hermanschaaf / go-mafan Goto Github PK

Right now the tree takes about 1s to build, which is not much better than the Python version. This is Go. We can do better.

It would be nice to add a wordwrap function for Chinese text.

Always more tests! Also, Mt Fuji is pretty from here.

It would be useful to experiment with possible space optimizations, both in terms of file size of the chinese dictionary file and in terms of memory usage while the program is running. Both these metrics should be minimized as far as possible.

For better prediction

Order words by decreasing frequency and use Mandelbrot's adapted version of zipf's law to save on space and memory usage.