Files
truf-server/app/keycheckers/zai/zaiKeycheck.py
T
2026-09-30 20:30:56 +03:00

509 lines
19 KiB
Python

import sys
sys.dont_write_bytecode = True
import argparse
import json
import os
import re
from urllib.parse import urlparse
import requests
sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
from keycheck_common import (
combined_provider_routing_hint,
commit_status_transaction,
default_input_file,
default_proxy_file,
ensure_output_files,
iter_findings,
keycheck_input_mode,
load_checked_statuses,
load_known_keys,
load_proxies,
mask_secret,
provider_routing_database_failed,
read_plain_keys,
record_validation_result,
recover_status_transaction,
require_provider_authority,
service_output_dir,
should_skip_key,
write_keycheck_event,
)
from keycheckers.provider_resolution import (
AMBIGUOUS_GENERIC_SK_HINT,
AMBIGUOUS_QWEN_DEEPSEEK_HINT,
resolve_provider_key,
)
SERVICE = "zai"
DETECTOR = "ZaiGLM"
OUTPUT_DIR = os.getenv("KEYCHECK_OUTPUT_DIR") or service_output_dir(SERVICE)
INPUT_FILE = os.getenv("KEYCHECK_INPUT_FILE") or default_input_file()
PROXY_FILE = os.getenv("KEYCHECK_PROXY_FILE") or default_proxy_file()
CHECKED_FILE = os.path.join(OUTPUT_DIR, "zaiChecked.txt")
RESULTS_FILE = os.path.join(OUTPUT_DIR, "zaiResults.jsonl")
STATUS_FILES = {
"VALID": os.path.join(OUTPUT_DIR, "zaiAlive.txt"),
"NO_BALANCE": os.path.join(OUTPUT_DIR, "zaiNoBalance.txt"),
"DEAD": os.path.join(OUTPUT_DIR, "zaiDead.txt"),
"RESTRICTED": os.path.join(OUTPUT_DIR, "zaiRestricted.txt"),
"LIMITED": os.path.join(OUTPUT_DIR, "zaiLimited.txt"),
"NETWORK": os.path.join(OUTPUT_DIR, "zaiNetwork.txt"),
"UNKNOWN": os.path.join(OUTPUT_DIR, "zaiUnknown.txt"),
}
DEFAULT_BASE_URLS = (
"https://api.z.ai/api/paas/v4",
"https://open.bigmodel.cn/api/paas/v4",
)
ZAI_KEY_REGEX = re.compile(
r"(?<![A-Za-z0-9_.-])(?:"
r"(?:zai|sk)-[A-Za-z0-9][A-Za-z0-9_-]{20,505}|"
r"[A-Fa-f0-9]{32}\.[A-Za-z0-9_-]{16,128}"
r")(?![A-Za-z0-9_.-])"
)
ZAI_DOTTED_KEY_REGEX = re.compile(r"^[A-Fa-f0-9]{32}\.[A-Za-z0-9_-]{16,128}$")
ZAI_CONTEXT_REGEX = re.compile(
r"(?:ZAI_API_KEY|GLM_API_KEY|ZHIPUAI_API_KEY|BIGMODEL_API_KEY|api\.z\.ai|"
r"open\.bigmodel\.cn|zhipuai|chatglm)",
re.IGNORECASE,
)
FOREIGN_KEY_PREFIXES = ("sk-ant-", "sk-or-", "sk-proj-", "sk-svcacct-", "sk-admin-")
AMBIGUOUS_HINTS = {AMBIGUOUS_QWEN_DEEPSEEK_HINT, AMBIGUOUS_GENERIC_SK_HINT}
AUTH_FAILURE_CODES = {"1000", "1001", "1003"}
AUTHENTICATED_NO_BALANCE_CODES = {"1113"}
AUTHENTICATED_LIMIT_CODES = {"1302", "1308", "1309", "1310", "1311"}
AUTHENTICATED_RESTRICTED_CODES = {"1005", "1220"}
PROBE_MODEL = "glm-5.2"
def normalize_base_url(value):
return str(value or "").strip().rstrip("/")
def split_csv(value):
if not value:
return []
values = value.split(",") if isinstance(value, str) else value
return [str(item).strip() for item in values if str(item).strip()]
def unique_ordered(values):
output = []
seen = set()
for value in values:
normalized = normalize_base_url(value)
if normalized and normalized not in seen:
seen.add(normalized)
output.append(normalized)
return output
def base_urls_from_environment(extra=None, include_defaults=True):
configured = []
for value in extra or ():
configured.extend(split_csv(value))
configured.extend(split_csv(os.getenv("ZAI_BASE_URLS") or os.getenv("ZHIPU_BASE_URLS")))
defaults = DEFAULT_BASE_URLS if include_defaults else ()
return unique_ordered([*configured, *defaults])
def endpoint_label(base_url):
parsed = urlparse(base_url)
return parsed.netloc or base_url
def key_from_text(*values):
for value in values:
match = ZAI_KEY_REGEX.search(str(value or ""))
if match:
return match.group(0)
return ""
def key_rejection_reason(key):
value = str(key or "")
try:
encoded = value.encode("utf-8", errors="strict")
except UnicodeEncodeError:
return "candidate is not valid UTF-8"
if len(encoded) > 512:
return "candidate exceeds the 512-byte key limit"
if value.startswith(FOREIGN_KEY_PREFIXES):
return "candidate has a foreign provider prefix"
if not ZAI_KEY_REGEX.fullmatch(value):
return "candidate does not match a bounded ZAI key format"
return ""
def finding_detector_names(finding):
if not isinstance(finding, dict):
return set()
extra = finding.get("ExtraData") if isinstance(finding.get("ExtraData"), dict) else {}
