FEAT: Add ActiveDirectoryServicePrincipal support for bulk copy

mssql_python/auth.py

@@ -130,6 +130,103 @@ def _acquire_token(auth_type: str) -> Tuple[bytes, str]:
             raise RuntimeError(f"Failed to create {credential_class.__name__}: {e}") from e
 
 
+def _parse_tenant_id(sts_url: str) -> Optional[str]:
+    """Extract tenant ID (GUID or domain) from a FedAuthInfo STS URL.
+
+    Expected formats:
+      https://login.microsoftonline.com/<tenant>/
+      https://login.microsoftonline.com/<tenant>/?...
+      https://login.microsoftonline.com/<tenant>
+    where <tenant> is either a GUID (e.g. ``aaaaaaaa-bbbb-cccc-dddd-eeeeeeeeeeee``)
+    or a verified domain (e.g. ``contoso.onmicrosoft.com``). Both forms are
+    accepted by ``azure.identity.ClientSecretCredential``.
+    """
+    # pylint: disable=import-outside-toplevel
+    from urllib.parse import urlparse
+
+    try:
+        parsed = urlparse(sts_url)
+    except (ValueError, AttributeError):
+        return None
+    path = (parsed.path or "").strip("/")
+    if not path:
+        return None
+    first_segment = path.split("/", 1)[0]
+    return first_segment or None
+
+
+class ServicePrincipalAuth:
+    """Builds an ``entra_id_token_factory`` callable for ActiveDirectoryServicePrincipal.
+
+    The bulkcopy path through mssql-py-core uses callback-based token
+    acquisition (FedAuth workflow ``0x02``) because tenant_id is only known
+    from the STS URL that the server returns during the TDS handshake.
+    """
+
+    @staticmethod
+    def make_token_factory(client_id: str, client_secret: str):
+        """Return a callable suitable for ``entra_id_token_factory``.
+
+        Signature: ``(spn: str, sts_url: str, auth_method: str) -> bytes``.
+        Returns the JWT encoded as UTF-16LE bytes (the TDS FedAuth wire format).
+        ``ClientSecretCredential`` is constructed inside the callable so
+        each invocation rebuilds it; tenant is parsed from ``sts_url``
+        on every call rather than captured.
+        """
+        if not client_id:
+            raise ValueError("ServicePrincipal auth requires a non-empty client_id (UID)")
+        if not client_secret:
+            raise ValueError("ServicePrincipal auth requires a non-empty client_secret (PWD)")
+
+        def _factory(spn: str, sts_url: str, auth_method: str) -> bytes:
+            # pylint: disable=import-outside-toplevel,unused-argument
+            try:
+                from azure.identity import ClientSecretCredential
+                from azure.core.exceptions import ClientAuthenticationError
+            except ImportError as e:
+                raise RuntimeError(
+                    "Azure authentication libraries are not installed. "
+                    "Please install with: pip install azure-identity azure-core"
+                ) from e
+
+            tenant_id = _parse_tenant_id(sts_url)
+            if not tenant_id:
+                raise RuntimeError(
+                    f"Could not extract tenant_id from STS URL: {sts_url!r}"
+                )
+
+            logger.info(
+                "ServicePrincipal token factory: acquiring token for tenant=%s, spn=%s",
+                tenant_id,
+                spn,
+            )
+            try:
+                credential = ClientSecretCredential(
+                    tenant_id=tenant_id,
+                    client_id=client_id,
+                    client_secret=client_secret,
+                )
+                # mssql-tds passes the resource SPN; azure-identity wants a scope.
+                scope = spn if spn.endswith("/.default") else spn.rstrip("/") + "/.default"
+                token = credential.get_token(scope).token
+                logger.info(
+                    "ServicePrincipal token factory: token acquired, length=%d chars",
+                    len(token),
+                )
+                return token.encode("utf-16-le")
+            except ClientAuthenticationError as e:
+                logger.error(
+                    "ServicePrincipal authentication failed: tenant=%s, error=%s",
+                    tenant_id,
+                    str(e),
+                )
+                raise RuntimeError(
+                    f"ServicePrincipal authentication failed for tenant '{tenant_id}': {e}"
+                ) from e
+
+        return _factory
+
+
 def process_auth_parameters(parameters: List[str]) -> Tuple[List[str], Optional[str]]:
     """
     Process connection parameters and extract authentication type.
@@ -180,6 +277,19 @@ def process_auth_parameters(parameters: List[str]) -> Tuple[List[str], Optional[
                 # Default authentication (uses DefaultAzureCredential)
                 logger.debug("process_auth_parameters: Default Azure authentication detected")
                 auth_type = "default"
+            elif value_lower == AuthType.SERVICE_PRINCIPAL.value:
+                # ServicePrincipal authentication. ODBC (msodbcsql 17.3+)
+                # handles this natively for regular queries, so leave
+                # auth_type=None to let ODBC own the query path.
+                # Bulkcopy still needs the auth type — extract_auth_type()
+                # propagates it as "serviceprincipal" so the bulkcopy path
+                # can register an entra_id_token_factory callback (Model B,
+                # required because tenant_id is only known from the STS URL
+                # that the server returns during the FedAuth handshake).
+                logger.debug(
+                    "process_auth_parameters: Service principal authentication detected"
+                )
+                auth_type = None
         modified_parameters.append(param)
 
