Add Fortran language support

.github/workflows/main.yml

@@ -26,6 +26,7 @@ jobs:
           - "c_sharp or visual_basic or f_sharp"
           - "r_lang"
           - "elixir"
+          - "fortran"
           - "dart or powershell"
     steps:
       - name: Checkout repository

Dockerfile

@@ -48,6 +48,7 @@ RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && \
         dotnet-sdk-6.0 \
         g++ \
         gcc \
+        gfortran \
         git \
         golang-go \
         haskell-platform \

README.md

@@ -40,6 +40,7 @@ make docker-build docker-pre-pr
 - C#
 - Dart
 - F#
+- Fortran90
 - Go
 - Haskell
 - Java

generated_code_examples/fortran/classification/decision_tree.f90

@@ -0,0 +1,29 @@
+module Model
+    implicit none
+contains
+    function score(input)
+        implicit none
+        double precision, dimension(:) :: input
+        double precision, dimension(3) :: var0
+        double precision, dimension(3) :: score
+        if (input(3) <= 2.449999988079071d0) then
+            var0 = (/ 1.0d0, 0.0d0, 0.0d0 /)
+        else
+            if (input(4) <= 1.75d0) then
+                if (input(3) <= 4.950000047683716d0) then
+                    if (input(4) <= 1.6500000357627869d0) then
+                        var0 = (/ 0.0d0, 1.0d0, 0.0d0 /)
+                    else
+                        var0 = (/ 0.0d0, 0.0d0, 1.0d0 /)
+                    end if
+                else
+                    var0 = (/ 0.0d0, 0.3333333333333333d0, 0.6666666666666666d0 /)
+                end if
+            else
+                var0 = (/ 0.0d0, 0.021739130434782608d0, 0.9782608695652174d0 /)
+            end if
+        end if
+        score(:) = var0
+        return
+    end function score
+end module Model

generated_code_examples/fortran/classification/lightgbm.f90

@@ -0,0 +1,111 @@
+module Model
+    implicit none
+contains
+    function score(input)
+        implicit none
+        double precision, dimension(:) :: input
+        double precision :: var0
+        double precision :: var1
+        double precision :: var2
+        double precision :: var3
+        double precision :: var4
+        double precision :: var5
+        double precision, dimension(3) :: score
+        if (input(3) > 3.1500000000000004d0) then
+            var0 = -1.1986122886681099d0
+        else
+            if (input(2) > 3.35d0) then
+                var0 = -0.8986122886681098d0
+            else
+                var0 = -0.9136122886681098d0
+            end if
+        end if
+        if (input(3) > 3.1500000000000004d0) then
+            if (input(3) > 4.450000000000001d0) then
+                var1 = -0.09503010837903424d0
+            else
+                var1 = -0.09563272415214283d0
+            end if
+        else
+            if (input(2) > 3.35d0) then
+                var1 = 0.16640323607832397d0
+            else
+                var1 = 0.15374604217339707d0
+            end if
+        end if
+        if (input(3) > 1.8d0) then
+            if (input(4) > 1.6500000000000001d0) then
+                var2 = -1.2055899476674514d0
+            else
+                var2 = -0.9500445227622534d0
+            end if
+        else
+            var2 = -1.2182214705715104d0
+        end if
+        if (input(4) > 0.45000000000000007d0) then
+            if (input(4) > 1.6500000000000001d0) then
+                var3 = -0.08146437273923739d0
+            else
+                var3 = 0.14244886188108738d0
+            end if
+        else
+            if (input(3) > 1.4500000000000002d0) then
+                var3 = -0.0950888159264695d0
+            else
+                var3 = -0.09438233722389686d0
+            end if
+        end if
+        if (input(4) > 1.6500000000000001d0) then
+            if (input(3) > 5.3500000000000005d0) then
+                var4 = -0.8824095771015287d0
+            else
+                var4 = -0.9121126703829481d0
+            end if
+        else
+            if (input(3) > 4.450000000000001d0) then
+                var4 = -1.1277829563828181d0
+            else
+                var4 = -1.1794405099157212d0
+            end if
+        end if
+        if (input(3) > 4.750000000000001d0) then
+            if (input(3) > 5.150000000000001d0) then
+                var5 = 0.16625543464258166d0
+            else
+                var5 = 0.09608601737074281d0
+            end if
+        else
+            if (input(1) > 4.950000000000001d0) then
+                var5 = -0.09644547407948921d0
+            else
+                var5 = -0.08181864271444342d0
+            end if
+        end if
+        score(:) = SOFTMAX((/ var0 + var1, var2 + var3, var4 + var5 /))
+        return
+    end function score
+    function softmax(x) result(res)
+        implicit none
+        double precision, dimension(:), intent(in) :: x
+        double precision, dimension(size(x)) :: res
+        double precision :: max_val, sum_val
+        integer :: i
+
+        ! Find maximum value in x
+        max_val = x(1)
+        do i = 2, size(x)
+            if (x(i) > max_val) then
+                max_val = x(i)
+            end if
+        end do
+
+        ! Compute softmax values
+        sum_val = 0.0d0
+        do i = 1, size(x)
+            res(i) = exp(x(i) - max_val)
+            sum_val = sum_val + res(i)
+        end do
+        res = res / sum_val
+
+    end function softmax
+end module Model