return {
name for name in (
str(finding.get("DetectorName") or finding.get("DetectorType") or "").strip().lower(),
str(extra.get("name") or "").strip().lower(),
) if name
}
def finding_provider_routing_hint(finding, key=""):
if not isinstance(finding, dict):
return "zai" if ZAI_DOTTED_KEY_REGEX.fullmatch(str(key or "")) else ""
context = finding.get("ScannerContext") if isinstance(finding.get("ScannerContext"), dict) else {}
persisted = str(context.get("provider_hint") or "").strip().lower()
if persisted:
return persisted
if "zaiglm" in finding_detector_names(finding) or ZAI_DOTTED_KEY_REGEX.fullmatch(str(key or "")):
return "zai"
text = "\n".join(str(context.get(name) or "") for name in ("nearby", "file"))
return "zai" if ZAI_CONTEXT_REGEX.search(text) else ""
def iter_candidate_decisions(input_file, plain_files, trusted_retry_files=None):
detectors = [
"ZaiGLM", "zaiglm", "CustomRegex", "QwenDashScope", "Qwen_DashScope",
"Qwen", "DashScope", "DeepSeek", "DeepSeekApiKey", "DeepSeek_API_Key",
"KimiMoonshot", "MoonshotAI", "Moonshot", "Kimi",
]
seen = set()
for item in iter_findings(input_file, detectors):
finding = item.get("finding") or {}
key = item.get("credential_secret_text") or key_from_text(
item.get("raw"), item.get("raw_v2"), finding.get("Raw"), finding.get("RawV2"),
)
if not key or key_rejection_reason(key):
continue
local_hint = finding_provider_routing_hint(finding, key)
hint = combined_provider_routing_hint(key, local_hint)
if provider_routing_database_failed():
raise RuntimeError("provider routing evidence lookup failed closed")
seen.add(key)
yield key, item.get("source") or input_file, finding, hint
owned_retry_paths = {
os.path.normcase(os.path.abspath(path)) for path in STATUS_FILES.values()
}
retry_paths = [
path for path in trusted_retry_files or ()
if os.path.normcase(os.path.abspath(path)) in owned_retry_paths
]
for item in read_plain_keys([*plain_files, *retry_paths], ZAI_KEY_REGEX):
key = item["key"]
if key in seen or key_rejection_reason(key):
continue
yield key, item["source"], {}, "zai"
def redact_text(value, key):
text = str(value or "")[:1000]
if key:
text = text.replace(key, "***REDACTED***")
return ZAI_KEY_REGEX.sub("***REDACTED***", text)
def parse_error(response, key):
try:
payload = response.json()
except ValueError:
payload = {}
error = payload.get("error") if isinstance(payload, dict) else {}
if not isinstance(error, dict):
error = {}
return {
"http_status": int(response.status_code),
"code": str(error.get("code") or (payload.get("code") if isinstance(payload, dict) else "") or ""),
"message": redact_text(
error.get("message") or error.get("msg") or (
payload.get("message") or payload.get("msg") if isinstance(payload, dict) else ""
) or response.text[:500],
key,
),
}
def classify_error(error):
http_status = int(error.get("http_status") or 0)
code = str(error.get("code") or "")
message = str(error.get("message") or "").lower()
if http_status == 402 or any(marker in message for marker in (
"insufficient balance", "balance is insufficient", "no balance", "account balance",
"recharge", "payment required", "billing arrears", "credit balance",
)):
return "NO_BALANCE", True
if code in AUTHENTICATED_NO_BALANCE_CODES:
return "NO_BALANCE", True
if any(marker in message for marker in (
"quota", "rate limit", "rate-limit", "too many requests", "resource exhausted",
"concurrency limit", "usage limit",
)):
return "LIMITED", True
if code in AUTHENTICATED_LIMIT_CODES:
return "LIMITED", True
if any(marker in message for marker in (
"permission denied", "access denied", "not authorized for", "model access", "forbidden",
)):
return "RESTRICTED", True
if code in AUTHENTICATED_RESTRICTED_CODES:
return "RESTRICTED", True
if http_status == 401 or code in AUTH_FAILURE_CODES:
return "DEAD", False
if 500 <= http_status <= 599 or code in {"1200", "1230", "1234", "1305"}:
return "NETWORK", False
if http_status == 429:
return "LIMITED", False
if http_status == 403:
return "RESTRICTED", False
return "UNKNOWN", False
def probe_chat_completion(key, base_url, model, proxy, timeout, debug=False):
if not model:
return {
"status": "UNKNOWN", "model": "",
"message": "no chat-capable model returned by /models",
