airbyte-source-azure-blob-storage 0.4.4__tar.gz → 0.5.0__tar.gz

{airbyte_source_azure_blob_storage-0.4.4 → airbyte_source_azure_blob_storage-0.5.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: airbyte-source-azure-blob-storage
-Version: 0.4.4
+Version: 0.5.0
 Summary: Source implementation for Azure Blob Storage.
 Home-page: https://airbyte.com
 License: MIT

{airbyte_source_azure_blob_storage-0.4.4 → airbyte_source_azure_blob_storage-0.5.0}/pyproject.toml

@@ -5,7 +5,7 @@ requires = [
 build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
-version = "0.4.4"
+version = "0.5.0"
 name = "airbyte-source-azure-blob-storage"
 description = "Source implementation for Azure Blob Storage."
 authors = [

{airbyte_source_azure_blob_storage-0.4.4 → airbyte_source_azure_blob_storage-0.5.0}/source_azure_blob_storage/config_migrations.py

@@ -9,19 +9,18 @@ from typing import Any, List, Mapping
 
 from airbyte_cdk import AirbyteEntrypoint, Source, create_connector_config_control_message
 
+
 logger = logging.getLogger("airbyte_logger")
 
 
 class MigrateConfig(ABC):
     @classmethod
     @abstractmethod
-    def should_migrate(cls, config: Mapping[str, Any]) -> bool:
-        ...
+    def should_migrate(cls, config: Mapping[str, Any]) -> bool: ...
 
     @classmethod
     @abstractmethod
-    def migrate_config(cls, config: Mapping[str, Any]) -> Mapping[str, Any]:
-        ...
+    def migrate_config(cls, config: Mapping[str, Any]) -> Mapping[str, Any]: ...
 
     @classmethod
     def modify_and_save(cls, config_path: str, source: Source, config: Mapping[str, Any]) -> Mapping[str, Any]:

{airbyte_source_azure_blob_storage-0.4.4 → airbyte_source_azure_blob_storage-0.5.0}/source_azure_blob_storage/spec.py

@@ -6,9 +6,10 @@
 from typing import Any, Dict, Literal, Optional, Union
 
 import dpath.util
+from pydantic import AnyUrl, BaseModel, Field
+
 from airbyte_cdk import OneOfOptionConfig
 from airbyte_cdk.sources.file_based.config.abstract_file_based_spec import AbstractFileBasedSpec
-from pydantic import AnyUrl, BaseModel, Field
 
 
 class Oauth2(BaseModel):
@@ -35,6 +36,25 @@ class Oauth2(BaseModel):
     )
 
 
+class ClientCredentials(BaseModel):
+    class Config(OneOfOptionConfig):
+        title = "Authenticate via Client Credentials"
+        discriminator = "auth_type"
+
+    auth_type: Literal["client_credentials"] = Field("client_credentials", const=True)
+    app_tenant_id: str = Field(title="Tenant ID", description="Tenant ID of the Microsoft Azure Application", airbyte_secret=True)
+    app_client_id: str = Field(
+        title="Client ID",
+        description="Client ID of your Microsoft developer application",
+        airbyte_secret=True,
+    )
+    app_client_secret: str = Field(
+        title="Client Secret",
+        description="Client Secret of your Microsoft developer application",
+        airbyte_secret=True,
+    )
+
+
 class StorageAccountKey(BaseModel):
     class Config(OneOfOptionConfig):
         title = "Authenticate via Storage Account Key"
@@ -60,7 +80,7 @@ class SourceAzureBlobStorageSpec(AbstractFileBasedSpec):
     def documentation_url(cls) -> AnyUrl:
         return AnyUrl("https://docs.airbyte.com/integrations/sources/azure-blob-storage", scheme="https")
 
-    credentials: Union[Oauth2, StorageAccountKey] = Field(
+    credentials: Union[Oauth2, ClientCredentials, StorageAccountKey] = Field(
         title="Authentication",
         description="Credentials for connecting to the Azure Blob Storage",
         discriminator="auth_type",

{airbyte_source_azure_blob_storage-0.4.4 → airbyte_source_azure_blob_storage-0.5.0}/source_azure_blob_storage/stream_reader.py

