"Strict Mode" for tree extension

[glaserL]

A reviewer suggested implementing a strict mode when parsing PTB-like annotations in conll.

Currently, we heuristically determine which columns contain bracketing annotations and treat them differently. The reviewer suggested implementing an option that would enforce which columns that should happen to, in order to guarantee avoiding miss-guessing.

My current idea would be, to offer a parameter -s/--strict followed by the column indices that should be treated that way.

So in our sample bracketing file instead of calling it like:
./run.sh CoNLLBrackets2RDF http://replace.me/ WORD POS_ON PARSE_ON POS_PTB PARSE_PTB
the call would look like:
./run.sh CoNLLBrackets2RDF http://replace.me/ WORD POS_ON PARSE_ON POS_PTB PARSE_PTB -s 2 4

Of course, this is just a suggestion; I don't know if this could create conflicts with any other parameters I am currently not thinking about.

[chiarcos]

My current idea would be, to offer a parameter -s/--strict followed by the column indices that should be treated that way. So in our sample bracketing file <https://github.com/acoli-repo/conll-rdf/blob/master/data/bracketing.sample.conll> instead of calling it like: ./run.sh CoNLLBrackets2RDF http://replace.me/ WORD POS_ON PARSE_ON POS_PTB PARSE_PTB the call would look like: ./run.sh CoNLLBrackets2RDF http://replace.me/ WORD POS_ON PARSE_ON POS_PTB PARSE_PTB -s 2 4 Of course, this is just a suggestion; I don't know if this could create conflicts with any other parameters I am currently not thinking about.

There are also other things we could be strict about, so this is ambiguous. If it defines columns with brackets, this parameter should not be called "-s[trict]" but "-b[rackets]" or the like. Otherwise, that proposal has my full support.