Data update

This commit is contained in:
Ingy döt Net 2025-06-11 20:16:52 -04:00
parent 72eb4943cb
commit 4d5544505c
2347 changed files with 62432 additions and 16731 deletions

View file

@ -0,0 +1,121 @@
module knuth_s
implicit none
private
public :: s_of_n_creator, free_sample, sample_state, s_of_n
! Type to hold sample state
type :: sample_state
integer :: n
integer :: count
integer, allocatable :: sample(:)
end type sample_state
contains
! Subroutine to create the s_of_n sampling function
subroutine s_of_n_creator(n, state)
integer, intent(in) :: n
type(sample_state), pointer :: state
allocate (state)
state % n = n
state % count = 0
allocate (state % sample(n))
end subroutine s_of_n_creator
! Function to perform sampling
function s_of_n(state, item) result(current_sample)
type(sample_state), pointer :: state
integer, intent(in) :: item
integer, pointer :: current_sample(:)
real :: r
integer :: j
! Add item to sample if we haven't reached n yet
if (state % count < state % n) then
state % count = state % count + 1
state % sample(state % count) = item
else
! Knuth's Algorithm S: select with probability n/i
call random_number(r)
if (r < real(state % n) / real(state % count + 1)) then
! Randomly replace one of the existing items
call random_number(r)
j = int(r * state % n) + 1
state % sample(j) = item
end if
state % count = state % count + 1
end if
current_sample => state % sample
end function s_of_n
! Cleanup function to deallocate sample state
subroutine free_sample(state)
type(sample_state), pointer :: state
deallocate (state % sample)
deallocate (state)
end subroutine free_sample
end module knuth_s
program test_knuth_s
use knuth_s
implicit none
type(sample_state), pointer :: state
integer, pointer :: sample(:)
integer :: i, j, k, reps
integer :: frequencies(0:9)
character(len=50) :: sample_str
integer :: n, timex(8)
integer, allocatable :: seed(:)
call random_seed(size=n)
allocate (seed(n))
call date_and_time(values=timex)
seed(1) = timex(5) * 3600000 + timex(6) * 60000 + timex(7) * 1000 + timex(8)
do i = 2, n
seed(i) = seed(1) + 37 * (i - 1)
end do
! Initialize random seed
call random_seed(put=seed)
! Show one run's sampling progression
print *, 'Single run samples for n = 3:'
call s_of_n_creator(3, state)
do i = 0, 9
sample => s_of_n(state, i)
! Format sample output
sample_str = ''
write (sample_str, '("[", *(I0, ", "))') sample(1:min(state % count, 3))
if (state % count > 0) then
sample_str = trim(sample_str)
sample_str = sample_str(1:len_trim(sample_str) - 1)//']'
else
sample_str = '[]'
end if
print '(A, I0, A, A)', ' Item: ', i, ' -> sample: ', trim(sample_str)
end do
call free_sample(state)
! Initialize frequency counter
frequencies = 0
! Run 100,000 repetitions to count final sample frequencies
reps = 100000
do i = 1, reps
call s_of_n_creator(3, state)
do j = 0, 9
sample => s_of_n(state, j)
end do
! Count frequencies from final sample
do k = 1, 3
frequencies(sample(k)) = frequencies(sample(k)) + 1
end do
call free_sample(state)
end do
! Print frequency results
print *, 'Test item frequencies for ', reps, ' runs:'
do i = 0, 9
print '(I2, ": ", I5)', i, frequencies(i)
end do
end program test_knuth_s

View file

@ -0,0 +1,98 @@
Type SampleState
sample(0 To 99) As Integer ' Array to store the sample (arbitrary maximum size)
sampleSize As Integer ' Current sample size
i As Integer ' Counter of processed elements
n As Integer ' Desired sample size
End Type
' Function to generate a random number between 0 and max-1
Function randInt(Byval max As Integer) As Integer
Return Int(Rnd() * max)
End Function
' Function to create a new sampler
Function createSampler(Byval n As Integer) As SampleState
Dim sampler As SampleState
sampler.sampleSize = 0
sampler.i = 0
sampler.n = n
Return sampler
End Function
' Function to process a new element and update the sample
Function sampleOfN(Byref sampler As SampleState, Byval item As Integer) As Integer Ptr
sampler.i += 1
If sampler.i <= sampler.n Then
sampler.sample(sampler.sampleSize) = item
sampler.sampleSize += 1
Else
If randInt(sampler.i) < sampler.n Then
sampler.sample(randInt(sampler.n)) = item
End If
End If
Return @sampler.sample(0)
End Function
Sub printArray(arr() As Integer, Byval size As Integer)
Print "["
For i As Integer = 0 To size-1
Print arr(i)
If i < size-1 Then Print ", "
Next
Print "]"
End Sub
Sub main()
Dim As Integer i, j
Dim As Integer frequency(0 To 9)
For i = 0 To 9
frequency(i) = 0
Next
Print "Single run samples for n = 3:"
Dim As SampleState sampler = createSampler(3)
Dim As Integer Ptr samplePtr
Dim As Integer sample(0 To 2)
For i = 0 To 9
samplePtr = sampleOfN(sampler, i)
For j = 0 To sampler.sampleSize-1
sample(j) = samplePtr[j]
Next
Print " Item: " & i & " -> sample: ";
Print "[ ";
For j = 0 To sampler.sampleSize-1
Print Chr(8) & sample(j);
If j < sampler.sampleSize-1 Then Print ", ";
Next
Print "]"
Next
For i = 0 To 99999
sampler = createSampler(3)
For j = 0 To 9
samplePtr = sampleOfN(sampler, j)
Next
For j = 0 To sampler.n-1
frequency(samplePtr[j]) += 1
Next
Next
Print !"\nTest item frequencies for 100000 runs:"
For i = 0 To 9
Print " " & i & ": " & frequency(i)
Next
End Sub
Randomize Timer
main()
Sleep

View file

@ -0,0 +1,36 @@
class SOfN {
constructor(n) {
this.m = this.n = n
this.s = []
}
add(item) {
if (this.s.length < this.n) {
this.s.push(item)
} else {
const rand = Math.floor(Math.random() * ++this.m)
if (rand < this.n) {
this.s[rand] = item
}
}
}
}
function main() {
for (const [n, m] of [[3, 3], [3, 10]]) {
const freqs = new Array(m).fill(0)
for (let i = 0; i < 1e5; ++i) {
const sOfN = new SOfN(n)
for (let x = 0; x < m; ++x) {
sOfN.add(x)
}
for (const d of sOfN.s) {
++freqs[d]
}
}
console.log(`Results for n=${n}, m=${m}: [${freqs.join(', ')}]`)
}
}
main()