upgrade to CDK 2 about nodes4knime HOT 12 OPEN

Have unsubscribed from all nodes3knime related updates now :-). So feel free to use this is a todo list

from nodes4knime.

not an "issue"... this is a ToDo

from nodes4knime.

Are there any updates on this?

from nodes4knime.

No :(

from nodes4knime.

I have recently looked into updating JChemPaint to CDK 2.1 and 2.0, but there are regressions. See JChemPaint/jchempaint#191

from nodes4knime.

Is JChemPaint the blocker and if so is this still the case?

Could we rebrand the existing nodes legacy nodes and create a new plugin based on CDK 2.x?

from nodes4knime.

@webbres, that's exactly what Daria and I were discussing yesterday.

from nodes4knime.

👍 happy to contribute to that effort if it's done

from nodes4knime.

I've forked a repo for this: https://github.com/cdk/nodes4knime-cdk2

But the wifi here is too slow to set up the KNIME SDK

from nodes4knime.

Why did you fork it Egon? The API should be compatible enough that it didn't need a fork. Any changes are likely unspecified invariants being missed by the callee.

from nodes4knime.

I think the main issue is JChemPaint and Ambit don't work with CDK 2, so upgrading loses that functionality which breaks backwards compatibility with existing workflows.

You can't deprecate the nodes as they don't work anymore.

from nodes4knime.

Why did you fork it Egon? The API should be compatible enough that it didn't need a fork. Any changes are likely unspecified invariants being missed by the callee.

Think of it as the migration path from Python2 to Python3: there were so many dependencies, many tools could not easily migrate to Python3. Similarly, there is functonality in the CDK-1.5-based KNIME nodes that are not easily migrated.

The idea of the fork is to have both world in parallel, allowing people to use CDK-2.x-based KNIME nodes whenever possible, but at the same time being able to install the CDK-1.5-based KNIME nodes when needed.

from nodes4knime.