Robert Feldt

Dynamic/online updating of graph?

1

It seems the main use case here is batch processing, i.e. compute the graph once and then use that as a fixed thing for NN queries. Is there some way...

NaN (or ArgumentError) from QGram distances for short strings

4

Some QGram distances today return NaN if one of the input strings are shorter than the q-gram length (q) while others don't: ```Julia julia> using StringDistances julia> isnan(Cosine(2)("", "bb")) true...

Cannot save model (weight vector + matrix) from CLI?

1

I can only see an argument to save the (test) predictions themselves but not for saving the trained model weights (the vector and matrix). Did I miss something or is...

set finalizer for XMLDocument by default

4

I propose most new users would rather prefer the finalizer is set by default. This does not solve the issue if people create their own XMLElement's but if they have...

I have a long string sl and an array of shorter strings S = [s1, s2, ..., sn] and need to know the length of the compressed strings one gets...

Plans to support Earley parsing?

2

Thanks for Lerche.jl, it looks great. The Python Lark supports also Earley(SPPF) parsing. Any plans to implement this also for Lerche.jl?

Stream data to html front end with VegaLite.jl generated graph?

1

Thanks for VegaLite.jl; it is superuseful. I see the upcoming feature of outputting html but is there also a plan for how we can stream data to the "front-end" and...