airbyte-cdk 6.55.0__py3-none-any.whl → 6.55.1.post11.dev15684355943__py3-none-any.whl

airbyte_cdk/cli/airbyte_cdk/_connector.py

@@ -101,7 +101,7 @@ def connector_cli_group() -> None:
     pass
 
 
-@connector_cli_group.command()
+@connector_cli_group.command("test")
 @click.argument(
     "connector",
     required=False,
@@ -114,10 +114,18 @@ def connector_cli_group() -> None:
     default=False,
     help="Only collect tests, do not run them.",
 )
-def test(
+@click.option(
+    "--pytest-arg",
+    "pytest_args",  # ← map --pytest-arg into pytest_args
+    type=str,
+    multiple=True,
+    help="Additional argument(s) to pass to pytest. Can be specified multiple times.",
+)
+def connector_test(
     connector: str | Path | None = None,
     *,
     collect_only: bool = False,
+    pytest_args: list[str] | None = None,
 ) -> None:
     """Run connector tests.
 
@@ -130,19 +138,36 @@ def test(
     directory. If the current working directory is not a connector directory (e.g. starting
     with 'source-') and no connector name or path is provided, the process will fail.
     """
+    click.echo("Connector test command executed.")
+    connector_name, connector_directory = resolve_connector_name_and_directory(connector)
+
+    pytest_args = pytest_args or []
+    if collect_only:
+        pytest_args.append("--collect-only")
+
+    run_connector_tests(
+        connector_name=connector_name,
+        connector_directory=connector_directory,
+        extra_pytest_args=pytest_args,
+    )
+
+
+def run_connector_tests(
+    connector_name: str,
+    connector_directory: Path,
+    extra_pytest_args: list[str],
+) -> None:
     if pytest is None:
         raise ImportError(
             "pytest is not installed. Please install pytest to run the connector tests."
         )
-    click.echo("Connector test command executed.")
-    connector_name, connector_directory = resolve_connector_name_and_directory(connector)
 
     connector_test_suite = create_connector_test_suite(
         connector_name=connector_name if not connector_directory else None,
         connector_directory=connector_directory,
     )
 
-    pytest_args: list[str] = []
+    pytest_args: list[str] = ["-p", "airbyte_cdk.test.standard_tests.pytest_hooks"]
     if connector_directory:
         pytest_args.append(f"--rootdir={connector_directory}")
         os.chdir(str(connector_directory))
@@ -158,8 +183,8 @@ def test(
     test_file_path.parent.mkdir(parents=True, exist_ok=True)
     test_file_path.write_text(file_text)
 
-    if collect_only:
-        pytest_args.append("--collect-only")
+    if extra_pytest_args:
+        pytest_args.extend(extra_pytest_args)
 
     pytest_args.append(str(test_file_path))
 
@@ -170,7 +195,6 @@ def test(
 
     click.echo(f"Running tests from connector directory: {connector_directory}...")
     click.echo(f"Test file: {test_file_path}")
-    click.echo(f"Collect only: {collect_only}")
     click.echo(f"Pytest args: {pytest_args}")
     click.echo("Invoking Pytest...")
     exit_code = pytest.main(

airbyte_cdk/cli/airbyte_cdk/_image.py

@@ -10,6 +10,7 @@ from pathlib import Path
 
 import rich_click as click
 
+from airbyte_cdk.cli.airbyte_cdk._connector import run_connector_tests
 from airbyte_cdk.models.connector_metadata import MetadataFile
 from airbyte_cdk.utils.connector_paths import resolve_connector_name_and_directory
 from airbyte_cdk.utils.docker import (
@@ -88,6 +89,81 @@ def build(
         sys.exit(1)
 
