v0.6.0

Project.toml

@@ -1,7 +1,7 @@
 name = "ClinicalTrialUtilities"
 uuid = "535c2557-d7d0-564d-8ff9-4ae146c18cfe"
 authors = ["Vladimir Arnautov ([email protected])"]
-version = "0.5.1"
+version = "0.6.0"
 
 [deps]
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"

change.log

@@ -1,9 +1,21 @@
+v0.6.0
+  - changes in PK parameters names
+  - nca! administration settings
+  - adjust kel range for nca! with adm = :iv
+  - MRTtauinf used for Vsstau calculation
+  - changes in plots
+  - descriptive, CI add
+  - export fix
+  - DataSet sort!
+  - documentation
+  - nca tests
+
 v0.5.1
   - DataFrames 1
   - SpecialFunctions 1
 
 v0.5.0
-  - StableRNGs test
+  - StableRNGs test+
   - rng keyword for randomtable, randomseq
   - auc_sparse for PKSubject
 

docs/src/nca.md

@@ -8,6 +8,72 @@ NCA analysis based on following steps:
 4. Exporting to DataFrame;
 5. Descriptive statistics / HTML export.
 
+### Description
+
+```julia
+include("ncatable.jl")
+```
+#### AUC
+
+```math
+AUC = \sum_{n=1}^N AUC_{n}
+```
+
+Where AUCn - partial AUC.
+
+Linear trapezoidal rule
+
+```math
+AUC\mid_{t_1}^{t_2} = \delta t \times \frac{C_1 + C_2}{2}
+
+AUMC\mid_{t_1}^{t_2} = \delta t \times \frac{t_1 \times C_1 + t_2 \times C_2}{2}
+```
+
+Logarithmic trapezoidal rule
+
+```math
+AUC\mid_{t_1}^{t_2} =   \delta t \times \frac{ C_2 - C_1}{ln(C_2/C_1)}
+
+AUMC\mid_{t_1}^{t_2} = \delta t \times \frac{t_2 \times C_2 - t_1 \times C_1}{ln(C_2/C_1)} -  \delta t^2 \times \frac{ C_2 - C_1}{ln(C_2/C_1)^2}
+```
+
+Linear interpolation rule
+
+```math
+C_x = C_1 + \frac{(t_x-t_1)\times(C_2 - C_1)}{t_2 - t_1}
+```
+
+Logarithmic interpolation rule
+
+```math
+C_x = exp\left(ln(C_1) + \frac{(t_x-t_1)\times(ln(C_2) - ln(C_1))}{t_2 - t_1}\right)
+```
+
+#### \lambda_z - elimination constant
+
+#### HL
+
+```math
+HL = ln(2) / \lambda_z
+```
+
+#### AUCinf
+
+```math
+AUCinf = AUClast + \frac{Clast}{\lambda_z}
+```
+
+#### AUMCinf
+
+```math
+AUMCinf =  AUMClast + \frac{tlast\times Clast}{\lambda_z} + \frac{Clast}{\lambda_z^2}
+```
+
+#### Accumulation index
+
+```math
+Accind = \frac{1}{1 - exp(-\lambda_z \tau)}
+```
 
