What’s new

Version 2.1

This version is the result of taking into account the feedback of v2.0’s users; thank you all!

The main focus is to make batch processing more flexible, for example by retrieving cutouts of different sizes, bypassing the download phase when working on the Datalabs, or calling Azulero directly from Python (notebooks). Additional upgrades include the revival of azul crop without Matplotlib and the improvement of the background rendering with azul process.

Below are a few details; for more, see the Change log.

For ESA Datalabs users

  • On Datalabs, no downloads or copies are performed.

  • A Python API implements core features of azul retrieve and azul process.

  • Credentials do not need to be stored in a file: an interactive mode prompts the user for passwords.

azul retrieve

  • In a single command, multiple radii are supported for cutout retrieval.

  • Queries and downloads are retried on failure.

azul process

  • The sky background is now darker by default, and less dependent on the white point, which makes it more stable with -w 0.

  • In incomplete tiles, the large empty regions are not inpainted anymore, which means the RAM usage is reasonable and the empty regions are rendered dark.

  • MEF files with one extension per band are accepted as targets.

Version 2.0

This version is an almost complete rewriting of Azulero. We have reworked every aspect of the software from the lowest level (pixel ordering, logging, code quality) to the highest level (parametrization, documentation, workflow). The result is a much cleaner and more extensible design, able to accommodate novel features more efficiently. Speaking of which, the purpose of this release is mass production of images. To this end, we introduce cutout retrieval and pipelining. In turn, they enable faster and parallel processing, as well as streamlined production workflows. In addition, Azulero 2.0 introduces new command azul arrange aimed at performing a very common post-processing stage: making collages from collections of images.

Cutout retrieval

Command azul retrieve can now download tile regions, which may save a significant amount of time and disk space. Typically, a 1’ x 1’ cutout is 1000x smaller and therefore 1000x faster to download and process than a WIDE tile.

Pipelining

Pipelining is a new way of chaining operations with Azulero. Relying on the Unix or Windows pipe operator, it is now possible to execute the various image production stages (e.g. downloading, rendering and collage) in a single command line.

Collage

New command azul arrange arranges input images into a grid. In combination with cutout retrieval, it offers a very convenient way to render images from catalogs.

All of these evolutions come with breaking changes, thus the major version number. Here is an overview (for more details, see the Change log):

Updated commands with breaking changes:

New command:

Deprecated commands:

  • azul find (merged into azul retrieve)

  • azul crop (to be updated later)

  • azul assemble (superseded by azul arrange)

  • azul overlay (to be updated later)