Tool for parsing the metadata of datasets as parameter values

[kostrykin]

As discussed today with @bgruening and @wm75, there might be some demand for a tool that parses parameter values from the metadata of datasets. I was about to create such a tool specifically for images:

https://github.com/BMCV/galaxy-image-analysis/pull/226/changes#diff-bfb2c10cf195c34a55ca9ead4f640efb91c9888715903f58648bc58c01fa092b

…but it was pointed out that having a more general tool will probably be more sustainable.

While I absolutely agree that having a general tool is much better, I have a bit doubts regarding usability of such a tool. For example, in the image-specific version linked above, we can simply hard-wire all the known metadata fields to corresponding outputs. However, in a general tool, I think this isn't possible, so the user would have to select what metadata fields they need to be parsed.

The issue is that—I think—it is not really user-friendly to require users to provide tokens like width or height in a free text field, tokens that are not really documented somewhere, tokens that users would need to look up in the Galaxy sources. Can this be done in a user-friendly manner somehow?

[bernt-matthias]

I think one would need to implement a new filter like https://github.com/galaxyproject/galaxy/blob/03e0b49756491fb1b7109bdbf180aed597f6db7a/lib/galaxy/tools/parameters/dynamic_options.py#L147 (it's a bit confusing, but some filters can add values). The data_meta filter can for instance add metadata values already.

Or maybe better a mechanism to generate values like from_dataset (eg from_dataset_metadata).

Was also thinking if the API might be able to provide metadata Infos. Then one could maybe use https://docs.galaxyproject.org/en/latest/dev/schema.html#from-url.