     logger.debug(
@@ -246,6 +356,7 @@ def extract_auth_type(connection_string: str) -> Optional[str]:
         AuthType.INTERACTIVE.value: "interactive",
         AuthType.DEVICE_CODE.value: "devicecode",
         AuthType.DEFAULT.value: "default",
+        AuthType.SERVICE_PRINCIPAL.value: "serviceprincipal",
     }
     for part in connection_string.split(";"):
         key, _, value = part.strip().partition("=")

mssql_python/constants.py

@@ -337,6 +337,7 @@ class AuthType(Enum):
     INTERACTIVE = "activedirectoryinteractive"
     DEVICE_CODE = "activedirectorydevicecode"
     DEFAULT = "activedirectorydefault"
+    SERVICE_PRINCIPAL = "activedirectoryserviceprincipal"
 
 
 class SQLTypes:

mssql_python/cursor.py

@@ -2934,24 +2934,58 @@ def bulkcopy(
         # Token acquisition — only thing cursor must handle (needs azure-identity SDK)
         if self.connection._auth_type:
             # Fresh token acquisition for mssql-py-core connection
-            from mssql_python.auth import AADAuth
-
-            try:
-                raw_token = AADAuth.get_raw_token(self.connection._auth_type)
-            except (RuntimeError, ValueError) as e:
-                raise RuntimeError(
-                    f"Bulk copy failed: unable to acquire Azure AD token "
-                    f"for auth_type '{self.connection._auth_type}': {e}"
-                ) from e
-            pycore_context["access_token"] = raw_token
-            # Token replaces credential fields — py-core's validator rejects
-            # access_token combined with authentication/user_name/password.
-            for key in ("authentication", "user_name", "password"):
-                pycore_context.pop(key, None)
-            logger.debug(
-                "Bulk copy: acquired fresh Azure AD token for auth_type=%s",
-                self.connection._auth_type,
-            )
+            from mssql_python.auth import AADAuth, ServicePrincipalAuth
+
+            if self.connection._auth_type == "serviceprincipal":
+                # Model B: callback-based. tenant_id is only known from the
+                # STS URL the server returns mid-handshake, so we register a
+                # factory that py-core invokes during FedAuth (workflow 0x02).
+                client_id = params.get("uid", "")
+                client_secret = params.get("pwd", "")
+                if not client_id or not client_secret:
+                    raise RuntimeError(
+                        "Bulk copy with Authentication=ActiveDirectoryServicePrincipal "
+                        "requires UID (client_id) and PWD (client_secret) in the "
+                        "connection string."
+                    )
+                try:
+                    factory = ServicePrincipalAuth.make_token_factory(
+                        client_id, client_secret
+                    )
+                except (RuntimeError, ValueError) as e:
+                    raise RuntimeError(
+                        f"Bulk copy failed: unable to build ServicePrincipal "
+                        f"token factory: {e}"
+                    ) from e
+                pycore_context["entra_id_token_factory"] = factory
+                # Keep authentication/user_name/password in pycore_context —
+                # py-core's auth validator + transformer need them to resolve
+                # the auth method to ActiveDirectoryServicePrincipal before
+                # the factory is dispatched at handshake time.
+                logger.debug(
+                    "Bulk copy: registered ServicePrincipal token factory for client_id=%s",
+                    client_id,
+                )
+            else:
+                # Model A: pre-acquired token. Used for Default, DeviceCode,
+                # Interactive (non-Windows), and any other AD method whose
+                # tenant_id is discoverable client-side via Azure Identity SDK.
+                try:
+                    raw_token = AADAuth.get_raw_token(self.connection._auth_type)
+                except (RuntimeError, ValueError) as e:
+                    raise RuntimeError(
+                        f"Bulk copy failed: unable to acquire Azure AD token "
+                        f"for auth_type '{self.connection._auth_type}': {e}"
+                    ) from e
+                pycore_context["access_token"] = raw_token
+                # Token replaces credential fields — py-core's validator rejects
+                # access_token combined with authentication/user_name/password.
+                for key in ("authentication", "user_name", "password"):
+                    pycore_context.pop(key, None)
+                logger.debug(
+                    "Bulk copy: acquired fresh Azure AD token for auth_type=%s",
+                    self.connection._auth_type,
+                )
 
         pycore_connection = None
         pycore_cursor = None