This plugin collects data about your usage of code completion, anonymises and sends it to us. Note, the plugin doesn't send any of your personal data or code. All we send is just numbers calculated based on your completion usage patterns.
Ever since 0.1.0, the plugin builds an index with frequencies of all 2-3-4-5 grams over the AST of the project source code files. This can increase the indexing time of the source code and is disabled by default. The plugin will ask for permission for each opened project. If you change your mind and want to forbid the indexing, please switch off the "Enable NGram indexing for completion" option (Help - Find Action...). As usual, no source code or personal data will be collected.
General usage instructions
Code completion is one of the most useful features of an IDE. In IntelliJ IDEA we have more than one types of code completion: Basic Completion, Smart Completion, Second Completion,Second Smart Completion, Statement Completion, Hippie Completion. Each of them in turn uses different algorithms and heuristics. However, none of them uses real machine learning yet.
This is what we’d like to change as we believe machine learning may improve completion efficiency. In order to implement a proper machine learning algorithm, we need more data from you.