So I was looking at the code here: https://github.com/biopython/biopython/blob/master/Bio/Cluster/cluster.c

And saw that the function "double median(int n, double x[])" can be improved. I wonder if this file is still in use or is it outdated?

This C implementation is ~~3.5 times faster on average when using g++ -O2 (2-5 times, depending on the input). And it can still be optimized more. The version for N-even is pretty much the same. https://drive.google.com/file/d/1uElcWd-Yf2pKdXP-azvFVUnS9AxmTcsX/view?usp=sharing

void medianOddCore(int n, double a[], int middle, int l, int r)
{  
    if (l >= r) return;

    int i = l, j = r;
    int x = (l + r) >> 1;
    while (i<=j)        
    {
        while (a[i] < a[x]) i++;
        while (a[j] > a[x]) j--;

        if (i<=j) {
            std::swap(a[i], a[j]);
            i++;
            j--;
        }  
    }

    if (l <= middle && middle <= j)
        medianOddCore(n, a, middle, l, j);
    else
        return medianOddCore(n, a, middle, i, r);
}

double medianOdd(int n, double a[])
{
    medianOddCore(n, a, n / 2, 0, n-1);
    return a[n / 2];
}

The google drive link contains all the sources code, which includes unit-testing for the function (using Biopython's median function as a baseline). It's in C++ to use the timer, file output for debug, ..; but the algorithm itself is purely C and does not need any library.

This is my first time looking at an open-source project, so I wonder if it can be added to Biopython? Thank you.