ThomasHoek / LRC-Dataset-Creator

Pipeline to transfer NLI datasets from CCG into unsupervised LRC datasets

# CCG to LRC pipeline

Initial commit for parse CCG trees into LRC trees.


todo:
1. Produce file to automatically generate .sen and .ccg files.
2. information for CandC 
3. Advanced filtering, using theory instead of all phrasal combination
4. config file


future todo?? Depends on plans: 
1. seperate PPDB using .sen.pl to generate dataset [50-50] dataset. 50% supervised, 50% unsupervised.
2. merge with template builder ??