New "upload" command!

Added a new non-batch single-folder no-zipping upload command, as it should be since the beginning.

Even more focused and stable.

No more NightLogs, Observation and Calibration syncing! Data syncing is always hard. Keeping the wish that Oort is able to sync such a sophisticated schema from local files to the cloud is way too risky, and makes Oort less reliable and focused.

Thanks to the newly-exposed tags of Datasets and Datafiles, such NightLogs, Observation and Calibrations objects, built upon the content of the DataFiles will be made later on, on the backend itself.

More stable than ever.

Changes:

• The zipping of data files is now an option (default = False) in the watch command.
• The possibility of providing a custom astronomer to upload on behalf of whom is temporarily removed.
• Added the possibility to login with the primary email address.

Improvements:

• The dropdown for selected a watched folder in the server webpage doesn't "jump" anymore when open.
• The watch command now correctly returns without being tight to the initial_walk of the folder watch. It is a non-trivial fix. No more duplicate PathObserver, one held by uploader process, and one help by the command line.
• The watch command warns for a folder already watched, and whether this is the home folder.
• A SqliteQueuedDatabase is used in place of the default SqliteDatabase to avoid database locked issues.
• Logs format have been improved and are more consistent throughout.
• Adding object types in various places to improve code checks.
• Improved the footer message about hidden files in the server webpage.
• The state of the uploader process appears in red in the server page if it is stopped.

Bugfixes:

• There is no conflict anymore between providing an upload_key and an organisation subdomain in the API requests. File will be directed to the organisation account of an organisation is provided, no matter what (unless uploader is not a member, of course).
• The check for an existing remote resource does not rely on the resource's name anymore, since this is a loosely-defined value anyway, and makes the search filtering less reliable.
• The target_name of the FITS/XISF file is correctly used instead of the dataset_name in the Observation and Calibration objects.
• The keywords DATE-OBS and DATE_OBS are checked before the DATE one in the FITS files to avoid the case where the latter is used for legacy date formats.
• Fixed the upload restart query by filtering the Upload objects also by the folder_path associated with the event_handler!
• Fixed the name and usage of config functions to avoid conflicting with the ones of the Arcsecond CLI.
• Important permission & filtering fix on the backend which prevented the upload of some files because their name already existed in another dataset.

More complete uploads

In this release, along with numerous small fixes, we allow to upload files that have no date in their headers (see documentation for more details), and files that zipped (with zip, gzip or bzip2).

1.0!

After much of polishing, here is Oort tool 1.0, the simplest way to upload your astronomical data to arcsecond.io.

Including server files

0.1.2

Including server files

First release to Pypi

0.1.1

First release to Pypi

First release

First release of Oort cloud!