Merge pull request #37 from PharmCat/dev

Dev

Project.toml

@@ -1,7 +1,7 @@
 authors = ["Vladimir Arnautov ([email protected])"]
 name = "ClinicalTrialUtilities"
 uuid = "535c2557-d7d0-564d-8ff9-4ae146c18cfe"
-version = "0.2.2"
+version = "0.2.3"
 
 [deps]
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
@@ -12,6 +12,8 @@ QuadGK = "1fd47b50-473d-5c70-9696-f719f8f3bcdc"
 SpecialFunctions = "276daf66-3868-5448-9aa4-cd146d93841b"
 Roots = "f2b01f46-fcfa-551c-844a-d8ac1e96c665"
 RecipesBase = "3cdcf5f2-1ef4-517c-9805-6587b60abb01"
+Reexport = "189a3867-3050-52da-a836-e630ba90ab69"
+CSV = "336ed68f-0bac-5ca0-87d4-7b16caf5d00b"
 
 [extras]
 CSV = "336ed68f-0bac-5ca0-87d4-7b16caf5d00b"
@@ -25,9 +27,10 @@ test = ["CSV", "Test", "Plots"]
 julia = "1.0, 1.1, 1.2, 1.3"
 Distributions = "0.16, 0.17, 0.18, 0.19, 0.20, 0.21, 0.22"
 StatsBase = "0.22, 0.23, 0.24, 0.25, 0.26, 0.27, 0.28, 0.29, 0.30, 0.31, 0.32"
-DataFrames = "0.19"
+DataFrames = "0.19, 0.20"
 QuadGK = "2.0, 2.1"
 SpecialFunctions = "0.8, 0.9"
 Roots = "0.7, 0.8"
 CSV = "0.5"
 RecipesBase = "0.7"
+Reexport = "0.1, 0.2"

cange.log

@@ -1,3 +1,12 @@
+v0.2.3
+
+  - ConTab struct
+  - Urine PK
+  - NCA Tau PK partial areas fix
+  - NCA test update
+  - Plotting improve
+  - minor fix
+
 
 v0.2.2
   - Distributions bump
@@ -8,7 +17,6 @@ v0.2.2
   - fisher & pirson non-public methods for frequencies
   - add Show methods
   - minor fix
-  - cosmetics
   - add sim
   - cosmetics
 

src/ClinicalTrialUtilities.jl

@@ -13,17 +13,21 @@
 # If you want to check and get R code you can find some here: http://powerandsamplesize.com/Calculators/
 __precompile__(true)
 module ClinicalTrialUtilities
-using Distributions, StatsBase, Random, Roots, QuadGK, DataFrames, RecipesBase
+
+using Distributions, Random, Roots, QuadGK, RecipesBase, Reexport
+
+@reexport using StatsBase
+
 import SpecialFunctions
 import Base.show
 import Base.showerror
 import Base.getindex
 import Base.length
 import Base.in
 import Base.iterate
+import Base.eltype
 import StatsBase.confint
-import DataFrames.DataFrame
-
+import DataFrames: DataFrame, DataFrames, names!, unstack, deleterows!
 
 function lgamma(x)
     return SpecialFunctions.logabsgamma(x)[1]

src/dataset.jl

@@ -7,3 +7,6 @@ end
 function Base.iterate(iter::DataSet, i::Int)
     return Base.iterate(iter.data, i)
 end
+function Base.eltype(obj::DataSet)
+    return eltype(obj.data)
+end

src/descriptives.jl

@@ -183,13 +183,21 @@ end
 
 #Statistics calculation
 
-@inline function descriptive_(data::Array{T,1}, stats::Union{Tuple{Vararg{Symbol}}, Array{Symbol,1}}) where T <: Real
+function notnan(x)
+    return !(x === NaN || x === nothing || x === missing)
+end
+
+@inline function descriptive_(data::Vector{T}, stats::Union{Tuple{Vararg{Symbol}}, Array{Symbol,1}}) where T <: Real
 
