Store ML training data in a separate HDF5 file

src/instance/file_instance.jl

@@ -15,7 +15,7 @@ mutable struct FileInstance <: Instance
     function FileInstance(filename::AbstractString, build_model::Function)::FileInstance
         instance = new(nothing, nothing, filename, nothing, build_model)
         instance.py = PyFileInstance(instance)
-        instance.h5 = Hdf5Sample(filename)
+        instance.h5 = Hdf5Sample("$filename.h5", mode="a")
         instance.filename = filename
         return instance
     end

test/instance/file_instance_test.jl

@@ -15,7 +15,7 @@ using Cbc
         solver = LearningSolver(Cbc.Optimizer)
         solve!(solver, instance)
 
-        h5 = Hdf5Sample(filename)
+        h5 = Hdf5Sample("$filename.h5")
         @test h5.get_scalar("mip_wallclock_time") > 0
     end
 

test/instance/jump_instance_test.jl

@@ -56,7 +56,7 @@ end
         instance = FileInstance(filename, build_model)
         solver = LearningSolver(Cbc.Optimizer)
         solve!(solver, instance)
-        h5 = MIPLearn.Hdf5Sample(filename)
+        h5 = MIPLearn.Hdf5Sample("$filename.h5")
         @test h5.get_array("mip_var_values") == [1.0, 0.0]
     end
 end