Multi-level clustering has been added, which allows you to distribute phrases into groups within already formed clusters to the full depth of cluster nesting.
Added the ability during clustering to set phrases that need to be perceived as a whole, without separating by words.
Improved clustering by conditions.
Added a better morphology for clustering for Russian and English.
Improved the algorithm for distributing phrases across clusters.
Added automatic detection of headers in CSV files.
Improved file cutting, significantly reduced the error in determining the required file size when cutting by the number of files.
The algorithm for searching for E-Mail addresses inside any files has been optimized.
Improved multithreaded file processing.
The cluster saving system has been improved, a format for importing clusters and tags into KeyCollector has been added.