generated_code_examples/fortran/classification/linear.f90

@@ -0,0 +1,12 @@
+module Model
+    implicit none
+contains
+    function score(input)
+        implicit none
+        double precision, dimension(:) :: input
+
+        double precision, dimension(3) :: score
+        score(:) = (/ 9.700311953536998d0 + input(1) * -0.4128360473754751d0 + input(2) * 0.9680426131053453d0 + input(3) * -2.498310603183548d0 + input(4) * -1.0723230787022542d0, 2.1575759475871163d0 + input(1) * 0.5400806228605453d0 + input(2) * -0.3245383349519669d0 + input(3) * -0.2034493200950831d0 + input(4) * -0.9338183426196143d0, -11.857887901124615d0 + input(1) * -0.12724457548509432d0 + input(2) * -0.6435042781533917d0 + input(3) * 2.7017599232786216d0 + input(4) * 2.006141421321863d0 /)
+        return
+    end function score
+end module Model

generated_code_examples/fortran/classification/random_forest.f90

@@ -0,0 +1,71 @@
+module Model
+    implicit none
+contains
+    function score(input)
+        implicit none
+        double precision, dimension(:) :: input
+        double precision, dimension(3) :: var0
+        double precision, dimension(3) :: var1
+        double precision, dimension(3) :: score
+        if (input(4) <= 0.75d0) then
+            var0 = (/ 1.0d0, 0.0d0, 0.0d0 /)
+        else
+            if (input(3) <= 4.75d0) then
+                var0 = (/ 0.0d0, 1.0d0, 0.0d0 /)
+            else
+                if (input(3) <= 5.049999952316284d0) then
+                    if (input(4) <= 1.75d0) then
+                        var0 = (/ 0.0d0, 0.8333333333333334d0, 0.16666666666666666d0 /)
+                    else
+                        var0 = (/ 0.0d0, 0.08333333333333333d0, 0.9166666666666666d0 /)
+                    end if
+                else
+                    var0 = (/ 0.0d0, 0.0d0, 1.0d0 /)
+                end if
+            end if
+        end if
+        if (input(4) <= 0.800000011920929d0) then
+            var1 = (/ 1.0d0, 0.0d0, 0.0d0 /)
+        else
+            if (input(1) <= 6.25d0) then
+                if (input(3) <= 4.8500001430511475d0) then
+                    var1 = (/ 0.0d0, 0.9487179487179487d0, 0.05128205128205128d0 /)
+                else
+                    var1 = (/ 0.0d0, 0.0d0, 1.0d0 /)
+                end if
+            else
+                if (input(4) <= 1.550000011920929d0) then
+                    var1 = (/ 0.0d0, 0.8333333333333334d0, 0.16666666666666666d0 /)
+                else
+                    var1 = (/ 0.0d0, 0.02564102564102564d0, 0.9743589743589743d0 /)
+                end if
+            end if
+        end if
+        score(:) = mul_vector_number(add_vectors(var0, var1), 0.5d0)
+        return
+    end function score
+    function add_vectors(v1, v2) result(res)
+        implicit none
+        double precision, dimension(:), intent(in) :: v1, v2
+        double precision, dimension(size(v1)) :: res
+        integer :: i
+
+        do i = 1, size(v1)
+            res(i) = v1(i) + v2(i)
+        end do
+
+    end function add_vectors
+
+    function mul_vector_number(v1, num) result(res)
+        implicit none
+        double precision, dimension(:), intent(in) :: v1
+        double precision, intent(in) :: num
+        double precision, dimension(size(v1)) :: res
+        integer :: i
+
+        do i = 1, size(v1)
+            res(i) = v1(i) * num
+        end do
+
+    end function mul_vector_number
+end module Model

