#!/usr/bin/env python3
"""Verify VerdictTank index-v5.1.html against v5.1 rules."""
import re, sys
from html.parser import HTMLParser

F = "/root/projects/verdicttank/index-v5.1.html"
html = open(F, encoding="utf-8").read()
fail = []

print("=== FORBIDDEN TOKENS (all must be 0) ===")
forbidden = ["\u2014", "GPT-5.6", "unlimited", "Unlimited", "UNLIMITED",
             "11 seats", "11-seat", "11 models", "11-judge", "9 vendors",
             "9-vendor", "eleven", "nine vendors", "nine-vendor",
             "nine rostered", "Three tiers", "$299"]
for pat in forbidden:
    n = html.count(pat)
    print(f"  {pat!r:22} {n}")
    if n:
        fail.append(f"forbidden token present: {pat!r} x{n}")

# $79 as a standalone price (not $799)
n79 = len(re.findall(r"\$79(?!\d)", html))
print(f"  {'$79 (standalone)':22} {n79}")
if n79:
    fail.append(f"stale $79 price x{n79}")

print("\n=== REQUIRED TOKENS (all must be >0) ===")
required = ["6-seat", "6 seats", "5 vendors", "4-seat reduced", "$5.20",
            "$249", "$799", "$1,499", "$207", "$666", "$1,249",
            "$15/review", "$10/review", "Most popular", "Pre-Review Coach",
            "v5.1", "verdicttank.com", "Moonshot AI", "Legal + Compliance",
            "Financial + Market", "Synthesis + Gate", "Cross-Check"]
for pat in required:
    n = html.count(pat)
    print(f"  {pat!r:22} {n}")
    if not n:
        fail.append(f"required token missing: {pat!r}")

print("\n=== COGS ARITHMETIC ===")
c = 5.20
for name, q, rev, claimed_cogs, claimed_m in [
        ("Pro", 5, 249, 26, 90), ("Enterprise", 30, 799, 156, 80),
        ("White-Label", 50, 1499, 260, 83)]:
    cogs = q * c
    m = (rev - cogs) / rev * 100
    ok = abs(cogs - claimed_cogs) < 0.01 and abs(round(m) - claimed_m) <= 1
    print(f"  {name:12} {q:2} x ${c} = ${cogs:7.2f} on ${rev:5} -> {m:.1f}% "
          f"(page claims ${claimed_cogs}/{claimed_m}%) {'OK' if ok else 'MISMATCH'}")
    if not ok:
        fail.append(f"COGS mismatch for {name}")

print(f"  overage $15 / ${c} = {15/c:.2f}x COGS (page: roughly 3x)")
print(f"  overage $10 / ${c} = {10/c:.2f}x COGS (page: roughly 2x)")
for base, ann in [(249, 207), (799, 666), (1499, 1249)]:
    d = (1 - ann / base) * 100
    ok = abs(d - 16.7) < 1.0
    print(f"  annual {base} -> {ann} = {d:.1f}% off {'OK' if ok else 'CHECK'}")
    if not ok:
        fail.append(f"annual discount off for {base}->{ann}: {d:.1f}%")
print(f"  AutogenAI $30K / Ent $9,588/yr  = {30000/(799*12):.2f}x (page: 3.1x)")
print(f"  AutogenAI $30K / WL  $17,988/yr = {30000/(1499*12):.2f}x (page: 1.7x)")

print("\n=== ANTHROPIC CONCENTRATION ===")
roster = re.search(r"<h3>The roster</h3>.*?</table>", html, re.S).group(0)
rows = re.findall(r"<tr><td>.*?</tr>", roster, re.S)
vendors = [re.findall(r"<td>([^<]*)</td>", r) for r in rows]
seats = [(v[1], v[2], v[3], v[4]) for v in vendors if len(v) >= 5]
print(f"  total seats: {len(seats)}")
for s in seats:
    print(f"    {s[0]:20} {s[1]:18} {s[2]:12} scores={s[3]}")
vs = {}
for s in seats:
    vs[s[2]] = vs.get(s[2], 0) + 1
print(f"  distinct vendors: {len(vs)} -> {vs}")
anth = vs.get("Anthropic", 0)
print(f"  Anthropic: {anth}/{len(seats)} = {anth/len(seats)*100:.0f}%")
if len(seats) != 6:
    fail.append(f"roster has {len(seats)} seats, expected 6")
if len(vs) != 5:
    fail.append(f"roster has {len(vs)} vendors, expected 5")
if anth != 2:
    fail.append(f"Anthropic holds {anth} seats, expected 2")

print("\n=== HTML STRUCTURE ===")
VOID = {"br", "img", "meta", "link", "hr", "input", "area", "base", "col",
        "embed", "source", "track", "wbr"}
class P(HTMLParser):
    def __init__(self):
        super().__init__(convert_charrefs=True)
        self.st, self.err = [], []
    def handle_starttag(self, t, a):
        if t not in VOID:
            self.st.append((t, self.getpos()))
    def handle_endtag(self, t):
        if t in VOID:
            return
        if not self.st:
            self.err.append(f"stray </{t}> at {self.getpos()}")
            return
        if self.st[-1][0] != t:
            self.err.append(f"mismatch </{t}> at {self.getpos()}, "
                            f"open <{self.st[-1][0]}> from {self.st[-1][1]}")
        else:
            self.st.pop()
p = P()
p.feed(html)
print(f"  unclosed tags: {p.st or 'none'}")
print(f"  structure errors: {p.err or 'none'}")
if p.st or p.err:
    fail.append("HTML structure problems")

print("\n=== PRICING LAYOUT ===")
cards = len(re.findall(r'class="price-card', html))
print(f"  price-card divs: {cards} (expected 4)")
if cards != 4:
    fail.append(f"{cards} price cards, expected 4")
grid4 = "grid-template-columns:repeat(4,1fr);gap:16px;margin:20px 0" in html
print(f"  .pricing-grid is 4-col: {grid4}")
if not grid4:
    fail.append("pricing-grid not 4 columns")
ribbon = html.count('<div class="ribbon">Most popular</div>')
print(f"  'Most popular' ribbon: {ribbon} (expected 1)")
if ribbon != 1:
    fail.append("ribbon count wrong")

# comparison table row labels
print("\n=== COMPARISON TABLE ROWS ===")
for label in ["Price", "Reviews per month", "Overage", "Panel", "Fix-Its",
              "Re-score loop", "Pre-Review Coach", "White-label", "Workspaces",
              "Corpus isolation", "Judge pool config", "Reseller model",
              "Annual billing (16.7% off)"]:
    present = f"<strong>{label}</strong>" in html
    print(f"  {label:30} {'OK' if present else 'MISSING'}")
    if not present:
        fail.append(f"comparison row missing: {label}")

print(f"\n  file size: {len(html)} bytes")
print("\n" + "=" * 50)
if fail:
    print(f"FAILED ({len(fail)}):")
    for f_ in fail:
        print(f"  - {f_}")
    sys.exit(1)
print("ALL CHECKS PASSED")