hydrat is a declarative framework for text classification tasks. Salient points:

• declarative, programmatic interface to data
• wrappers for a variety of external classifiers
• wrappers for a variety of text processing tools
• automatic extraction of a variety of textual features
• advanced data management via a structured representation and metadata system to avoid expensive recomputation
• storage and analysis of results

Please see

• Installation
• GettingStarted