generated_code_examples/fortran/classification/svm.f90

@@ -0,0 +1,67 @@
+module Model
+    implicit none
+contains
+    function score(input)
+        implicit none
+        double precision, dimension(:) :: input
+        double precision :: var0
+        double precision :: var1
+        double precision :: var2
+        double precision :: var3
+        double precision :: var4
+        double precision :: var5
+        double precision :: var6
+        double precision :: var7
+        double precision :: var8
+        double precision :: var9
+        double precision :: var10
+        double precision :: var11
+        double precision :: var12
+        double precision :: var13
+        double precision :: var14
+        double precision :: var15
+        double precision :: var16
+        double precision :: var17
+        double precision :: var18
+        double precision :: var19
+        double precision :: var20
+        double precision :: var21
+        double precision :: var22
+        double precision :: var23
+        double precision :: var24
+        double precision :: var25
+        double precision :: var26
+        double precision :: var27
+        double precision, dimension(3) :: score
+        var0 = EXP(-0.06389634699048878d0 * ((5.1d0 - input(1)) ** 2.0d0 + (2.5d0 - input(2)) ** 2.0d0 + (3.0d0 - input(3)) ** 2.0d0 + (1.1d0 - input(4)) ** 2.0d0))
+        var1 = EXP(-0.06389634699048878d0 * ((4.9d0 - input(1)) ** 2.0d0 + (2.4d0 - input(2)) ** 2.0d0 + (3.3d0 - input(3)) ** 2.0d0 + (1.0d0 - input(4)) ** 2.0d0))
+        var2 = EXP(-0.06389634699048878d0 * ((6.3d0 - input(1)) ** 2.0d0 + (2.5d0 - input(2)) ** 2.0d0 + (4.9d0 - input(3)) ** 2.0d0 + (1.5d0 - input(4)) ** 2.0d0))
+        var3 = EXP(-0.06389634699048878d0 * ((5.4d0 - input(1)) ** 2.0d0 + (3.0d0 - input(2)) ** 2.0d0 + (4.5d0 - input(3)) ** 2.0d0 + (1.5d0 - input(4)) ** 2.0d0))
+        var4 = EXP(-0.06389634699048878d0 * ((6.2d0 - input(1)) ** 2.0d0 + (2.2d0 - input(2)) ** 2.0d0 + (4.5d0 - input(3)) ** 2.0d0 + (1.5d0 - input(4)) ** 2.0d0))
+        var5 = EXP(-0.06389634699048878d0 * ((5.6d0 - input(1)) ** 2.0d0 + (2.9d0 - input(2)) ** 2.0d0 + (3.6d0 - input(3)) ** 2.0d0 + (1.3d0 - input(4)) ** 2.0d0))
+        var6 = EXP(-0.06389634699048878d0 * ((6.7d0 - input(1)) ** 2.0d0 + (3.0d0 - input(2)) ** 2.0d0 + (5.0d0 - input(3)) ** 2.0d0 + (1.7d0 - input(4)) ** 2.0d0))
+        var7 = EXP(-0.06389634699048878d0 * ((5.0d0 - input(1)) ** 2.0d0 + (2.3d0 - input(2)) ** 2.0d0 + (3.3d0 - input(3)) ** 2.0d0 + (1.0d0 - input(4)) ** 2.0d0))
