airbyte-cdk 6.21.0.dev0__tar.gz → 6.21.1.dev0__tar.gz

{airbyte_cdk-6.21.0.dev0 → airbyte_cdk-6.21.1.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: airbyte-cdk
-Version: 6.21.0.dev0
+Version: 6.21.1.dev0
 Summary: A framework for writing Airbyte Connectors.
 License: MIT
 Keywords: airbyte,connector-development-kit,cdk

{airbyte_cdk-6.21.0.dev0 → airbyte_cdk-6.21.1.dev0}/airbyte_cdk/sources/declarative/declarative_component_schema.yaml

@@ -1316,6 +1316,7 @@ definitions:
             - "$ref": "#/definitions/KeysToLower"
             - "$ref": "#/definitions/KeysToSnakeCase"
             - "$ref": "#/definitions/FlattenFields"
+            - "$ref": "#/definitions/DpathFlattenFields"
             - "$ref": "#/definitions/KeysReplace"
       state_migrations:
         title: State Migrations
@@ -1513,7 +1514,6 @@ definitions:
         anyOf:
           - "$ref": "#/definitions/JsonDecoder"
           - "$ref": "#/definitions/XmlDecoder"
-          - "$ref": "#/definitions/CompositeRawDecoder"
       $parameters:
         type: object
         additionalProperties: true
@@ -1789,6 +1789,10 @@ definitions:
           - type: array
             items:
               type: string
+      condition:
+        type: string
+        interpolation_context:
+          - raw_schema
   SchemaTypeIdentifier:
     title: Schema Type Identifier
     description: (This component is experimental. Use at your own risk.) Identifies schema details for dynamic schema extraction and processing.
@@ -1862,6 +1866,7 @@ definitions:
             - "$ref": "#/definitions/KeysToLower"
             - "$ref": "#/definitions/KeysToSnakeCase"
             - "$ref": "#/definitions/FlattenFields"
+            - "$ref": "#/definitions/DpathFlattenFields"
             - "$ref": "#/definitions/KeysReplace"
       schema_type_identifier:
         "$ref": "#/definitions/SchemaTypeIdentifier"
@@ -1966,6 +1971,33 @@ definitions:
       $parameters:
         type: object
         additionalProperties: true
+  DpathFlattenFields:
+    title: Dpath Flatten Fields
+    description: A transformation that flatten field values to the to top of the record.
+    type: object
+    required:
+      - type
+      - field_path
+    properties:
+      type:
+        type: string
+        enum: [DpathFlattenFields]
+      field_path:
+        title: Field Path
+        description: A path to field that needs to be flattened.
+        type: array
+        items:
+          - type: string
+        examples:
+          - ["data"]
+          - ["data", "*", "field"]
+      delete_origin_value:
+        title: Delete Origin Value
+        description: Whether to delete the origin value or keep it. Default is False.
+        type: boolean
+      $parameters:
+        type: object
+        additionalProperties: true
   KeysReplace:
     title: Keys Replace
     description: A transformation that replaces symbols in keys.
@@ -2068,24 +2100,6 @@ definitions:
       $parameters:
         type: object
         additionalProperties: true
-  ZipfileDecoder:
-    title: Zipfile Decoder
-    description: Decoder for response data that is returned as zipfile(s).
-    type: object
-    additionalProperties: true
-    required:
-      - type
-      - parser
-    properties:
-      type:
-        type: string
-        enum: [ZipfileDecoder]
-      parser:
-        title: Parser
-        description: Parser to parse the decompressed data from the zipfile(s).
-        anyOf:
-          - "$ref": "#/definitions/GzipParser"
-          - "$ref": "#/definitions/JsonParser"
   ListPartitionRouter:
     title: List Partition Router
     description: A Partition router that specifies a list of attributes where each attribute describes a portion of the complete data set for a stream. During a sync, each value is iterated over and can be used as input to outbound API requests.
@@ -2914,7 +2928,6 @@ definitions:
           - "$ref": "#/definitions/XmlDecoder"
           - "$ref": "#/definitions/GzipJsonDecoder"
           - "$ref": "#/definitions/CompositeRawDecoder"
-          - "$ref": "#/definitions/ZipfileDecoder"
       $parameters:
         type: object
         additionalProperties: true
@@ -3113,8 +3126,6 @@ definitions:
           - "$ref": "#/definitions/IterableDecoder"
           - "$ref": "#/definitions/XmlDecoder"
           - "$ref": "#/definitions/GzipJsonDecoder"
-          - "$ref": "#/definitions/CompositeRawDecoder"
-          - "$ref": "#/definitions/ZipfileDecoder"
       download_decoder:
         title: Download Decoder
         description: Component decoding the download response so records can be extracted.
@@ -3125,8 +3136,6 @@ definitions:
           - "$ref": "#/definitions/IterableDecoder"
           - "$ref": "#/definitions/XmlDecoder"
           - "$ref": "#/definitions/GzipJsonDecoder"
-          - "$ref": "#/definitions/CompositeRawDecoder"
-          - "$ref": "#/definitions/ZipfileDecoder"
       $parameters:
         type: object
         additionalProperties: true

