def load_bigann()

in lib/data.py [0:0]

• 17 lines of code
• 3 McCabe index (conditional complexity)

def load_bigann(device, size = 10 ** 6, test=True, qsize=10 ** 5):
    basedir = getBasedir("bigann")

    dbsize = int(size / 10 ** 6)
    xt = mmap_bvecs(join(basedir, 'bigann_learn.bvecs'))
    if test:
        xb = mmap_bvecs(join(basedir, 'bigann_base.bvecs'))
        xq = mmap_bvecs(join(basedir, 'bigann_query.bvecs'))
        # trim xb to correct size
        xb = xb[:dbsize * 1000 * 1000]
        gt = ivecs_read(join(basedir, 'gnd/idx_%dM.ivecs' % dbsize))
    else:
        xb = xt[:size]
        xq = xt[size:size+qsize]
        xt = xt[size+qsize:]

    xb, xq = sanitize(xb), sanitize(xq)
    if not test:
        gt = get_nearestneighbors(xq, xb, 100, device)

    return xt, xb, xq, gt