 ### nca!
 ```@docs

docs/src/ncatable.jl

@@ -0,0 +1 @@
+

src/dataset.jl

@@ -80,6 +80,11 @@ function Base.deleteat!(a::DataSet{T}, inds::Dict) where T
     return a
 end
 
+function Base.sort!(a::DataSet; by)
+    sort!(a.data; by = by)
+    a
+end
+
 ################################################################################
 #=
 function Tables.istable(table::DataSet)

src/descriptives.jl

@@ -128,10 +128,12 @@ Descriptive statistics.
 function descriptive(data;
     sort::Union{Symbol, Array{T,1}} = Array{Symbol,1}(undef,0),
     vars = [],
-    stats::Union{Symbol, Array{T,1}, Tuple{Vararg{Symbol}}} = :default)::DataSet{Descriptive} where T <: Union{Symbol, String}
+    stats::Union{Symbol, Array{T,1}, Tuple{Vararg{Symbol}}} = :default, level = 0.95)::DataSet{Descriptive} where T <: Union{Symbol, String}
 
     stats = checkstats(stats)
-
+    if isa(vars, UnitRange{Int64})
+        vars = names(data)[vars]
+    end
     if isa(vars, Array) && length(vars) == 0
         vars = filter(x -> x ∉ sort, names(data))
         del  = []
@@ -158,7 +160,7 @@ function descriptive(data;
         #pushvardescriptive!(d, vars, Matrix(data[:,vars]), Dict(), stats)
         mx = Matrix(data[:,vars])
         for v  = 1:length(vars)  #For each variable in list
-            push!(d, Descriptive(vars[v], nothing, Dict(:Variable => v), descriptive_(mx[:, v], stats)))
+            push!(d, Descriptive(vars[v], nothing, Dict(:Variable => v), descriptive_(mx[:, v], stats, level)))
         end
         return DataSet(d)
     end
@@ -173,30 +175,32 @@ function descriptive(data;
         for v  = 1:length(vars)  #For each variable in list
             dict = Dict{Symbol, Any}(sort .=> sortval)
             dict[:Variable]   =  vars[v]
-            push!(d, Descriptive(vars[v], nothing, dict, descriptive_(mx[:, v], stats)))
+            push!(d, Descriptive(vars[v], nothing, dict, descriptive_(mx[:, v], stats, level)))
         end
         #pushvardescriptive!(d, vars, mx, sortval, stats)  #push variable descriprives for mx
     end
     return DataSet(d)
 end
-function descriptive(data::Array{T, 1}; stats::Union{Symbol, Vector, Tuple} = :default, var = nothing, varname = nothing, sort = Dict())::Descriptive where T <: Real
+function descriptive(data::Array{T, 1}; stats::Union{Symbol, Vector, Tuple} = :default, var = nothing, varname = nothing, sort = Dict(), level = 0.95)::Descriptive where T <: Real
     stats = checkstats(stats)
-    return Descriptive(var, varname, sort, descriptive_(data, stats))
+    return Descriptive(var, varname, sort, descriptive_(data, stats, level))
 
 end
 
 """
 Check if all statistics in allstat list. return stats tuple
 """
 @inline function checkstats(stats::Union{Symbol, Array{T,1}, Tuple{Vararg{Symbol}}})::Tuple{Vararg{Symbol}} where T <: Union{Symbol, String}
-    allstat = (:n, :min, :max, :range, :mean, :var, :sd, :sem, :cv, :harmmean, :geomean, :geovar, :geosd, :geocv, :skew, :ses, :kurt, :sek, :uq, :median, :lq, :iqr, :mode)
+    allstat = (:n, :min, :max, :range, :mean, :var, :sd, :sem, :cv, :harmmean, :geomean, :geovar, :geosd, :geocv, :skew, :ses, :kurt, :sek, :uq, :median, :lq, :iqr, :mode, :meanci)
     if isa(stats, Symbol)
         if stats == :default stats = (:n, :mean, :sd, :sem, :uq, :median, :lq)
         elseif stats == :all stats = allstat
         else stats = Tuple(stats) end
     end
     stats = Tuple(Symbol.(stats))
-    if any(x -> x  ∉  allstat, stats) throw(ArgumentError("stats element not in allstat list")) end
+    if any(x -> x  ∉  allstat, stats)
+        throw(ArgumentError("stats element ($(findall(x -> x  ∉  allstat, stats))) not in allstat list"))
+    end
     return stats
 end
 """
@@ -237,7 +241,7 @@ 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
+@inline function descriptive_(data::Vector{T}, stats::Union{Tuple{Vararg{Symbol}}, Array{Symbol,1}}, level) where T <: Real
 
     #=
     dlist = findall(x -> x === NaN || x === nothing || x === missing, data)
@@ -395,6 +399,13 @@ end
             dict[s] = abs(duq-dlq)
         elseif s == :mode
             dict[s] = mode(data)
+        elseif s == :meanci
+            if dmean === nothing dmean = mean(data) end
+            if dvar === nothing dvar = var(data) end
+            if dn === nothing dn = length(data) end
+            e = sqrt(dvar/dn)*quantile(TDist(dn-1), 1-(1-level)/2)
+            dict[Symbol(string(s)*"L"*string(level))] = dmean-e
+            dict[Symbol(string(s)*"U"*string(level))] = dmean+e
         end
     end
     return dict

src/export.jl

@@ -196,7 +196,7 @@ HTLM export.
             while s
                 s = false
                 for r = 2:rown
-                    if tablematrix[r,c] !=0 && data[r,c] === data[r-1,c]
+                    if tablematrix[r,c] !=0 && !ismissing(data[r,c]) && !ismissing(data[r-1,c]) && data[r,c] == data[r-1,c]
                         tablematrix[r,c] -= 1;
                         tablematrix[r-1,c] += 1;
                         s = true;