Add simple values() method in Tensorics to retrieve all the values of a Tensor

[tensorics]

Add simple values() method in Tensorics to retrieve all the values of a Tensor

[kaifox]

yes, then its clear!

Potentially uniqueValuesOf(...) as the other methods in Tensorics?

public static final <T> Set<T> Tensorics.uniqueValuesOf(Tensor<T> tensor);

In general, I clients should void using also the asMap() method directly ... In general, Tensorics.mapFrom(Tensor t) should be used. This calls asMap() if available, which makes it more efficient....

[kaifox]

Hi @delph74

What would you expect as a return type of such a method?

[delph74]

Hello @kaifox ,
We could return a Set, and in this case maybe call the method uniqueValues().
What do you think?

[agorzawski]

Hi @delph74 and @kaifox

As far I remember, a tensor has no 'values' like method. But (if implements Mappable, and this is the case for the ImmutableTensor) it has .asMap()... which you follow with values().

The practical use of such values collection (only values) may be very limited, as this may vary with their position in the Tensor (coordinate).

[kaifox]

Hey @agorzawski !!! Cool to see you here!

[agorzawski]

@kaifox I just want to recall, that I have never left this project ;-)

[delph74]

Yes @kaifox , something like that would be perfect.
I often use values for Map and I think for a Tensor we may have many usecase where such method would be needed!

[michi42]

Maybe one more thing - there is Tensorics.stream(Tensor) that allows processing Position-Value pairs via the Java 8 streaming API, and TensorStreamFilters and TensorStreamMappers to help with filtering/mapping. Also there are some collectors to build tensors or tensorbackeds with streams in TensorStreams.

[delph74]

yes of course it it possible and not to difficult to get all the values of a Tensor, like that for example
Tensorics.stream(betaStarParameters).map(Map.Entry::getValue).collect(toSet());

So nothing urgent or really missing, just for convenience and make it easier.

[kaifox]

Good point @michi42 ! ... Still, convenience methods can still be justified... I think there is nothing against adding it.

[kaifox]

@agorzawski , good to hear!

[michi42]

Yes sure, I did not mean to object that :-)

[kaifox]

I know ;-)

[kaifox]

Just wanted to encourage @delph74 to add it if she feels like ;-)

[delph74]

Yes why not, I would then add my first contribution to tensoric project ✌️ 😃

[kaifox]

Cooool!!!!!