MAGPIE: A sense-annotated corpus of potentially idiomatic expressions

Idiom Principle on MAGPIE dataset

forked from bloa/magpie.git amd jimbou/magpie enhacenments. Extended with UMDA search algorithm, corpuses, scenarios and results.

Using Databricks and Apache Spark (PySpark) to perform distributed processing on the Wikipedia corpus (54MB, 400,000+ entries), combined with the MAGPIE idiom corpus for NLP and semantic structure analysis.