More compact parse tables

[XanderVertegaal]

Closes #49

This branch was developed using changes from this PR and this PR on langpro-container. Those two PRs should be merged before this PR can be reviewed. I have developed this PR based on a local branch where I've integrated the changes from both PRs.

This PR enables the final three CCG parse artifacts (i.e. the CCG Term, CCG Corrected Term and LLF). In a meeting with @kovvalsky, the following guidelines for styling the nodes were laid out.

• The chunker output (e.g. "I-NP") in leaf nodes can be omitted.
• For each node, the server returns a string consisting of a "rule" and "content", in the format rule[content]. These brackets surrounding are to be removed from the output.
• Rules may be formatted in a grey box, as long as they are put on the same line as the content (to avoid the table from becoming taller than it needs to be).
• The rule @ is common and trivial and should be filtered out.
• A colon : marks the start of a 'feature', which is set in subscript. The end of a string or a special character ()[]/\- terminates the feature.
• Node content should appear in uppercase, but only in "CCG Tree" output, except for those parts of the content that are a 'feature' or are found in a list of whitelisted elements (i.e. 'period' or 'conj').
• Node content should not wrap across multiple lines.