Try to download all files at once

[m-mohr]

My suggestion in regard of user convenience would be the possibility to download the whole layerstack by integrating this opportunity into the action buttons in the web editor. As for now, this opens a window where it is possible to download one layer after another.

Try to add a download button to the modal where you can download all files at once. Related to #265 (don't download metadata by default)

[m-mohr]

Closing as this is restricted by the browsers (either memory or number of parallel downloads) and would basically need a server-side solution which would need to zip the files and offer them for download.

[m-mohr]

There seems to be a Chrome only API: https://github.com/WICG/file-system-access

Anyway, no solution seems to work across all browsers in a reliable way.

[Patrick1G]

@m-mohr some option for downloading potentially hundreds of files from the editor would be valuable, e.g. here: eodcgmbh/sen2like#3 (comment)

if the browser does not support this, could the save_results process be adapted or a new process be defined e.g. save_custom_results()... For example having the ability to stack a time series by spectral bands would be super valuable..
Or even simply adding all result files into a single zip file would be valuable..

[m-mohr]

@patrick-griffiths I understand and agree that this would be valuable and I would be very happy to implement this if there would be a technical solution available in the Browser. Unfortunately, it looks like there is non (yet). We also discussed this recently in the openEO meeting with a wide variety of developers and we agreed that there is no good solution due to Browser limitations. It needs a server side implementation if we want to provide a user-friendly solution, even for adding them to a ZIP file. So this should be discussed as part of the overall architecture. Maybe the aggregator can do this as an added functionality in openEO Platform based on normal content-negotiation mechanisms (happy to clarify this in the API).

[Patrick1G]

@m-mohr - then lets propose this (server side implementation, via the aggregator) as one of the evolution enhancements we could still do within the platform contract during the team call on Thursday... Wonder if you could come up with a cost estimate for such a task...

In the meantime, I wonder if we could create a workaround in the Editor using the UDP functionality?

[m-mohr]

In the meantime, I wonder if we could create a workaround in the Editor using the UDP functionality?

I'm not aware of any implemented process that would allow this.

[Patrick1G]

@m-mohr - maybe along the python client solution that @ValentinaHutter had suggested:

results = job.get_results()
metadata = results.get_metadata()
results.download_files("./data/")

[m-mohr]

I don't understand how that should work?! I can't run Python in the Browser.

[Patrick1G]

I was wonder if we could pack the python functionality for downloading all files at once into a User Defined Process that can then be (re)used in the editor to download all files at once. But maybe the problem with the browser would still persist?

[m-mohr]

Some smart people seem to have found a solution that we could try an re-use: https://github.com/jimmywarting/StreamSaver.js
-> Saving multiple files as a zip

[Patrick1G]

@m-mohr - lets put in a dedicated task for our final SAP!

[m-mohr]

It has limitations though, which seem to be around 500MB - 1 GB in total. So a server-side solution is still preferrable.