide.




2D/3D/4D HDF5 with lazy per-frame loading; 1D, 2D and radial methods in 2θ, q or d. The application only prepares the frame — binning is pyFAI's, so the numbers match a direct pyFAI call.
Overlay, background subtraction, averaging, an interactive two-curve merge, Guinier analysis, and per-curve scale/offset/q-crop.
Build a reference phase from a CIF (all seven crystal systems) or a built-in calibrant, predict its low-angle rings from the unit cell and overlay them on the measured profile; gemmi ships with the package for full space-group absences (glide/screw included).
Prominence-based detection plus pseudo-Voigt fitting, single peak or a joint fit of all peaks (background thread, cancellable, a cancelled run is discarded); match observed peaks to a reference phase by d-spacing and rank the candidates.
Lattice least squares on the matched peak positions, reporting refined values with uncertainties; optionally GSAS-II for Le Bail/Rietveld, isolated in a subprocess and cancellable.
Build a Series from multi-frame or in-situ data: heatmap, waterfall, frame-by-frame peak tracking and batch analysis, with frame indices and provenance kept.
.dat/.h5/TIFF plus PNG/PDF/SVG plots; geometry, integration parameters, fit settings and software version are written into the files and the session, which can be saved and restored.
Run Python analysis scripts inside the application: ctx exposes the current pattern and series, and results can be applied back to the UI or exported.
| Version | Platform | File size | Published | Downloads | |
|---|---|---|---|---|---|
| No package yet — use it online in TORCH. | |||||
Contact the maintainers above for usage, adaptation or customisation.
