Quick and dirty script to split the huge XML file with all the wikipedia pages into individual files. No actual XML parsing is done, it handles the input as a text file and parses it line by line, so it doesn't need much memory to run.

wikisplit.py

#!/usr/bin/env python3

import os, sys, shutil

infile = sys.argv[1]
outdir = sys.argv[2]

shutil.rmtree(outdir, ignore_errors=True)
os.mkdir(outdir)

with open(infile) as i:
	p = False
	n = 0
	l = i.readline()
	while l:
		if l.strip() == '<page>':
			p = True
			n += 1
			o = open(f'{outdir}{os.sep}{n:010d}.xml', 'w')
			print(f'\r{n:010d}', end='', flush=True)
		elif l.strip() == '</page>':
			p = False
			o.write(l)
			o.close()
		if p:
			o.write(l)
		l = i.readline()