 
+@image_cli_group.command("test")
+@click.argument(
+    "connector",
+    required=False,
+    type=str,
+    metavar="[CONNECTOR]",
+)
+@click.option(
+    "--image",
+    help="Image to test, instead of building a new one.",
+)
+def image_test(  # "image test" command
+    connector: str | None = None,
+    *,
+    image: str | None = None,
+) -> None:
+    """Test a connector Docker image.
+
+    [CONNECTOR] can be a connector name (e.g. 'source-pokeapi'), a path to a connector directory, or omitted to use the current working directory.
+    If a string containing '/' is provided, it is treated as a path. Otherwise, it is treated as a connector name.
+
+    If an image is provided, it will be used for testing instead of building a new one.
+
+    Note: You should run `airbyte-cdk secrets fetch` before running this command to ensure
+    that the secrets are available for the connector tests.
+    """
+    if not verify_docker_installation():
+        click.echo(
+            "Docker is not installed or not running. Please install Docker and try again.", err=True
+        )
+        sys.exit(1)
+
+    connector_name, connector_directory = resolve_connector_name_and_directory(connector)
+
+    # Select only tests with the 'image_tests' mark
+    pytest_args = ["-m", "image_tests"]
+    if not image:
+        metadata_file_path: Path = connector_directory / "metadata.yaml"
+        try:
+            metadata = MetadataFile.from_file(metadata_file_path)
+        except (FileNotFoundError, ValueError) as e:
+            click.echo(
+                f"Error loading metadata file '{metadata_file_path}': {e!s}",
+                err=True,
+            )
+            sys.exit(1)
+
+        tag = "dev-latest"
+        image = f"{metadata.data.dockerRepository}:{tag}"
+        click.echo(f"Building Image for Connector: {image}")
+        try:
+            image = build_connector_image(
+                connector_directory=connector_directory,
+                connector_name=connector_name,
+                metadata=metadata,
+                tag=tag,
+                no_verify=True,
+            )
+        except ConnectorImageBuildError as e:
+            click.echo(
+                f"Error building connector image: {e!s}",
+                err=True,
+            )
+            sys.exit(1)
+
+    pytest_args.extend(["--connector-image", image])
+
+    click.echo(f"Testing Connector Image: {image}")
+    run_connector_tests(
+        connector_name=connector_name,
+        connector_directory=connector_directory,
+        extra_pytest_args=pytest_args,
+    )
+
+
 __all__ = [
     "image_cli_group",
 ]

airbyte_cdk/cli/airbyte_cdk/_secrets.py

@@ -99,12 +99,12 @@ def secrets_cli_group() -> None:
     help="Print GitHub CI mask for secrets.",
     type=bool,
     is_flag=True,
-    default=False,
+    default=None,
 )
 def fetch(
     connector: str | Path | None = None,
     gcp_project_id: str = GCP_PROJECT_ID,
-    print_ci_secrets_masks: bool = False,
+    print_ci_secrets_masks: bool | None = None,
 ) -> None:
     """Fetch secrets for a connector from Google Secret Manager.
 
@@ -181,22 +181,23 @@ def fetch(
         if secret_count == 0:
             raise exceptions[0]
 
-    if not print_ci_secrets_masks:
-        return
-
-    if not os.environ.get("CI", None):
+    if print_ci_secrets_masks and "CI" not in os.environ:
         click.echo(
             "The `--print-ci-secrets-masks` option is only available in CI environments. "
             "The `CI` env var is either not set or not set to a truthy value. "
             "Skipping printing secret masks.",
             err=True,
         )
-        return
-
-    # Else print the CI mask
-    _print_ci_secrets_masks(
-        secrets_dir=secrets_dir,
-    )
+        print_ci_secrets_masks = False
+    elif print_ci_secrets_masks is None:
+        # If not explicitly set, we check if we are in a CI environment
+        # and set to True if so.
+        print_ci_secrets_masks = os.environ.get("CI", "") != ""
+
+    if print_ci_secrets_masks:
+        _print_ci_secrets_masks(
+            secrets_dir=secrets_dir,
+        )
 
 
 @secrets_cli_group.command("list")

airbyte_cdk/models/airbyte_protocol_serializers.py

@@ -4,9 +4,11 @@ from typing import Any, Dict
 from serpyco_rs import CustomType, Serializer
 
 from .airbyte_protocol import (  # type: ignore[attr-defined] # all classes are imported to airbyte_protocol via *
+    AirbyteCatalog,
     AirbyteMessage,
     AirbyteStateBlob,
     AirbyteStateMessage,
+    AirbyteStream,
     AirbyteStreamState,
     ConfiguredAirbyteCatalog,
     ConfiguredAirbyteStream,
@@ -30,6 +32,8 @@ def custom_type_resolver(t: type) -> CustomType[AirbyteStateBlob, Dict[str, Any]
     return AirbyteStateBlobType() if t is AirbyteStateBlob else None
 
 
+AirbyteCatalogSerializer = Serializer(AirbyteCatalog, omit_none=True)
+AirbyteStreamSerializer = Serializer(AirbyteStream, omit_none=True)
 AirbyteStreamStateSerializer = Serializer(
     AirbyteStreamState, omit_none=True, custom_type_resolver=custom_type_resolver
 )

airbyte_cdk/models/connector_metadata.py

@@ -34,6 +34,15 @@ class ConnectorBuildOptions(BaseModel):
     )
 
 
+class SuggestedStreams(BaseModel):
+    """Suggested streams from metadata.yaml."""
+
+    streams: list[str] = Field(
+        default=[],
+        description="List of suggested streams for the connector",
+    )
+
+
 class ConnectorMetadata(BaseModel):
     """Connector metadata from metadata.yaml."""
 