{airbyte_cdk-6.21.0.dev0 → airbyte_cdk-6.21.1.dev0}/airbyte_cdk/sources/declarative/decoders/__init__.py

@@ -2,12 +2,7 @@
 # Copyright (c) 2023 Airbyte, Inc., all rights reserved.
 #
 
-from airbyte_cdk.sources.declarative.decoders.composite_raw_decoder import (
-    CompositeRawDecoder,
-    GzipParser,
-    JsonParser,
-    Parser,
-)
+from airbyte_cdk.sources.declarative.decoders.composite_raw_decoder import CompositeRawDecoder
 from airbyte_cdk.sources.declarative.decoders.decoder import Decoder
 from airbyte_cdk.sources.declarative.decoders.json_decoder import (
     GzipJsonDecoder,
@@ -20,18 +15,15 @@ from airbyte_cdk.sources.declarative.decoders.pagination_decoder_decorator impor
     PaginationDecoderDecorator,
 )
 from airbyte_cdk.sources.declarative.decoders.xml_decoder import XmlDecoder
-from airbyte_cdk.sources.declarative.decoders.zipfile_decoder import ZipfileDecoder
 
 __all__ = [
     "Decoder",
     "CompositeRawDecoder",
     "JsonDecoder",
-    "JsonParser",
     "JsonlDecoder",
     "IterableDecoder",
     "GzipJsonDecoder",
     "NoopDecoder",
     "PaginationDecoderDecorator",
     "XmlDecoder",
-    "ZipfileDecoder",
 ]

{airbyte_cdk-6.21.0.dev0 → airbyte_cdk-6.21.1.dev0}/airbyte_cdk/sources/declarative/models/declarative_component_schema.py

@@ -719,6 +719,7 @@ class HttpResponseFilter(BaseModel):
 class TypesMap(BaseModel):
     target_type: Union[str, List[str]]
     current_type: Union[str, List[str]]
+    condition: Optional[str]
 
 
 class SchemaTypeIdentifier(BaseModel):
@@ -791,6 +792,25 @@ class FlattenFields(BaseModel):
     parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
 
 
+class DpathFlattenFields(BaseModel):
+    type: Literal["DpathFlattenFields"]
+    field_path: List[str] = Field(
+        ...,
+        description="A path to field that needs to be flattened.",
+        examples=[
+            ["data"],
+            ["data", "*", "field"],
+        ],
+        title="Field Path",
+    )
+    delete_origin_value: Optional[bool] = Field(
+        False,
+        description="Whether to delete the origin value or keep it. Default is False.",
+        title="Delete Origin Value",
+    )
+    parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
+
+
 class KeysReplace(BaseModel):
     type: Literal["KeysReplace"]
     old: str = Field(
@@ -1222,6 +1242,9 @@ class LegacySessionTokenAuthenticator(BaseModel):
 
 
 class JsonParser(BaseModel):
+    class Config:
+        extra = Extra.allow
+
     type: Literal["JsonParser"]
     encoding: Optional[str] = "utf-8"
 
