Why does Galaxy use a 'history' and not a 'directory' to manage user content?

Question: Why does Galaxy use a 'history' and not a 'directory' to manage user content?

2.2 years ago by

steve • 30

NYC

steve • 30 wrote:

Why was this implementation chosen, at least on the front-end if not on the back-end as well? I understand that files in a user's History are currently contained as a blob in a database (correct?). Does this have an advantage over simply storing it as plain files in a user-directory created by Galaxy? Is there any plan to implement a "directory view" on the front-end to manage history files easier when the history grows very large?

In the context of my own use cases, a lot of our current pipelines are very 'file-oriented', so the prospect of trying to get them implemented in Galaxy seems a little daunting.

admin history galaxy database dataset • 742 views

ADD COMMENT • link •

modified 2.2 years ago by James Taylor ♦ 470 • written 2.2 years ago by steve • 30

2.2 years ago by

James Taylor ♦ 470

United States

James Taylor ♦ 470 wrote:

While Galaxy stores metadata about a dataset (name, datatype, tags) in its database, the data are actually stored in files on disk. This is what makes it possible to integrate existing command line tools that alrady work on files.

The history is one view of a set of files in Galaxy that corresponds to the order in which analysis steps were performed. Datasets can also be viewed in a hierarchical "Data Library".

We do not store the file in a user's directory because datasets in Galaxy can be shared by many users.

ADD COMMENT • link written 2.2 years ago by James Taylor ♦ 470

OK thanks, I see the page about this here. I have poked around and found some of the files I've generated in Galaxy here: /galaxy-central/database/files/. One of my concerns is that non-admin Galaxy users will be requesting downstream analysis (outside of Galaxy) on files they have generated within Galaxy. Is there an easy way for them to find the direct paths to their files so that our non-Galaxy Bioinformaticians can easily access the files on the server? Hoping to avoid unneccessary import/export of data in Galaxy if we can just access it directly from the file system. EDIT: Just found that clicking 'View Details' on a History item does show the full path.

ADD REPLY • link modified 2.2 years ago • written 2.2 years ago by steve • 30

Similar posts • Search »