}
url = f"{normalize_base_url(base_url)}/chat/completions"
headers = {"Authorization": f"Bearer {key}", "Content-Type": "application/json"}
payload = {
"model": model,
"messages": [{"role": "user", "content": "ping"}],
"max_tokens": 1,
"stream": False,
}
try:
response = requests.post(url, headers=headers, json=payload, proxies=proxy, timeout=timeout)
except requests.RequestException as exc:
return {
"status": "NETWORK", "model": model,
"message": redact_text(exc, key),
}
if debug:
print(
f" DEBUG {endpoint_label(base_url)} chat probe {model}: HTTP {response.status_code}: "
f"{redact_text(response.text[:500], key)}"
)
if response.status_code == 200:
try:
response_payload = response.json()
except ValueError as exc:
return {
"status": "UNKNOWN", "model": model, "http_status": response.status_code,
"message": f"invalid chat completion response: {exc}",
}
if isinstance(response_payload, dict) and response_payload.get("choices"):
return {
"status": "GENERATION_OK", "model": model,
"http_status": response.status_code, "message": "chat completion accepted",
}
error = parse_error(response, key)
status, authenticated = classify_error(error)
return {
"status": status, "authenticated": authenticated, "model": model,
"http_status": response.status_code, "business_code": error.get("code") or "",
"error": error, "message": error.get("message") or "",
}
def check_base_url(key, base_url, proxy, timeout, debug=False):
url = f"{normalize_base_url(base_url)}/models"
headers = {"Authorization": f"Bearer {key}", "Accept": "application/json"}
try:
response = requests.get(url, headers=headers, proxies=proxy, timeout=timeout)
except requests.RequestException as exc:
return {
"status": "NETWORK", "base_url": base_url,
"region": endpoint_label(base_url), "message": redact_text(exc, key),
}
if debug:
print(
f" DEBUG {endpoint_label(base_url)} /models: HTTP {response.status_code}: "
f"{redact_text(response.text[:500], key)}"
)
if response.status_code == 200:
try:
payload = response.json()
data = payload.get("data") if isinstance(payload, dict) else None
if not isinstance(data, list):
raise ValueError("missing data model list")
models = sorted({
str(item.get("id") or item.get("name") or "")
for item in data if isinstance(item, dict) and (item.get("id") or item.get("name"))
})
except (TypeError, ValueError, json.JSONDecodeError) as exc:
return {
"status": "UNKNOWN", "base_url": base_url,
"region": endpoint_label(base_url), "message": f"invalid models response: {exc}",
}
probe_model = PROBE_MODEL
probe = probe_chat_completion(key, base_url, probe_model, proxy, timeout, debug)
probe_status = probe.get("status") or "UNKNOWN"
if probe_status == "GENERATION_OK":
status = "VALID"
elif probe_status in {"NO_BALANCE", "LIMITED", "RESTRICTED", "NETWORK"}:
status = probe_status
elif probe_status == "DEAD":
status = "RESTRICTED"
else:
status = "UNKNOWN"
probe_message = probe.get("message") or json.dumps(probe.get("error") or {}, ensure_ascii=False)
message = (
f"chat probe ok; model={probe_model}; models={len(data)}"
if status == "VALID"
else f"models authenticated; generation_probe={probe_status}; model={probe_model}; "
f"models={len(data)}; {probe_message}"
)
return {
"status": status, "authenticated": True, "base_url": base_url,
"region": endpoint_label(base_url), "model_count": len(data),
"models": models[:30], "llm_probe_status": probe_status,
"llm_probe_model": probe.get("model") or probe_model,
"llm_probe_http_status": probe.get("http_status"),
"business_code": probe.get("business_code") or "",
"probe": probe, "error": probe.get("error") or {},
"message": message[:1000],
}
error = parse_error(response, key)
status, authenticated = classify_error(error)
return {
"status": status, "authenticated": authenticated, "base_url": base_url,
"region": endpoint_label(base_url), "http_status": response.status_code,
"business_code": error.get("code") or "", "error": error,
"message": error.get("message") or "",
}
def check_key(key, base_urls, proxy, timeout, debug=False):
rejection = key_rejection_reason(key)
if rejection:
return {"status": "UNKNOWN", "message": rejection, "candidate_rejected": True}
attempts = []
for base_url in base_urls:
