Truncation should be done in a smarter way

[fatadel]

In the graph type visualizations (jsonld-vis), the truncation should be done in a smarter way. The following screenshot shows the problem with the current approach - the string is not shown completely despite the fact it is already truncated.

[egekorkan]

This is visible only in some specific cases like the example above. The max width is the problem, this TD has the same string in two places but only the one on the right is truncated.

[fatadel]

The logic for truncation contains some magic numbers, like an assumption of 1 character being 9 pixels long. The original logic for this type of visualization was adapted from here, which is used in the JSON-LD Playground.
Maybe we can constrain each text in size and let CSS decide for truncation... Though, we also need to make sure that the text direction is correctly taken into account.