@@ -47,6 +56,11 @@ class ConnectorMetadata(BaseModel):
         description="List of tags for the connector",
     )
 
+    suggestedStreams: SuggestedStreams | None = Field(
+        default=None,
+        description="Suggested streams for the connector",
+    )
+
     @property
     def language(self) -> ConnectorLanguage:
         """Get the connector language."""

airbyte_cdk/sources/declarative/declarative_component_schema.yaml

@@ -4186,6 +4186,46 @@ definitions:
       - type
       - stream_config
       - components_mapping
+  StreamParametersDefinition:
+    title: Stream Parameters Definition
+    description: (This component is experimental. Use at your own risk.) Represents a stream parameters definition to set up dynamic streams from defined values in manifest.
+    type: object
+    required:
+      - type
+      - list_of_parameters_for_stream
+    properties:
+      type:
+        type: string
+        enum: [StreamParametersDefinition]
+      list_of_parameters_for_stream:
+        title: Stream Parameters
+        description: A list of object of parameters for stream, each object in the list represents params for one stream.
+        type: array
+        items:
+          type: object
+        examples:
+          - [{"name": "test stream", "$parameters": {"entity": "test entity"}, "primary_key": "test key"}]
+  ParametrizedComponentsResolver:
+    type: object
+    title: Parametrized Components Resolver
+    description: (This component is experimental. Use at your own risk.) Resolves and populates dynamic streams from defined parametrized values in manifest.
+    properties:
+      type:
+        type: string
+        enum: [ParametrizedComponentsResolver]
+      stream_parameters:
+        "$ref": "#/definitions/StreamParametersDefinition"
+      components_mapping:
+        type: array
+        items:
+          "$ref": "#/definitions/ComponentMappingDefinition"
+      $parameters:
+        type: object
+        additionalProperties: true
+    required:
+      - type
+      - stream_parameters
+      - components_mapping
   DynamicDeclarativeStream:
     type: object
     description: (This component is experimental. Use at your own risk.) A component that described how will be created declarative streams based on stream template.
@@ -4212,6 +4252,7 @@ definitions:
         anyOf:
           - "$ref": "#/definitions/HttpComponentsResolver"
           - "$ref": "#/definitions/ConfigComponentsResolver"
+          - "$ref": "#/definitions/ParametrizedComponentsResolver"
     required:
       - type
       - stream_template

airbyte_cdk/sources/declarative/models/declarative_component_schema.py

@@ -1512,6 +1512,31 @@ class ConfigComponentsResolver(BaseModel):
     parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
 
 
+class StreamParametersDefinition(BaseModel):
+    type: Literal["StreamParametersDefinition"]
+    list_of_parameters_for_stream: List[Dict[str, Any]] = Field(
+        ...,
+        description="A list of object of parameters for stream, each object in the list represents params for one stream.",
+        examples=[
+            [
+                {
+                    "name": "test stream",
+                    "$parameters": {"entity": "test entity"},
+                    "primary_key": "test key",
+                }
+            ]
+        ],
+        title="Stream Parameters",
+    )
+
+
+class ParametrizedComponentsResolver(BaseModel):
+    type: Literal["ParametrizedComponentsResolver"]
+    stream_parameters: StreamParametersDefinition
+    components_mapping: List[ComponentMappingDefinition]
+    parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
+
+
 class RequestBodyPlainText(BaseModel):
     type: Literal["RequestBodyPlainText"]
     value: str
@@ -2943,7 +2968,9 @@ class DynamicDeclarativeStream(BaseModel):
     stream_template: Union[DeclarativeStream, StateDelegatingStream] = Field(
         ..., description="Reference to the stream template.", title="Stream Template"
     )
-    components_resolver: Union[HttpComponentsResolver, ConfigComponentsResolver] = Field(
+    components_resolver: Union[
+        HttpComponentsResolver, ConfigComponentsResolver, ParametrizedComponentsResolver
+    ] = Field(
         ...,
         description="Component resolve and populates stream templates with components values.",
         title="Components Resolver",