@@ -1657,18 +1680,6 @@ class CompositeErrorHandler(BaseModel):
     parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
 
 
-class ZipfileDecoder(BaseModel):
-    class Config:
-        extra = Extra.allow
-
-    type: Literal["ZipfileDecoder"]
-    parser: Union[GzipParser, JsonParser] = Field(
-        ...,
-        description="Parser to parse the decompressed data from the zipfile(s).",
-        title="Parser",
-    )
-
-
 class CompositeRawDecoder(BaseModel):
     type: Literal["CompositeRawDecoder"]
     parser: Union[GzipParser, JsonParser, JsonLineParser, CsvParser]
@@ -1818,6 +1829,7 @@ class DeclarativeStream(BaseModel):
                 KeysToLower,
                 KeysToSnakeCase,
                 FlattenFields,
+                DpathFlattenFields,
                 KeysReplace,
             ]
         ]
@@ -1874,7 +1886,7 @@ class SessionTokenAuthenticator(BaseModel):
         description="Authentication method to use for requests sent to the API, specifying how to inject the session token.",
         title="Data Request Authentication",
     )
-    decoder: Optional[Union[JsonDecoder, XmlDecoder, CompositeRawDecoder]] = Field(
+    decoder: Optional[Union[JsonDecoder, XmlDecoder]] = Field(
         None, description="Component used to decode the response.", title="Decoder"
     )
     parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
@@ -1993,6 +2005,7 @@ class DynamicSchemaLoader(BaseModel):
                 KeysToLower,
                 KeysToSnakeCase,
                 FlattenFields,
+                DpathFlattenFields,
                 KeysReplace,
             ]
         ]
@@ -2079,7 +2092,6 @@ class SimpleRetriever(BaseModel):
             XmlDecoder,
             GzipJsonDecoder,
             CompositeRawDecoder,
-            ZipfileDecoder,
         ]
     ] = Field(
         None,
@@ -2156,8 +2168,6 @@ class AsyncRetriever(BaseModel):
             IterableDecoder,
             XmlDecoder,
             GzipJsonDecoder,
-            CompositeRawDecoder,
-            ZipfileDecoder,
         ]
     ] = Field(
         None,
@@ -2172,8 +2182,6 @@ class AsyncRetriever(BaseModel):
             IterableDecoder,
             XmlDecoder,
             GzipJsonDecoder,
-            CompositeRawDecoder,
-            ZipfileDecoder,
         ]
     ] = Field(
         None,

{airbyte_cdk-6.21.0.dev0 → airbyte_cdk-6.21.1.dev0}/airbyte_cdk/sources/declarative/parsers/model_to_component_factory.py

