Improve documentation on segment filter rules

[imagejan]

It's not really clear to me from the node descriptions whether the rules for filter on segment labels as well as filter overlapping segment labels in the Segment Cropper and Feature Calculator nodes take regular expressions or wild cards, or if that behavior can be configured in any way, similar to the Row Filter node.

By trial and error with the Feature Calculator (BETA) downstream of an Interactive Annotator node, I noticed that:

• a filter rule like Segment: \d+ works,
• a filter rule like Segment* as well, but
• Segment.* produces an empty table, despite being a valid regex to match e.g. "Segment: 1"

It would be nice if that was clearly documented somewhere, or if the choice would be left to the user, just like in the Row Filter and other KNIME core nodes.

[imagejan]

I also wonder why this is necessary:

https://github.com/knime-ip/knip/blob/8aa1488e09dafcbdfca903fd9ed4ed67d24e5efd/org.knime.knip.core/src/org/knime/knip/core/ui/imgviewer/events/RulebasedLabelFilter.java#L283-L298

Isn't that conflating wildcard and regex syntax? At least in some cases, it breaks correct regex matching.

[tibuch]

@imagejan thanks for pointing it out! I opened a PR which should fix this.

The new dialog works now like the Row Filter node and looks like this: