chipkent

from deephaven.TableTools import emptyTable

t = emptyTable(10).update("X=i")

chipkent

# add JDK to path (otherwise jnius gives DLL load error)
import os
os.environ['PATH'] = os.environ['PATH'] + ";C:\\Program Files\\Java\jdk1.8.0_72\\jre\\bin\\server"
os.environ['PATH'] = os.environ['PATH'] + ";C:\\Program Files\\Java\jdk1.8.0_60\\jre\\bin\\server"
print(os.environ['PATH'])

# add jpy to the python path
import sys
if sys.version_info.major == 2 and sys.version_info.minor == 7:

chipkent

using Distributions
using BenchmarkTools

function lratio2{FT}(dd::Dict{FT,Tuple{Float64, ContinuousUnivariateDistribution}})

  num::Float64 = 0.0
  denom::Float64 = 0

  for (k, v) in dd

chipkent

using NamedArrays

function compute_betas{T <: AbstractFloat}(factors::AbstractArray{T,2}, returns::AbstractArray{T,2})
    # make sure the dimension labels match
    if typeof(factors) <: NamedArrays.NamedArray && typeof(returns) <: NamedArrays.NamedArray
        @assert NamedArrays.names(factors,1) == NamedArrays.names(returns,1)
        @assert factors.dimnames[1] == returns.dimnames[1]
    end

chipkent

using DataFrames

symbol = ["A","A","A","A","A","B","B","B","B","B"]
date = ["d1","d2","d3","d4","d5","d1","d2","d3","d4","d5"]
value = Float64[1,2,3,4,5,6,7,8,9,10]

df = DataFrame(S=symbol,D=date,V=value)
df[2,3] = NA
df[8,3] = NA

chipkent

using NamedArrays

function compute_covariance{T <: AbstractFloat}( M::AbstractArray{T,2}, I::AbstractArray{T,2})
    return make_positive_definite( (M'*M) ./ (I'*I) )
end

make_positive_definite{T <: AbstractFloat}( m::AbstractArray{T,2}, epsilon::Real=1e-8 ) = nearPSD(m,epsilon)

nearPSD{T <: AbstractFloat}( A::NamedArrays.NamedArray{T,2}, epsilon::Real=1e-12 ) =

chipkent

module XYZ

export
    SERDataStore

abstract AbstractDataStore{K,V} <: Associative{K,V}

immutable SERDataStore{K<:AbstractString,V} <: AbstractDataStore{K,V}
    rootdir::AbstractString #TODO rename