@@ -66,7 +66,6 @@ from airbyte_cdk.sources.declarative.decoders import (
     JsonlDecoder,
     PaginationDecoderDecorator,
     XmlDecoder,
-    ZipfileDecoder,
 )
 from airbyte_cdk.sources.declarative.decoders.composite_raw_decoder import (
     CompositeRawDecoder,
@@ -211,6 +210,9 @@ from airbyte_cdk.sources.declarative.models.declarative_component_schema import
 from airbyte_cdk.sources.declarative.models.declarative_component_schema import (
     DpathExtractor as DpathExtractorModel,
 )
+from airbyte_cdk.sources.declarative.models.declarative_component_schema import (
+    DpathFlattenFields as DpathFlattenFieldsModel,
+)
 from airbyte_cdk.sources.declarative.models.declarative_component_schema import (
     DynamicSchemaLoader as DynamicSchemaLoaderModel,
 )
@@ -357,9 +359,6 @@ from airbyte_cdk.sources.declarative.models.declarative_component_schema import
 from airbyte_cdk.sources.declarative.models.declarative_component_schema import (
     XmlDecoder as XmlDecoderModel,
 )
-from airbyte_cdk.sources.declarative.models.declarative_component_schema import (
-    ZipfileDecoder as ZipfileDecoderModel,
-)
 from airbyte_cdk.sources.declarative.partition_routers import (
     CartesianProductStreamSlicer,
     ListPartitionRouter,
@@ -434,6 +433,9 @@ from airbyte_cdk.sources.declarative.transformations import (
     RemoveFields,
 )
 from airbyte_cdk.sources.declarative.transformations.add_fields import AddedFieldDefinition
+from airbyte_cdk.sources.declarative.transformations.dpath_flatten_fields import (
+    DpathFlattenFields,
+)
 from airbyte_cdk.sources.declarative.transformations.flatten_fields import (
     FlattenFields,
 )
@@ -542,6 +544,7 @@ class ModelToComponentFactory:
             KeysToSnakeCaseModel: self.create_keys_to_snake_transformation,
             KeysReplaceModel: self.create_keys_replace_transformation,
             FlattenFieldsModel: self.create_flatten_fields,
+            DpathFlattenFieldsModel: self.create_dpath_flatten_fields,
             IterableDecoderModel: self.create_iterable_decoder,
             XmlDecoderModel: self.create_xml_decoder,
             JsonFileSchemaLoaderModel: self.create_json_file_schema_loader,
@@ -575,7 +578,6 @@ class ModelToComponentFactory:
             ConfigComponentsResolverModel: self.create_config_components_resolver,
             StreamConfigModel: self.create_stream_config,
             ComponentMappingDefinitionModel: self.create_components_mapping_definition,
-            ZipfileDecoderModel: self.create_zipfile_decoder,
         }
 
         # Needed for the case where we need to perform a second parse on the fields of a custom component
@@ -677,6 +679,19 @@ class ModelToComponentFactory:
             flatten_lists=model.flatten_lists if model.flatten_lists is not None else True
         )
 
+    def create_dpath_flatten_fields(
+        self, model: DpathFlattenFieldsModel, config: Config, **kwargs: Any
+    ) -> DpathFlattenFields:
+        model_field_path: List[Union[InterpolatedString, str]] = [x for x in model.field_path]
+        return DpathFlattenFields(
+            config=config,
+            field_path=model_field_path,
+            delete_origin_value=model.delete_origin_value
+            if model.delete_origin_value is not None
+            else False,
+            parameters=model.parameters or {},
+        )
+
     @staticmethod
     def _json_schema_type_name_to_type(value_type: Optional[ValueType]) -> Optional[Type[Any]]:
         if not value_type:
@@ -1701,7 +1716,11 @@ class ModelToComponentFactory:
 
     @staticmethod
     def create_types_map(model: TypesMapModel, **kwargs: Any) -> TypesMap:
-        return TypesMap(target_type=model.target_type, current_type=model.current_type)
+        return TypesMap(
+            target_type=model.target_type,
+            current_type=model.current_type,
+            condition=model.condition if model.condition is not None else "True",
+        )
 
     def create_schema_type_identifier(
         self, model: SchemaTypeIdentifierModel, config: Config, **kwargs: Any
@@ -1801,12 +1820,6 @@ class ModelToComponentFactory:
     ) -> GzipJsonDecoder:
         return GzipJsonDecoder(parameters={}, encoding=model.encoding)
 
-    def create_zipfile_decoder(
-        self, model: ZipfileDecoderModel, config: Config, **kwargs: Any
-    ) -> ZipfileDecoder:
-        parser = self._create_component_from_model(model=model.parser, config=config)
-        return ZipfileDecoder(parser=parser)
-
     def create_gzip_parser(
         self, model: GzipParserModel, config: Config, **kwargs: Any
     ) -> GzipParser:

{airbyte_cdk-6.21.0.dev0 → airbyte_cdk-6.21.1.dev0}/airbyte_cdk/sources/declarative/schema/dynamic_schema_loader.py

@@ -10,6 +10,7 @@ from typing import Any, List, Mapping, MutableMapping, Optional, Union
 import dpath
 from typing_extensions import deprecated
 