+        var8 = EXP(-0.06389634699048878d0 * ((6.0d0 - input(1)) ** 2.0d0 + (2.7d0 - input(2)) ** 2.0d0 + (5.1d0 - input(3)) ** 2.0d0 + (1.6d0 - input(4)) ** 2.0d0))
+        var9 = EXP(-0.06389634699048878d0 * ((5.9d0 - input(1)) ** 2.0d0 + (3.2d0 - input(2)) ** 2.0d0 + (4.8d0 - input(3)) ** 2.0d0 + (1.8d0 - input(4)) ** 2.0d0))
+        var10 = EXP(-0.06389634699048878d0 * ((5.7d0 - input(1)) ** 2.0d0 + (2.6d0 - input(2)) ** 2.0d0 + (3.5d0 - input(3)) ** 2.0d0 + (1.0d0 - input(4)) ** 2.0d0))
+        var11 = EXP(-0.06389634699048878d0 * ((5.0d0 - input(1)) ** 2.0d0 + (3.0d0 - input(2)) ** 2.0d0 + (1.6d0 - input(3)) ** 2.0d0 + (0.2d0 - input(4)) ** 2.0d0))
+        var12 = EXP(-0.06389634699048878d0 * ((5.4d0 - input(1)) ** 2.0d0 + (3.4d0 - input(2)) ** 2.0d0 + (1.7d0 - input(3)) ** 2.0d0 + (0.2d0 - input(4)) ** 2.0d0))
+        var13 = EXP(-0.06389634699048878d0 * ((5.7d0 - input(1)) ** 2.0d0 + (3.8d0 - input(2)) ** 2.0d0 + (1.7d0 - input(3)) ** 2.0d0 + (0.3d0 - input(4)) ** 2.0d0))
+        var14 = EXP(-0.06389634699048878d0 * ((4.8d0 - input(1)) ** 2.0d0 + (3.4d0 - input(2)) ** 2.0d0 + (1.9d0 - input(3)) ** 2.0d0 + (0.2d0 - input(4)) ** 2.0d0))
+        var15 = EXP(-0.06389634699048878d0 * ((4.5d0 - input(1)) ** 2.0d0 + (2.3d0 - input(2)) ** 2.0d0 + (1.3d0 - input(3)) ** 2.0d0 + (0.3d0 - input(4)) ** 2.0d0))
+        var16 = EXP(-0.06389634699048878d0 * ((5.7d0 - input(1)) ** 2.0d0 + (4.4d0 - input(2)) ** 2.0d0 + (1.5d0 - input(3)) ** 2.0d0 + (0.4d0 - input(4)) ** 2.0d0))
+        var17 = EXP(-0.06389634699048878d0 * ((5.1d0 - input(1)) ** 2.0d0 + (3.8d0 - input(2)) ** 2.0d0 + (1.9d0 - input(3)) ** 2.0d0 + (0.4d0 - input(4)) ** 2.0d0))
+        var18 = EXP(-0.06389634699048878d0 * ((5.1d0 - input(1)) ** 2.0d0 + (3.3d0 - input(2)) ** 2.0d0 + (1.7d0 - input(3)) ** 2.0d0 + (0.5d0 - input(4)) ** 2.0d0))
+        var19 = EXP(-0.06389634699048878d0 * ((6.2d0 - input(1)) ** 2.0d0 + (2.8d0 - input(2)) ** 2.0d0 + (4.8d0 - input(3)) ** 2.0d0 + (1.8d0 - input(4)) ** 2.0d0))
+        var20 = EXP(-0.06389634699048878d0 * ((7.2d0 - input(1)) ** 2.0d0 + (3.0d0 - input(2)) ** 2.0d0 + (5.8d0 - input(3)) ** 2.0d0 + (1.6d0 - input(4)) ** 2.0d0))
+        var21 = EXP(-0.06389634699048878d0 * ((6.1d0 - input(1)) ** 2.0d0 + (3.0d0 - input(2)) ** 2.0d0 + (4.9d0 - input(3)) ** 2.0d0 + (1.8d0 - input(4)) ** 2.0d0))
+        var22 = EXP(-0.06389634699048878d0 * ((6.0d0 - input(1)) ** 2.0d0 + (3.0d0 - input(2)) ** 2.0d0 + (4.8d0 - input(3)) ** 2.0d0 + (1.8d0 - input(4)) ** 2.0d0))