airbyte_cdk/sources/declarative/parsers/model_to_component_factory.py

@@ -351,6 +351,9 @@ from airbyte_cdk.sources.declarative.models.declarative_component_schema import
 from airbyte_cdk.sources.declarative.models.declarative_component_schema import (
     PageIncrement as PageIncrementModel,
 )
+from airbyte_cdk.sources.declarative.models.declarative_component_schema import (
+    ParametrizedComponentsResolver as ParametrizedComponentsResolverModel,
+)
 from airbyte_cdk.sources.declarative.models.declarative_component_schema import (
     ParentStreamConfig as ParentStreamConfigModel,
 )
@@ -504,7 +507,9 @@ from airbyte_cdk.sources.declarative.resolvers import (
     ComponentMappingDefinition,
     ConfigComponentsResolver,
     HttpComponentsResolver,
+    ParametrizedComponentsResolver,
     StreamConfig,
+    StreamParametersDefinition,
 )
 from airbyte_cdk.sources.declarative.retrievers import (
     AsyncRetriever,
@@ -738,6 +743,7 @@ class ModelToComponentFactory:
             AsyncRetrieverModel: self.create_async_retriever,
             HttpComponentsResolverModel: self.create_http_components_resolver,
             ConfigComponentsResolverModel: self.create_config_components_resolver,
+            ParametrizedComponentsResolverModel: self.create_parametrized_components_resolver,
             StreamConfigModel: self.create_stream_config,
             ComponentMappingDefinitionModel: self.create_components_mapping_definition,
             ZipfileDecoderModel: self.create_zipfile_decoder,
@@ -3861,6 +3867,29 @@ class ModelToComponentFactory:
             parameters=model.parameters or {},
         )
 
+    def create_parametrized_components_resolver(
+        self, model: ParametrizedComponentsResolverModel, config: Config
+    ) -> ParametrizedComponentsResolver:
+        stream_parameters = StreamParametersDefinition(
+            list_of_parameters_for_stream=model.stream_parameters.list_of_parameters_for_stream
+        )
+        components_mapping = [
+            self._create_component_from_model(
+                model=components_mapping_definition_model,
+                value_type=ModelToComponentFactory._json_schema_type_name_to_type(
+                    components_mapping_definition_model.value_type
+                ),
+                config=config,
+            )
+            for components_mapping_definition_model in model.components_mapping
+        ]
+        return ParametrizedComponentsResolver(
+            stream_parameters=stream_parameters,
+            config=config,
+            components_mapping=components_mapping,
+            parameters=model.parameters or {},
+        )
+
     _UNSUPPORTED_DECODER_ERROR = (
         "Specified decoder of {decoder_type} is not supported for pagination."
         "Please set as `JsonDecoder`, `XmlDecoder`, or a `CompositeRawDecoder` with an inner_parser of `JsonParser` or `GzipParser` instead."

airbyte_cdk/sources/declarative/resolvers/__init__.py

@@ -12,6 +12,9 @@ from airbyte_cdk.sources.declarative.models import (
 from airbyte_cdk.sources.declarative.models import (
     HttpComponentsResolver as HttpComponentsResolverModel,
 )
+from airbyte_cdk.sources.declarative.models import (
+    ParametrizedComponentsResolver as ParametrizedComponentsResolverModel,
+)
 from airbyte_cdk.sources.declarative.resolvers.components_resolver import (
     ComponentMappingDefinition,
     ComponentsResolver,
@@ -24,10 +27,15 @@ from airbyte_cdk.sources.declarative.resolvers.config_components_resolver import
 from airbyte_cdk.sources.declarative.resolvers.http_components_resolver import (
     HttpComponentsResolver,
 )
+from airbyte_cdk.sources.declarative.resolvers.parametrized_components_resolver import (
+    ParametrizedComponentsResolver,
+    StreamParametersDefinition,
+)
 
 COMPONENTS_RESOLVER_TYPE_MAPPING: Mapping[str, type[BaseModel]] = {
     "HttpComponentsResolver": HttpComponentsResolverModel,
     "ConfigComponentsResolver": ConfigComponentsResolverModel,
+    "ParametrizedComponentsResolver": ParametrizedComponentsResolverModel,
 }
 
