universal-autoreg: initial public release (restored, redacted, hardened)

providers/cleanapis_oauth.py

@@ -0,0 +1,105 @@
+# cleanapis.com mass reg: GitHub OAuth -> free Starter plan -> api key -> verify -> save
+import sys, re, json, time, random, urllib.parse, requests, os
+sys.path.insert(0, 'C:/Users/User/tmp/regfarm/github_oauth')
+from gh_http import GHSession
+
+CID = "Ov23li48tsmVvDWqWfhs"
+SCOPE = "user:email,read:user"
+GH_POOL = "C:/Users/User/tmp/gh_accounts.json"
+OUT = "cleanapis_accounts.jsonl"
+USED = "ca_used_gh.txt"
+UA = {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/130.0.0.0 Safari/537.36'}
+
+def load_pool():
+    d = json.load(open(GH_POOL, encoding='utf-8'))
+    used = set()
+    try: used = set(open(USED).read().split())
+    except Exception: pass
+    return [a for a in d if a.get('login') not in used]
+
+def mark_used(login):
+    open(USED, 'a').write(login + '\n')
+
+def reg_one(acc):
+    res = {'gh_login': acc['login'], 'ok': False, 'provider': 'github_oauth', 'ts': int(time.time())}
+    s = requests.Session(); s.headers.update(UA)
+    r = s.get('https://cleanapis.com/auth/github/redirect', allow_redirects=False, timeout=30)
+    q = urllib.parse.parse_qs(urllib.parse.urlparse(r.headers.get('Location', '')).query)
+    state = q.get('state', [None])[0]
+    if not state:
+        res['err'] = 'no state'; return res
+    ghs = GHSession(bind=False)
+    ok, msg = ghs.login(acc['login'], acc['password'], acc.get('totp'))
+    if not ok:
+        res['err'] = 'gh login: ' + str(msg)[:120]; return res
+    code, err = ghs.oauth_code(CID, state, scope=SCOPE)
+    if not code:
+        res['err'] = 'oauth: ' + str(err)[:150]; return res
+    cb = s.get('https://cleanapis.com/auth/github/callback?code=%s&state=%s' % (code, state), allow_redirects=True, timeout=40)
+    res['cb'] = cb.status_code
+    if '/login' in cb.url and 'dashboard' not in cb.url:
+        res['err'] = 'callback bounced'; return res
+    # activate free starter plan
+    g = s.get('https://cleanapis.com/dashboard/get-started?plan=starter', timeout=25)
+    csrf = re.findall(r'name="_token" value="([^"]+)"', g.text)
+    if csrf:
+        act = s.post('https://cleanapis.com/dashboard/billing/checkout/4',
+                     data={'_token': csrf[0], 'billing_cycle': 'monthly', 'coupon_code': '', 'payment_method': 'paddle'},
+                     allow_redirects=True, timeout=40,
+                     headers={'Origin': 'https://cleanapis.com', 'Referer': 'https://cleanapis.com/dashboard/get-started?plan=starter'})
+        res['plan_ok'] = 'billing/success' in act.url or 'plan' in act.url
+    # create api key
+    t = s.get('https://cleanapis.com/dashboard/api-keys', timeout=25)
+    csrf2 = re.findall(r'name="_token" value="([^"]+)"', t.text)
+    if not csrf2:
+        res['err'] = 'no csrf api-keys'; return res
+    nm = 'farm%d' % random.randint(10000, 99999)
+    cr = s.post('https://cleanapis.com/dashboard/api-keys',
+                data={'_token': csrf2[0], 'name': nm, 'scopes[]': ['inference', 'models:read', 'embeddings']},
+                allow_redirects=True, timeout=40,
+                headers={'Origin': 'https://cleanapis.com', 'Referer': 'https://cleanapis.com/dashboard/api-keys'})
+    ids = re.findall(r'api-keys.{0,4}?(\d{4,7}).{0,4}?reveal', cr.text)
+    ids = list(dict.fromkeys(ids))
+    if not ids:
+        res['err'] = 'no key created'; return res
+    # reveal newest key (highest id)
+    kid = max(ids, key=int)
+    rv = s.get('https://cleanapis.com/dashboard/api-keys/%s/reveal' % kid, headers={'Accept': 'application/json'}, timeout=25)
+    try: d = rv.json()
+    except Exception: res['err'] = 'reveal non-json ' + rv.text[:80]; return res
+    if not d.get('ok'):
+        res['err'] = 'reveal fail: ' + str(d)[:100]; return res
+    k = d['key']
+    mr = requests.get('https://cleanapis.com/v1/models', headers={'Authorization': 'Bearer ' + k}, timeout=30)
+    res['models'] = mr.status_code
+    res['api_key'] = k
+    res['key_id'] = kid
+    res['ok'] = (mr.status_code == 200)
+    mark_used(acc['login'])
+    return res
+
+def main():
+    n = int(sys.argv[1]) if len(sys.argv) > 1 else 1
+    start = int(sys.argv[2]) if len(sys.argv) > 2 else 0
+    pool = load_pool()
+    print('pool free:', len(pool), 'start:', start, flush=True)
+    okc = fail429 = 0
+    for i, acc in enumerate(pool[start:start+n], start=start):
+        try:
+            r = reg_one(acc)
+        except Exception as e:
+            r = {'idx': i, 'gh_login': acc.get('login'), 'ok': False, 'err': 'EXC %s: %s' % (type(e).__name__, str(e)[:150])}
+        if r.get('ok'): okc += 1
+        if '429' in str(r.get('err', '')):
+            fail429 += 1
+            if fail429 >= 3:
+                print('3x429 -> sleep 1800s', flush=True)
+                time.sleep(1800); fail429 = 0
+        line = json.dumps(r, ensure_ascii=False)
+        print('[%d] %s' % (i, line), flush=True)
+        open(OUT, 'a', encoding='utf-8').write(line + '\n')
+        time.sleep(random.uniform(20, 35))
+    print('DONE ok=%d/%d' % (okc, n), flush=True)
+
+if __name__ == '__main__':
+    main()