Additional helper methods on Taggables about octue-sdk-python HOT 8 CLOSED

And/Or some similar useful APIs, yes, that'd be elegant. contains might be a good one too.

Perhaps this is a good opportunity to implement a chainable filter on TagGroup which could work like a list except with extra methods to filter further - so you could go, for example:

<thing>.getTags.startswith('meta').contains('sys2').subTags()[-1]

which, if applied to a Taggable with tags="tag1 tag2 meta:sys1:1234 meta:sys2:3456" would result in 3456.

from octue-sdk-python.

@thclark I've implemented chained filtering 💯 Isn't the proposed has method the same as __contains__?

It certainly can be, but I don't know whether we want to use __contains__ for that. Probably don't want to overload such an important method in a mixin.

For example, a Dataset is a Taggable but you'd want to reserve __contains__ for much more obvious uses like if <DataFile> in Dataset.

from octue-sdk-python.

@time-trader you might want to follow this

from octue-sdk-python.

@thclark do you want the startswith etc methods to be able to work on subtags too? Or just be able to get subtags from a TagGroup?

from octue-sdk-python.

The other side of the problem is needing to be able to test against tags, for searching matching files in a manifest.

This is where we really should be using a search index and lucene syntax, which we could do trivially if this were online-only. But there aren't any native python implementations (everything requires a JVM, and I foresee trouble ahead with that...), and being able to do it locally would be helpful.

But, we could implement a has method on the taggable class that returns T/F given the filters above.

That'd allow us to implement a match_files() method on a a Dataset which would implement a crude O(MN) search (M = no. tags, N = no. files) for files meeting those criteria.

from octue-sdk-python.

@thclark I've implemented chained filtering 💯 Isn't the proposed has method the same as __contains__?

from octue-sdk-python.

Maybe we use something more descriptive though, like has_tag or has_tags

from octue-sdk-python.

I've swapped __contains__ for has_tag but I've just realised that it doesn't matter if we have loads of magic methods for TagGroup as it's not a mixin - its instances are used as an attribute of a mixin.

from octue-sdk-python.