drop 1.6

.github/workflows/Tier1.yml

@@ -30,7 +30,6 @@ jobs:
     strategy:
       matrix:
         version:
-          - '1.6'
           - '1.8'
           - '1'
         os:

docs/src/index.md

@@ -38,6 +38,10 @@ Implemented covariance structures:
 
 Actually Metida can fit datasets with wore than  160k observation and 40k subjects levels on PC with 64 GB RAM. This is not "hard-coded" limitation, but depends on your model and data structure. Fitting of big datasets can take a lot of time. Optimal dataset size is less than 100k observations with maximum length of block less than 400.
 
+!!! note
+
+    Julia v1.8 or higher required.
+
 ## Contents
 
 ```@contents

src/Metida.jl

@@ -3,7 +3,6 @@
 
 __precompile__()
 module Metida
-using Compat
 using ProgressMeter, LinearAlgebra, ForwardDiff, DiffResults, Random, Optim, LineSearches, MetidaBase#, SparseArrays#, Polyester#, LoopVectorization
 import StatsBase, StatsModels, Distributions
 

src/varstruct.jl

@@ -263,7 +263,7 @@ struct CovStructure{T, T2} <: AbstractCovarianceStructure
                 if isa(random[i].model, ConstantTerm) # if only ConstantTerm in the model - data_ - first is collumn (responce)
                     data_     = data[[first(keys(data))]] 
                 else
-                    data_     = data[Tuple(StatsModels.termvars(random[i].model))] # only collumns for model
+                    data_     = data[StatsModels.termvars(random[i].model)] # only collumns for model
                 end
                 if isa(random[i].covtype.s, ZERO)
                     schema[i] = InterceptTerm{false}()
@@ -306,7 +306,7 @@ struct CovStructure{T, T2} <: AbstractCovarianceStructure
             if isa(repeated[i].model, ConstantTerm) # if only ConstantTerm in the model - data_ - first is collumn (responce)
                 data_     = data[[first(keys(data))]] 
             else
-                data_     = data[Tuple(StatsModels.termvars(repeated[i].model))] # only collumns for model
+                data_     = data[StatsModels.termvars(repeated[i].model)] # only collumns for model
             end
 
             schema[rn + i] = apply_schema(repeated[i].model, StatsModels.schema(data_, repeated[i].coding))