make manifest accessible

[FreakyBytes]

Trac Ticket #141

component: code
owner: somebody
reporter: anonymous
created: 2016-06-07 03:29:24
milestone:
type: feature request
version:
keywords:

For CellML people, when generating a COMBINE archive with WebCAT,

it would be nice to be able to download the manifest.xml only (instead of downloading the whole archive and extracting the manifest.xml from there). This is because they can generate the archive in PMR based on a manifest.xml).

Second thing: Would you be able to offer a web service that returns a manifest.xml file for a CellML exposure?

Dagmar (on behalf of Tommy)

I came from WEBCAT-INTERFACE
(feel free to delete this if it is not related to your report)

comment 1

time: 2016-06-07 17:21:27
author: martin

I don't know if I understand correctly, but do you want a pipeline/API that git-clones from PMR, creates an archive from the contents and returns the generated manifest.xml? Just to create a second version of the archive at PMR?

That sounds like creating an archive in by the most complicated means possible. In German we'd call it "von hintern durch die Brust ins Auge" ;-)
Moreover, you would be missing the meta data. So we would need to create another API to also allow for extracting the meta data file. That would take you two calls. Each involving multiple NZ-GER network connections and your users would need to wait for seconds...

I would rather suggest you add a link to an instance of the !Combine/Archive-Web interface like

http://webcat.sems.uni-rostock.de/rest/import?name=YOURTITLE&type=GIT&remote=URLTOPMR

This link is super easy to generate, you can even think about doing it using /JavaScript on the users site. If users follow that link they end up the !Combine/Archive-Web interface and can further modify it over there or download the resulting archive.

If you don't execute the above link with a browser you'll see that is also returns a some machine processable information, like

• archiveName`: that's actually what you provided with the `name=YOURTITLE above
• shareLink: that's the link to the archive if you want it to share with some one
• archiveId: the id of the archive
• workspaceId: the id of the workspace

You'll be interested in the archiveId` and the `workspaceId, as you can create a URL to the actual COMBINE archive from it:

http://webcat.sems.uni-rostock.de/cat/download/archive/workspaceId/archiveId.omex

Thus, users do not need to visit the web interface and immediately get the generated archive.

And remember, it is super easy to setup your own instance of the !Combine/Archive-Web interface, if you have privacy or network-time concerns. Just deploy a pre-compiled binary from http://bin.sems.uni-rostock.de//CombineArchiveWeb/ or see /BuildAndInstall to compile and build your own instance.

comment 2

time: 2016-06-09 22:58:23
author: martin

We now also provide DOCKER images for the /CombineArchive Web interface: https://hub.docker.com/r/binfalse/webcat/

Thus, setting up the web interface was never easier nor cleaner. And you can safe the privacy of your users, and may get better performance from your web servers. :)