@@ -3,22 +3,62 @@
 
 import logging
 from io import IOBase
-from typing import Iterable, List, Optional, Union
+from typing import Any, Iterable, List, Mapping, MutableMapping, Optional, Union
 
 import pytz
+from azure.core.credentials import AccessToken, TokenCredential
+from azure.core.exceptions import ResourceNotFoundError
+from azure.storage.blob import BlobServiceClient, ContainerClient
+from smart_open import open
+
 from airbyte_cdk import AirbyteTracedException, FailureType
 from airbyte_cdk.sources.file_based.file_based_stream_reader import AbstractFileBasedStreamReader, FileReadMode
 from airbyte_cdk.sources.file_based.remote_file import RemoteFile
 from airbyte_cdk.sources.streams.http.requests_native_auth import Oauth2Authenticator
-from azure.core.credentials import AccessToken
-from azure.core.exceptions import ResourceNotFoundError
-from azure.storage.blob import BlobServiceClient, ContainerClient
-from smart_open import open
 
 from .spec import SourceAzureBlobStorageSpec
 
 
-class AzureOauth2Authenticator(Oauth2Authenticator):
+class AzureClientCredentialsAuthenticator(Oauth2Authenticator, TokenCredential):
+    def __init__(self, tenant_id: str, client_id: str, client_secret: str, **kwargs):
+        super().__init__(
+            token_refresh_endpoint=f"https://login.microsoftonline.com/{tenant_id}/oauth2/v2.0/token",
+            client_id=client_id,
+            client_secret=client_secret,
+            grant_type="client_credentials",
+            scopes=["https://storage.azure.com/.default"],
+            refresh_token=None,
+        )
+
+    def build_refresh_request_body(self) -> Mapping[str, Any]:
+        """
+        Returns the request body to set on the refresh request
+
+        Override to define additional parameters
+        """
+        payload: MutableMapping[str, Any] = {
+            "grant_type": self.get_grant_type(),
+            "client_id": self.get_client_id(),
+            "client_secret": self.get_client_secret(),
+        }
+
+        if self.get_scopes():
+            payload["scope"] = " ".join(self.get_scopes())
+
+        if self.get_refresh_request_body():
+            for key, val in self.get_refresh_request_body().items():
+                # We defer to existing oauth constructs over custom configured fields
+                if key not in payload:
+                    payload[key] = val
+
+        return payload
+
+    def get_token(self, *args, **kwargs) -> AccessToken:
+        """Parent class handles Oauth Refresh token logic."""
+        return AccessToken(token=self.get_access_token(), expires_on=int(self.get_token_expiry_date().timestamp()))
+
+
+class AzureOauth2Authenticator(Oauth2Authenticator, TokenCredential):
     """
     Authenticator for Azure Blob Storage SDK to align with azure.core.credentials.TokenCredential protocol
     """
@@ -62,17 +102,24 @@ class SourceAzureBlobStorageStreamReader(AbstractFileBasedStreamReader):
         return BlobServiceClient(self.account_url, credential=self._credentials)
 
     @property
-    def azure_credentials(self) -> Union[str, AzureOauth2Authenticator]:
+    def azure_credentials(self) -> Union[str, AzureOauth2Authenticator, AzureClientCredentialsAuthenticator]:
         if not self._credentials:
             if self.config.credentials.auth_type == "storage_account_key":
                 self._credentials = self.config.credentials.azure_blob_storage_account_key
-            else:
+            elif self.config.credentials.auth_type == "oauth2":
                 self._credentials = AzureOauth2Authenticator(
                     token_refresh_endpoint=f"https://login.microsoftonline.com/{self.config.credentials.tenant_id}/oauth2/v2.0/token",
                     client_id=self.config.credentials.client_id,
                     client_secret=self.config.credentials.client_secret,
                     refresh_token=self.config.credentials.refresh_token,
                 )
+            elif self.config.credentials.auth_type == "client_credentials":
+                self._credentials = AzureClientCredentialsAuthenticator(
+                    tenant_id=self.config.credentials.app_tenant_id,
+                    client_id=self.config.credentials.app_client_id,
+                    client_secret=self.config.credentials.app_client_secret,
+                )
+
         return self._credentials
 
     def get_matching_files(