+from airbyte_cdk.sources.declarative.interpolation.interpolated_boolean import InterpolatedBoolean
 from airbyte_cdk.sources.declarative.interpolation.interpolated_string import InterpolatedString
 from airbyte_cdk.sources.declarative.retrievers.retriever import Retriever
 from airbyte_cdk.sources.declarative.schema.schema_loader import SchemaLoader
@@ -53,6 +54,7 @@ class TypesMap:
 
     target_type: Union[List[str], str]
     current_type: Union[List[str], str]
+    condition: Optional[str]
 
 
 @deprecated("This class is experimental. Use at your own risk.", category=ExperimentalClassWarning)
@@ -177,7 +179,7 @@ class DynamicSchemaLoader(SchemaLoader):
             if field_type_path
             else "string"
         )
-        mapped_field_type = self._replace_type_if_not_valid(raw_field_type)
+        mapped_field_type = self._replace_type_if_not_valid(raw_field_type, raw_schema)
         if (
             isinstance(mapped_field_type, list)
             and len(mapped_field_type) == 2
@@ -194,14 +196,22 @@ class DynamicSchemaLoader(SchemaLoader):
             )
 
     def _replace_type_if_not_valid(
-        self, field_type: Union[List[str], str]
+        self,
+        field_type: Union[List[str], str],
+        raw_schema: MutableMapping[str, Any],
     ) -> Union[List[str], str]:
         """
         Replaces a field type if it matches a type mapping in `types_map`.
         """
         if self.schema_type_identifier.types_mapping:
             for types_map in self.schema_type_identifier.types_mapping:
-                if field_type == types_map.current_type:
+                # conditional is optional param, setting to true if not provided
+                condition = InterpolatedBoolean(
+                    condition=types_map.condition if types_map.condition is not None else "True",
+                    parameters={},
+                ).eval(config=self.config, raw_schema=raw_schema)
+
+                if field_type == types_map.current_type and condition:
                     return types_map.target_type
         return field_type
 

airbyte_cdk-6.21.1.dev0/airbyte_cdk/sources/declarative/transformations/dpath_flatten_fields.py

@@ -0,0 +1,55 @@
+from dataclasses import InitVar, dataclass
+from typing import Any, Dict, List, Mapping, Optional, Union
+
+import dpath
+
+from airbyte_cdk.sources.declarative.interpolation.interpolated_string import InterpolatedString
+from airbyte_cdk.sources.declarative.transformations import RecordTransformation
+from airbyte_cdk.sources.types import Config, StreamSlice, StreamState
+
+
+@dataclass
+class DpathFlattenFields(RecordTransformation):
+    """
+    Flatten fields only for provided path.
+
+    field_path: List[Union[InterpolatedString, str]] path to the field to flatten.
+    delete_origin_value: bool = False whether to delete origin field or keep it. Default is False.
+
+    """
+
+    config: Config
+    field_path: List[Union[InterpolatedString, str]]
+    parameters: InitVar[Mapping[str, Any]]
+    delete_origin_value: bool = False
+
+    def __post_init__(self, parameters: Mapping[str, Any]) -> None:
+        self._field_path = [
+            InterpolatedString.create(path, parameters=parameters) for path in self.field_path
+        ]
+        for path_index in range(len(self.field_path)):
+            if isinstance(self.field_path[path_index], str):
+                self._field_path[path_index] = InterpolatedString.create(
+                    self.field_path[path_index], parameters=parameters
+                )
+
+    def transform(
+        self,
+        record: Dict[str, Any],
+        config: Optional[Config] = None,
+        stream_state: Optional[StreamState] = None,
+        stream_slice: Optional[StreamSlice] = None,
+    ) -> None:
+        path = [path.eval(self.config) for path in self._field_path]
+        if "*" in path:
+            matched = dpath.values(record, path)
+            extracted = matched[0] if matched else None
+        else:
+            extracted = dpath.get(record, path, default=[])
+
+        if isinstance(extracted, dict):
+            conflicts = set(extracted.keys()) & set(record.keys())
+            if not conflicts:
+                if self.delete_origin_value:
+                    dpath.delete(record, path)
+                record.update(extracted)

{airbyte_cdk-6.21.0.dev0 → airbyte_cdk-6.21.1.dev0}/airbyte_cdk/sources/utils/transform.py