 __all__ = [
@@ -38,4 +46,6 @@ __all__ = [
     "StreamConfig",
     "ConfigComponentsResolver",
     "COMPONENTS_RESOLVER_TYPE_MAPPING",
+    "ParametrizedComponentsResolver",
+    "StreamParametersDefinition",
 ]

airbyte_cdk/sources/declarative/resolvers/parametrized_components_resolver.py

@@ -0,0 +1,125 @@
+#
+# Copyright (c) 2024 Airbyte, Inc., all rights reserved.
+#
+
+from copy import deepcopy
+from dataclasses import InitVar, dataclass, field
+from typing import Any, Dict, Iterable, List, Mapping
+
+import dpath
+import yaml
+from typing_extensions import deprecated
+from yaml.parser import ParserError
+
+from airbyte_cdk.sources.declarative.interpolation import InterpolatedString
+from airbyte_cdk.sources.declarative.resolvers.components_resolver import (
+    ComponentMappingDefinition,
+    ComponentsResolver,
+    ResolvedComponentMappingDefinition,
+)
+from airbyte_cdk.sources.source import ExperimentalClassWarning
+from airbyte_cdk.sources.types import Config
+
+
+@deprecated("This class is experimental. Use at your own risk.", category=ExperimentalClassWarning)
+@dataclass
+class StreamParametersDefinition:
+    """
+    Represents a stream parameters definition to set up dynamic streams from defined values in manifest.
+    """
+
+    list_of_parameters_for_stream: List[Dict[str, Any]]
+
+
+@deprecated("This class is experimental. Use at your own risk.", category=ExperimentalClassWarning)
+@dataclass
+class ParametrizedComponentsResolver(ComponentsResolver):
+    """
+    Resolves and populates dynamic streams from defined parametrized values in manifest.
+    """
+
+    stream_parameters: StreamParametersDefinition
+    config: Config
+    components_mapping: List[ComponentMappingDefinition]
+    parameters: InitVar[Mapping[str, Any]]
+    _resolved_components: List[ResolvedComponentMappingDefinition] = field(
+        init=False, repr=False, default_factory=list
+    )
+
+    def __post_init__(self, parameters: Mapping[str, Any]) -> None:
+        """
+        Initializes and parses component mappings, converting them to resolved definitions.
+
+        Args:
+            parameters (Mapping[str, Any]): Parameters for interpolation.
+        """
+
+        for component_mapping in self.components_mapping:
+            if isinstance(component_mapping.value, (str, InterpolatedString)):
+                interpolated_value = (
+                    InterpolatedString.create(component_mapping.value, parameters=parameters)
+                    if isinstance(component_mapping.value, str)
+                    else component_mapping.value
+                )
+
+                field_path = [
+                    InterpolatedString.create(path, parameters=parameters)
+                    for path in component_mapping.field_path
+                ]
+
+                self._resolved_components.append(
+                    ResolvedComponentMappingDefinition(
+                        field_path=field_path,
+                        value=interpolated_value,
+                        value_type=component_mapping.value_type,
+                        create_or_update=component_mapping.create_or_update,
+                        parameters=parameters,
+                    )
+                )
+            else:
+                raise ValueError(
+                    f"Expected a string or InterpolatedString for value in mapping: {component_mapping}"
+                )
+
+    def resolve_components(
+        self, stream_template_config: Dict[str, Any]
+    ) -> Iterable[Dict[str, Any]]:
+        kwargs = {"stream_template_config": stream_template_config}
+
+        for components_values in self.stream_parameters.list_of_parameters_for_stream:
+            updated_config = deepcopy(stream_template_config)
+            kwargs["components_values"] = components_values  # type: ignore[assignment] # component_values will always be of type Mapping[str, Any]
+            for resolved_component in self._resolved_components:
+                valid_types = (
+                    (resolved_component.value_type,) if resolved_component.value_type else None
+                )
+                value = resolved_component.value.eval(
+                    self.config, valid_types=valid_types, **kwargs
+                )
+                path = [path.eval(self.config, **kwargs) for path in resolved_component.field_path]
+                parsed_value = self._parse_yaml_if_possible(value)
+                # https://github.com/dpath-maintainers/dpath-python/blob/master/dpath/__init__.py#L136
+                # dpath.set returns the number of changed elements, 0 when no elements changed
+                updated = dpath.set(updated_config, path, parsed_value)
+
+                if parsed_value and not updated and resolved_component.create_or_update:
+                    dpath.new(updated_config, path, parsed_value)
+
+            yield updated_config
+
+    @staticmethod
+    def _parse_yaml_if_possible(value: Any) -> Any:
+        """
+        Try to turn value into a Python object by YAML-parsing it.
+
+        * If value is a `str` and can be parsed by `yaml.safe_load`,
+          return the parsed result.
+        * If parsing fails (`yaml.parser.ParserError`) – or value is not
+          a string at all – return the original value unchanged.
+        """
+        if isinstance(value, str):
+            try:
+                return yaml.safe_load(value)
+            except ParserError:  # "{{ record[0] in ['cohortActiveUsers'] }}"   # not valid YAML
+                return value
+        return value