Utility modules

• cleartk-test-util: test case base classes and utilities for testing licenses, parameter names, etc.
• cleartk-util: simple type-system agnostic readers and writers, and various utilities used by other ClearTK modules

Machine learning modules

• cleartk-ml: the core machine learning APIs: classifiers, features, instances, feature extractors, feature encoders, etc.
• cleartk-ml-opennlp-maxent: wrappers to use OpenNLP-maxent as a ClearTK classifier
• cleartk-ml-svmlight: wrappers to use SVM-light as a ClearTK classifier
• cleartk-ml-tksvmlight: wrappers to use Tree Kernels in SVM-light as a ClearTK classifier
• cleartk-ml-libsvm: wrappers to use LIBSVM as a ClearTK classifier
• cleartk-ml-mallet: wrappers to use MALLET and GRMM as ClearTK classifiers

Evaluation module

• cleartk-eval: classes for evaluating pipelines - train/test, cross-validation, etc.

Type system

• cleartk-type-system: the official UIMA type system for ClearTK

Wrappers for external components

• cleartk-snowball: a wrapper around the Snowball stemmer
• cleartk-opennlp-tools: wrappers around the OpenNLP sentence segmenter, part-of-speech tagger, and syntactic parser
• cleartk-berkeleyparser: a wrapper around the Berkeley syntactic parser
• cleartk-clearparser: a wrapper around the Clear dependency parser, stemmer, and semantic role-labeler. Requires models available for download here.
• cleartk-maltparser: a wrapper around the Malt dependency parser
• cleartk-stanford-corenlp: a wrapper around the Stanford CoreNLP sentence segmenter, tokenizer, part-of-speech tagger, named-entity tagger, syntactic parser, dependency parser and coreference resolution system.

Home-grown components

• cleartk-token: sentence segmenters, tokenizers and part-of-speech taggers, including a segmenter based on java.text.BreakIterator, a PennTreebankTokenizer, and readers for Genia part of speech tags.
• cleartk-named-entity: readers and writers for named-entity tagged corpora, e.g. CoNLL 2003 and ACE 2005.
• cleartk-syntax: readers for Penn Treebank format trees, utility classes for working with constituent and dependency tree annotations, and feature extractors for tree paths, etc.
• cleartk-semantic-roles: readers for PropBank and CoNLL 2005, and an incomplete implementation of a semantic-role labeler using ClearTK classifiers.
• cleartk-timeml: readers and writers for TimeBank and TempEval, along with state-of-the-art models trained on these corpora.
• cleartk-examples: examples of using ClearTK

Official release module

• cleartk-release: lists the official versions of ClearTK modules for the current release - ensures that all dependencies declared in ClearTK are compatible.