Web interface for Azulero by A. Basset and Eummy by M. Schirmer - two pipelines designed to produce color images from Euclid space telescope data.

Overview

Azumy wraps the azul and eummy command-line pipelines into a browser-based interface. It covers the full workflow:

Step	Command	Description
Find	azul find	Locate tile indices from object names or sky coordinates
Retrieve	azul retrieve	Download MER FITS files from ESA archives
Crop	azul crop	Visually select a region of interest on the VIS channel
Process	azul process or eummy	Render a color sRGB image from IYJH channels

Screenshot of the interface:

Requirements

• Docker and Docker Compose
• Internet access (ESA data archives)
• Optionally: ESA IDR or EC DSS credentials for private data access

Quick start

git clone https://github.com/Rolgim/Azumy.git
cd Azumy
chmod +x run.sh
./run.sh

On first run, the script will ask:

1. Workspace path — where FITS files and output images will be stored
2. IDR credentials — username/password for easidr.esac.esa.int (optional)
3. DSS credentials — username/password for eas-dps-rest-ops.esac.esa.int (optional)

Then open http://localhost:8000 in your browser.

SAS public access (Euclid Q1 data) works without credentials.

Usage

Find

• Click on the sky map (Aladin Lite) to select coordinates, or type an object name (e.g. M82)
• Upload a tiling GeoJSON file (DpdMerFinalCatalog.geojson) to display tile footprints on the map and enable tile lookup
• Click ▶ azul find — found tiles appear as chips below the terminal
• Select tiles and click → Send to Retrieve

The tiling GeoJSON for DR1 is available at (Euclid consortium members only):

https://gitlab.euclid-sgs.uk/sy-tools/ST_SMT_DATA/-/raw/DR1/data/DpdMerFinalCatalog.geojson

Retrieve

• Tile indices are pre-filled from Find
• Choose a provider: IDR (private), SAS (public), or DSS
• Click ▶ azul retrieve — files download one by one with a live progress indicator
• Click ✕ Cancel to interrupt at any time

Crop

• Enter the tile index and click ▶ Load VIS preview
• Draw a rectangle on the image to select a region
• The NumPy slicing spec is computed automatically (rounded to 500px increments)
• Click → Send to Process

Process

• The tile spec is pre-filled from Crop
• Select an engine from azul and eummy
• Expand Parameters to adjust rendering settings (zero points, scaling, stretch, hue, saturation...)
• Click ▶ process — output image is displayed when ready

Project structure

Azumy/
├── main.py              # FastAPI app
├── utils.py             # Shared utilities (stream_command, build_cmd)
├── routers/
│   ├── find.py          # azul find + GeoJSON tiling endpoint
│   ├── retrieve.py      # azul retrieve with cancel support
│   ├── crop.py          # VIS preview + slicing computation
│   ├── process.py       # azul process
│   └── workspace.py     # Workspace introspection
├── frontend/
│   ├── index.html
│   ├── styles.css
│   ├── aladin.js        # Aladin Lite (vendored)
│   └── js/
│       ├── main.js
│       ├── find.js
│       ├── retrieve.js
│       ├── crop.js
│       ├── process.js
│       ├── map.js
│       ├── settings.js
│       ├── terminal.js
│       ├── progress.js
│       └── websocket.js
├── Dockerfile
├── docker-compose.yml
├── run.sh               # First-launch setup script
└── .env                 # Generated by run.sh (gitignored)

Docker commands

# Start
./run.sh

# Stop
docker compose down

# Logs
docker compose logs -f

# Rebuild after code changes
docker compose up --build -d

# Reset credentials and workspace config
rm .env secrets/netrc
./run.sh

Credits

• Azulero - Antoine Basset (CNES) et al.
• Eummy - Mischa Schirmer
• Aladin Lite - CDS Strasbourg
• Image data - ESA Euclid / Euclid Consortium / NASA