@@ -9,6 +9,7 @@ from typing import Any, Callable, Dict, Generator, Mapping, Optional, cast
 
 from jsonschema import Draft7Validator, RefResolver, ValidationError, Validator, validators
 
+MAX_NESTING_DEPTH = 3
 json_to_python_simple = {
     "string": str,
     "number": float,
@@ -225,6 +226,31 @@ class TypeTransformer:
             logger.warning(self.get_error_message(e))
 
     def get_error_message(self, e: ValidationError) -> str:
-        instance_json_type = python_to_json[type(e.instance)]
-        key_path = "." + ".".join(map(str, e.path))
-        return f"Failed to transform value {repr(e.instance)} of type '{instance_json_type}' to '{e.validator_value}', key path: '{key_path}'"
+        """
+        Construct a sanitized error message from a ValidationError instance.
+        """
+        field_path = ".".join(map(str, e.path))
+        type_structure = self._get_type_structure(e.instance)
+
+        return f"Failed to transform value from type '{type_structure}' to type '{e.validator_value}' at path: '{field_path}'"
+
+    def _get_type_structure(self, input_data: Any, current_depth: int = 0) -> Any:
+        """
+        Get the structure of a given input data for use in error message construction.
+        """
+        # Handle null values
+        if input_data is None:
+            return "null"
+
+        # Avoid recursing too deep
+        if current_depth >= MAX_NESTING_DEPTH:
+            return "object" if isinstance(input_data, dict) else python_to_json[type(input_data)]
+
+        if isinstance(input_data, dict):
+            return {
+                key: self._get_type_structure(field_value, current_depth + 1)
+                for key, field_value in input_data.items()
+            }
+
+        else:
+            return python_to_json[type(input_data)]

{airbyte_cdk-6.21.0.dev0 → airbyte_cdk-6.21.1.dev0}/pyproject.toml

@@ -23,7 +23,7 @@ classifiers = [
 keywords = ["airbyte", "connector-development-kit", "cdk"]
 
 # Python CDK uses dynamic versioning: https://github.com/mtkennerly/poetry-dynamic-versioning
-version = "6.21.0.dev0"  # Version will be calculated dynamically.
+version = "6.21.1dev0"  # Version will be calculated dynamically.
 
 [tool.poetry-dynamic-versioning]
 enable = false

airbyte_cdk-6.21.0.dev0/airbyte_cdk/sources/declarative/decoders/zipfile_decoder.py

@@ -1,50 +0,0 @@
-#
-# Copyright (c) 2024 Airbyte, Inc., all rights reserved.
-#
-
-import logging
-import zipfile
-from dataclasses import dataclass
-from io import BytesIO
-from typing import Any, Generator, MutableMapping
-
-import orjson
-import requests
-
-from airbyte_cdk.models import FailureType
-from airbyte_cdk.sources.declarative.decoders import Decoder
-from airbyte_cdk.sources.declarative.decoders.composite_raw_decoder import (
-    Parser,
-)
-from airbyte_cdk.utils import AirbyteTracedException
-
-logger = logging.getLogger("airbyte")
-
-
-@dataclass
-class ZipfileDecoder(Decoder):
-    parser: Parser
-
-    def is_stream_response(self) -> bool:
-        return False
-
-    def decode(
-        self, response: requests.Response
-    ) -> Generator[MutableMapping[str, Any], None, None]:
-        try:
-            zip_file = zipfile.ZipFile(BytesIO(response.content))
-        except zipfile.BadZipFile as e:
-            logger.error(
-                f"Received an invalid zip file in response to URL: {response.request.url}. "
-                f"The size of the response body is: {len(response.content)}"
-            )
-            raise AirbyteTracedException(
-                message="Received an invalid zip file in response.",
-                internal_message=f"Received an invalid zip file in response to URL: {response.request.url}.",
-                failure_type=FailureType.system_error,
-            ) from e
-
-        for file_name in zip_file.namelist():
-            unzipped_content = zip_file.read(file_name)
-            buffered_content = BytesIO(unzipped_content)
-            yield from self.parser.parse(buffered_content)