infiniflow/ragflow · error · Error

message.compileNotSupported

Error message

message.compileNotSupported

What it means

After creds exist, validation requires a primary admin email for delegated Drive API calls; _primary_admin_email is only set inside load_credentials from the credential JSON. None here means creds were loaded through a path that skipped that assignment (or load_credentials never ran to completion).

Source

Thrown at web/src/hooks/use-dataset-generate.ts:205

  const queryClient = useQueryClient();
  const { id } = useParams();
  const { handleUnbindTask } = useUnBindTask();
  const { t } = useTranslation();

  const {
    data,
    isPending: loading,
    mutateAsync,
  } = useMutation({
    mutationKey: [DatasetKey.generate],
    mutationFn: async ({ type }: { type: GenerateType }) => {
      // Go/hybrid: dataset compilation is driven automatically by the scheduler
      // on document completion; there is no manual RunIndex trigger, so a manual
      // "generate" must NOT pretend to succeed. The UI hides/disables the
      // control; if it is ever invoked, reject loudly so callers don't mistake
      // it for a real run (plan v4.1 §4.2).
      if (isGoDatasetBackend()) {
        throw new Error(t('message.compileNotSupported'));
      }
      const { data } = await runIndex(id!, TraceTypeMap[type]);
      if (data.code === 0) {
        message.success(t('message.operated'));
        queryClient.invalidateQueries({
          queryKey: DatasetGenerateKeys.traceById(type, id),
        });
      }
      return data;
    },
  });

  const { mutateAsync: pauseGenerate } = useMutation({
    mutationKey: [DatasetKey.pauseGenerate],
    mutationFn: async ({
      task_id,
      type,
    }: {

View on GitHub (pinned to 554fb1133a)

Solutions

1. Re-run the full load_credentials() with a payload containing the primary admin key (see error 627)
2. If subclassing or patching, set both _creds and _primary_admin_email
3. Validate the credential JSON shape before handing it to the connector

Example fix

# before (test hack)
connector._creds = mocked_creds
connector.validate_connector_settings()  # boom: admin email None

# after
connector.load_credentials({**token_payload, "primary_admin_email": "admin@acme.com"})
connector.validate_connector_settings()

def admin_email_loaded(connector) -> bool:
    return connector._primary_admin_email is not None

from common.data_source.exceptions import ConnectorValidationError

try:
    connector.validate_connector_settings()
except ConnectorValidationError as e:
    if "primary_admin_email" in str(e).lower() or "DB_CREDENTIALS_PRIMARY_ADMIN_KEY" in str(e):
        connector.load_credentials({**raw_creds, "primary_admin_email": fetch_admin_email()})
        connector.validate_connector_settings()

When it happens

Trigger: load_credentials() raised before assigning _primary_admin_email (missing key → ValueError) but a partial state persisted, custom code assigning _creds manually, or a credentials payload where the key mapped to None.

Common situations: Custom tests monkey-patching _creds, subclasses overriding load_credentials without setting the admin email, credential dicts built with null values.

Related errors

• main() returned a non-JSON-serializable value.
• main() must return a value. Use null for an empty result.
• Failed to fetch memory detail
• Failed to update memory
• Failed to fetch search detail

AI-assisted analysis of infiniflow/ragflow@554fb1133a (2026-08-15). Data as JSON: /api/errors/35552f897be8de0b. Report an issue: GitHub.