Generate metadata for each document in a dataset
Develop a pipeline to create a custom sentiment dictionary with contrast analysis