result = check_base_url(key, base_url, proxy, timeout, debug)
attempts.append(result)
if result.get("authenticated"):
return {**result, "attempts": attempts}
statuses = [attempt.get("status") for attempt in attempts]
status = next(
(candidate for candidate in ("NETWORK", "LIMITED", "RESTRICTED", "UNKNOWN", "DEAD") if candidate in statuses),
"UNKNOWN",
)
selected = next((attempt for attempt in attempts if attempt.get("status") == status), {})
return {**selected, "status": status, "attempts": attempts}
def ensure_files():
ensure_output_files([CHECKED_FILE, RESULTS_FILE, *STATUS_FILES.values()])
recover_status_transaction(CHECKED_FILE, STATUS_FILES)
def write_result(key, result, source, finding):
status = result.get("status") or "UNKNOWN"
write_keycheck_event(SERVICE, RESULTS_FILE, key, result, source, finding, DETECTOR)
commit_status_transaction(
CHECKED_FILE, STATUS_FILES, key, status,
result.get("message", ""), result.get("region") or source,
)
record_validation_result(SERVICE, key, result, source, finding, DETECTOR)
def retry_statuses_from_args(args):
statuses = set()
for enabled, status in (
(args.retry_network, "NETWORK"),
(args.retry_limited, "LIMITED"),
(args.retry_unknown, "UNKNOWN"),
(args.retry_restricted, "RESTRICTED"),
(args.retry_no_balance, "NO_BALANCE"),
(args.retry_valid, "VALID"),
):
if enabled:
statuses.add(status)
return statuses
def retry_input_files_from_args(args):
if args.recheck_all:
statuses = list(STATUS_FILES)
else:
statuses = list(retry_statuses_from_args(args))
return [STATUS_FILES[status] for status in statuses]
def parse_args():
parser = argparse.ArgumentParser(description="ZAI / Zhipu GLM key checker")
parser.add_argument("--input", default=INPUT_FILE)
parser.add_argument("--plain", action="append", default=[])
parser.add_argument("--proxy-file", default=PROXY_FILE)
parser.add_argument("--timeout", type=int, default=15)
parser.add_argument("--max-keys", type=int, default=0)
parser.add_argument("--base-url", action="append", default=[])
parser.add_argument("--no-default-base-urls", action="store_true")
parser.add_argument("--retry-network", action="store_true")
parser.add_argument("--retry-limited", action="store_true")
parser.add_argument("--retry-unknown", action="store_true")
parser.add_argument("--retry-restricted", action="store_true")
parser.add_argument("--retry-no-balance", action="store_true")
parser.add_argument("--retry-valid", action="store_true")
parser.add_argument("--recheck-all", action="store_true")
parser.add_argument("--debug", action="store_true")
return parser.parse_args()
def main():
require_provider_authority(SERVICE)
args = parse_args()
ensure_files()
proxy_cycler = load_proxies(args.proxy_file)
checked = load_checked_statuses(CHECKED_FILE)
known = load_known_keys(CHECKED_FILE, STATUS_FILES)
retry_statuses = retry_statuses_from_args(args)
retry_files = retry_input_files_from_args(args)
base_urls = base_urls_from_environment(args.base_url, not args.no_default_base_urls)
if not base_urls:
raise SystemExit("No ZAI base URLs configured")
processed = 0
skipped = 0
postgres_mode = keycheck_input_mode() == "postgres"
for key, source, finding, routing_hint in iter_candidate_decisions(args.input, args.plain, retry_files):
is_ambiguous = routing_hint in AMBIGUOUS_HINTS
if routing_hint != "zai" and not (postgres_mode and is_ambiguous):
skipped += 1
continue
if should_skip_key(
key, checked, known, args, retry_statuses, service=SERVICE,
source=source, finding=finding, detector=DETECTOR,
):
skipped += 1
continue
if args.max_keys and processed >= args.max_keys:
break
processed += 1
print(f"\n[{processed}] ZAI candidate {mask_secret(key)} from {source}")
proxy = next(proxy_cycler) if proxy_cycler else None
if is_ambiguous:
result = resolve_provider_key(
key, finding, proxy, args.timeout, args.debug,
hint=routing_hint, origin_service=SERVICE,
)
else:
result = check_key(key, base_urls, proxy, args.timeout, args.debug)
print(f" STATUS: {result['status']} | {result.get('message', '')[:200]}")
write_result(key, result, source, finding)
known.add(key)
checked[key] = result["status"]
print(f"\nDone. Processed={processed}, skipped={skipped}, results={RESULTS_FILE}")
if __name__ == "__main__":
main()