+    #=
     dlist = findall(x -> x === NaN || x === nothing || x === missing, data)
     if length(dlist) > 0
         data = copy(data)
         deleteat!(data, dlist)
     end
+    =#
+
+    data       = data[notnan.(data)]
 
     dn         = nothing
     dmin       = nothing

src/freque.jl

@@ -1,4 +1,26 @@
 
+
+struct ConTab{Int32, Int32}
+    tab::Matrix{Int}
+    row::Vector
+    col::Vector
+
+    function ConTab(m::Matrix{Int})
+        row = Vector{Symbol}(undef, size(m, 1)) .= Symbol("")
+        col = Vector{Symbol}(undef, size(m, 2)) .= Symbol("")
+        new{size(m, 1), size(m, 2)}(m, row, col)
+    end
+    function ConTab(m::Matrix{Int}, row, col)
+        new{size(m, 1), size(m, 2)}(m, row, col)
+    end
+end
+
+struct Freque
+    val
+    n
+end
+
+
 function freque(data::DataFrame; vars::Symbol, alpha = 0.05)::DataFrame
     result = DataFrame(value = Any[], n = Int[], p = Float64[], cil = Float64[], ciu = Float64[])
     list   = unique(data[:, vars])
@@ -12,46 +34,56 @@ function freque(data::DataFrame; vars::Symbol, alpha = 0.05)::DataFrame
     return result
 end
 
-function contab(data::DataFrame; row::Symbol, col::Symbol)::Matrix
+function contab(data::DataFrame; row::Symbol, col::Symbol)::ConTab
     clist = unique(data[:, col])
     rlist = unique(data[:, row])
     cn    = length(clist)
     rn    = length(rlist)
-    dfs   = Array{Int, 2}(undef, rn, cn)
+    dfs   = Matrix{Int}(undef, rn, cn)
     for ri = 1:rn
         rowl  = data[data[:, row] .== rlist[ri], col]
         for ci = 1:cn
             cnt = count(x -> x == clist[ci], rowl)
             dfs[ri, ci] = cnt
         end
     end
-    return dfs
+    return ConTab(dfs, rlist, clist)
 end
 
 
 function pirson(a::Matrix{Int})
-    n = length(a[:,1])
-    m = length(a[1,:])
-    tm = sum(a, dims=1)[1,:]
-    tn = sum(a, dims=2)[:,1]
+    n   = length(a[:,1])
+    m   = length(a[1,:])
+    tm  = sum(a, dims=1)[1,:]
+    tn  = sum(a, dims=2)[:,1]
     num = sum(tm)
-    ae = Array{Real, 2}(undef, n, m)
+    ae  = Array{Real, 2}(undef, n, m)
     for im = 1:m
         for in = 1:n
             ae[in, im] = tn[in]*tm[im]/num
         end
     end
-    chsq = sum(((a .- ae) .^2 ) ./ ae)
+    chsq  = sum(((a .- ae) .^2 ) ./ ae)
     chsqy = sum(((abs.((a .- ae)) .- 0.5) .^2 ) ./ ae)
-    ml = 2 * sum( a .* log.( a ./ ae ))
-    df   = (n - 1)*(m - 1)
-    ϕ = sqrt(chsq / (num*(n - 1)*(m - 1)))
-    C = sqrt(chsq/(chsq+num))
-    K = sqrt(chsq/num/sqrt((n - 1)*(m - 1)))
+    ml    = 2 * sum( a .* log.( a ./ ae ))
+    df    = (n - 1)*(m - 1)
+    ϕ     = sqrt(chsq / (num*(n - 1)*(m - 1)))
+    C     = sqrt(chsq/(chsq+num))
+    K     = sqrt(chsq/num/sqrt((n - 1)*(m - 1)))
     return chsq, chsqy, ml, 1-cdf(Chisq(df), chsq)
 end
 
 function fisher(a::Matrix{Int})
     dist  = Hypergeometric(sum(a[1, :]), sum(a[2, :]), sum(a[:, 1]))
     value = min(2 * min(cdf(dist, a[1, 1]), ccdf(dist, a[1, 1])), 1.0)
 end
+
+
+function fisher(t::ConTab{2, 2})
+    fisher(t.tab)
+end
+
+#=
+function StatsBase.confint(t::ConTab{2, 2})
+end
+=#