How michaeldalbrow/dazzle works
Project type and purpose
• From the file tree and README, Dazzle is a Python library/CLI tool (not a web or distributed microservice) for over‐sampled image construction, difference imaging, transient detection, and PSF photometry, targeted at the Nancy Grace Roman Space Telescope data.
File structure
• Top‐level items:
• LICENSE, README.md, PSFs/ (data folder), dazzel_paper.pdf (paper)
• Package directory “dazzle/” with four core modules:
– dazzel.py (entry points and orchestration)
– detect.py (transient detection logic)
– make_psf_grid.py (PSF oversampling)
– photometry.py (difference‐image PSF photometry)
– utils.py (shared utilities, I/O, FITS handling)
• tests/ directory with JSON configs, test scripts, and fixtures.
README for additional insights
• Workflow is linear and script‐driven:
1) difference‐image construction
2) transient detection
3) photometry
• Configuration via JSON that controls input data paths, image splits, iteration counts.
• Parallel processing by splitting large images into blocks.