+        var23 = EXP(-0.06389634699048878d0 * ((4.9d0 - input(1)) ** 2.0d0 + (2.5d0 - input(2)) ** 2.0d0 + (4.5d0 - input(3)) ** 2.0d0 + (1.7d0 - input(4)) ** 2.0d0))
+        var24 = EXP(-0.06389634699048878d0 * ((7.9d0 - input(1)) ** 2.0d0 + (3.8d0 - input(2)) ** 2.0d0 + (6.4d0 - input(3)) ** 2.0d0 + (2.0d0 - input(4)) ** 2.0d0))
+        var25 = EXP(-0.06389634699048878d0 * ((5.6d0 - input(1)) ** 2.0d0 + (2.8d0 - input(2)) ** 2.0d0 + (4.9d0 - input(3)) ** 2.0d0 + (2.0d0 - input(4)) ** 2.0d0))
+        var26 = EXP(-0.06389634699048878d0 * ((6.0d0 - input(1)) ** 2.0d0 + (2.2d0 - input(2)) ** 2.0d0 + (5.0d0 - input(3)) ** 2.0d0 + (1.5d0 - input(4)) ** 2.0d0))
+        var27 = EXP(-0.06389634699048878d0 * ((6.3d0 - input(1)) ** 2.0d0 + (2.8d0 - input(2)) ** 2.0d0 + (5.1d0 - input(3)) ** 2.0d0 + (1.5d0 - input(4)) ** 2.0d0))
+        score(:) = (/ 0.11172510039290856d0 + var0 * -0.8898986041811555d0 + var1 * -0.8898986041811555d0 + var2 * -0.0d0 + var3 * -0.0d0 + var4 * -0.0d0 + var5 * -0.756413813553974d0 + var6 * -0.0d0 + var7 * -0.8898986041811555d0 + var8 * -0.0d0 + var9 * -0.0d0 + var10 * -0.8898986041811555d0 + var11 * 0.04218875216876044d0 + var12 * 0.7142250613852136d0 + var13 * 0.0d0 + var14 * 0.8898986041811555d0 + var15 * 0.8898986041811555d0 + var16 * 0.0d0 + var17 * 0.8898986041811555d0 + var18 * 0.8898986041811555d0, -0.04261957451303831d0 + var19 * -0.37953658977037247d0 + var20 * -0.0d0 + var21 * -0.0d0 + var22 * -0.37953658977037247d0 + var23 * -0.37953658977037247d0 + var24 * -0.26472396872040066d0 + var25 * -0.3745962010653211d0 + var26 * -0.10077618026650095d0 + var27 * -0.0d0 + var11 * 0.0d0 + var12 * 0.0d0 + var13 * 0.37953658977037247d0 + var14 * 0.37953658977037247d0 + var15 * 0.3044555865539922d0 + var16 * 0.05610417372785803d0 + var17 * 0.37953658977037247d0 + var18 * 0.37953658977037247d0, 1.8136162062461285d0 + var19 * -110.34516826676301d0 + var20 * -13.999391039896215d0 + var21 * -108.44329471899991d0 + var22 * -110.34516826676301d0 + var23 * -22.21095753342801d0 + var24 * -0.0d0 + var25 * -0.0d0 + var26 * -65.00217641452454d0 + var27 * -110.34516826676301d0 + var0 * 0.0d0 + var1 * 0.0d0 + var2 * 110.34516826676301d0 + var3 * 62.115561183470184d0 + var4 * 37.19509025661546d0 + var5 * 0.0d0 + var6 * 110.34516826676301d0 + var7 * 0.0d0 + var8 * 110.34516826676301d0 + var9 * 110.34516826676301d0 + var10 * 0.0d0 /)
+        return
+    end function score
+end module Model