Allow base64-service-account-json key auth Issue: #923

.changes/unreleased/Features-20240516-125735.yaml

@@ -0,0 +1,6 @@
+kind: Features
+body: Add support for base 64 encoded json keyfile credentials
+time: 2024-05-16T12:57:35.383416-07:00
+custom:
+    Author: robeleb1
+    Issue: "923"

dbt/adapters/bigquery/connections.py

@@ -40,9 +40,11 @@
 from dbt.adapters.events.types import SQLQuery
 from dbt_common.events.functions import fire_event
 from dbt.adapters.bigquery import __version__ as dbt_version
+from dbt.adapters.bigquery.utility import is_base64, base64ToString
 
 from dbt_common.dataclass_schema import ExtensibleDbtClassMixin, StrEnum
 
+
 logger = AdapterLogger("BigQuery")
 
 BQ_QUERY_JOB_SPLIT = "-----Query Job SQL Follows-----"
@@ -125,7 +127,7 @@ class BigQueryCredentials(Credentials):
     job_creation_timeout_seconds: Optional[int] = None
     job_execution_timeout_seconds: Optional[int] = None
 
-    # Keyfile json creds
+    # Keyfile json creds (unicode or base 64 encoded)
     keyfile: Optional[str] = None
     keyfile_json: Optional[Dict[str, Any]] = None
 
@@ -332,6 +334,8 @@ def get_google_credentials(cls, profile_credentials) -> GoogleCredentials:
 
         elif method == BigQueryConnectionMethod.SERVICE_ACCOUNT_JSON:
             details = profile_credentials.keyfile_json
+            if is_base64(profile_credentials.keyfile_json):
+                details = base64ToString(details)
             return creds.from_service_account_info(details, scopes=profile_credentials.scopes)
 
         elif method == BigQueryConnectionMethod.OAUTH_SECRETS:

dbt/adapters/bigquery/utility.py

@@ -1,3 +1,5 @@
+import base64
+import binascii
 import json
 from typing import Any, Optional
 
@@ -43,3 +45,39 @@ def sql_escape(string):
     if not isinstance(string, str):
         raise dbt_common.exceptions.CompilationError(f"cannot escape a non-string: {string}")
     return json.dumps(string)[1:-1]
+
+
+def is_base64(s: str | bytes) -> bool:
+    """
+    Checks if the given string or bytes object is valid Base64 encoded.
+
+    Args:
+        s: The string or bytes object to check.
+
+    Returns:
+        True if the input is valid Base64, False otherwise.
+    """
+
+    if isinstance(s, str):
+        # For strings, ensure they consist only of valid Base64 characters
+        if not s.isascii():
+            return False
+        # Convert to bytes for decoding
+        s = s.encode("ascii")
+
+    try:
+        # Use the 'validate' parameter to enforce strict Base64 decoding rules
+        base64.b64decode(s, validate=True)
+        return True
+    except TypeError:
+        return False
+    except binascii.Error:  # Catch specific errors from the base64 module
+        return False
+
+
+def base64ToString(b):
+    return base64.b64decode(b).decode("utf-8")
+
+
+def stringToBase64(s):
+    return base64.b64encode(s.encode("utf-8"))

tests/conftest.py

@@ -1,6 +1,7 @@
 import pytest
 import os
 import json
+from dbt.adapters.bigquery.utility import is_base64, base64ToString
 
 # Import the fuctional fixtures as a plugin
 # Note: fixtures with session scope need to be local
@@ -38,6 +39,8 @@ def oauth_target():
 
 def service_account_target():
     credentials_json_str = os.getenv("BIGQUERY_TEST_SERVICE_ACCOUNT_JSON").replace("'", "")
+    if is_base64(credentials_json_str):
+        credentials_json_str = base64ToString(credentials_json_str)
     credentials = json.loads(credentials_json_str)
     project_id = credentials.get("project_id")
     return {

tests/functional/adapter/test_json_keyfile.py

@@ -0,0 +1,82 @@
+import json
+import pytest
+from dbt.adapters.bigquery.utility import stringToBase64, is_base64
+
+
+@pytest.fixture
+def example_json_keyfile():
+    keyfile = json.dumps(
+        {
+            "type": "service_account",
+            "project_id": "",
+            "private_key_id": "",
+            "private_key": "-----BEGIN PRIVATE KEY----------END PRIVATE KEY-----\n",
+            "client_email": "",
+            "client_id": "",
+            "auth_uri": "https://accounts.google.com/o/oauth2/auth",
+            "token_uri": "https://oauth2.googleapis.com/token",
+            "auth_provider_x509_cert_url": "https://www.googleapis.com/oauth2/v1/certs",
+            "client_x509_cert_url": "",
+        }
+    )
+
+    return keyfile
+
+
+@pytest.fixture
+def example_json_keyfile_b64():
+    keyfile = json.dumps(
+        {
+            "type": "service_account",
+            "project_id": "",
+            "private_key_id": "",
+            "private_key": "-----BEGIN PRIVATE KEY----------END PRIVATE KEY-----\n",
+            "client_email": "",
+            "client_id": "",
+            "auth_uri": "https://accounts.google.com/o/oauth2/auth",
+            "token_uri": "https://oauth2.googleapis.com/token",
+            "auth_provider_x509_cert_url": "https://www.googleapis.com/oauth2/v1/certs",
+            "client_x509_cert_url": "",
+        }
+    )
+
+    return stringToBase64(keyfile)
+
+
+def test_valid_base64_strings(example_json_keyfile_b64):
+    valid_strings = [
+        "SGVsbG8gV29ybGQh",  # "Hello World!"
+        "Zm9vYmFy",  # "foobar"
+        "QUJDREVGR0hJSktMTU5PUFFSU1RVVldYWVowMTIzNDU2Nzg5",  # A long string
+        "",  # Empty string
+        example_json_keyfile_b64.decode("utf-8"),
+    ]
+
+    for s in valid_strings:
+        assert is_base64(s) is True
+
+
+def test_valid_base64_bytes(example_json_keyfile_b64):
+    valid_bytes = [
+        b"SGVsbG8gV29ybGQh",  # "Hello World!"
+        b"Zm9vYmFy",  # "foobar"
+        b"QUJDREVGR0hJSktMTU5PUFFSU1RVVldYWVowMTIzNDU2Nzg5",  # A long string
+        b"",  # Empty bytes
+        example_json_keyfile_b64,
+    ]
+    for s in valid_bytes:
+        assert is_base64(s) is True
+
+
+def test_invalid_base64(example_json_keyfile):
+    invalid_inputs = [
+        "This is not Base64",
+        "SGVsbG8gV29ybGQ",  # Incorrect padding
+        "Invalid#Base64",
+        12345,  # Not a string or bytes
+        b"Invalid#Base64",
+        "H\xffGVsbG8gV29ybGQh",  # Contains invalid character \xff
+        example_json_keyfile,
+    ]
+    for s in invalid_inputs:
+        assert is_base64(s) is False