feat: Adds ZipfileDecoder component

airbyte_cdk/sources/declarative/declarative_component_schema.yaml

@@ -2067,6 +2067,24 @@ 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"
+          - "$ref": "#/definitions/JsonParser"
   ListPartitionRouter:
     title: List Partition Router
     description: A Partition router that specifies a list of attributes where each attribute describes a portion of the complete data set for a stream. During a sync, each value is iterated over and can be used as input to outbound API requests.
@@ -2895,6 +2913,7 @@ definitions:
           - "$ref": "#/definitions/XmlDecoder"
           - "$ref": "#/definitions/GzipJsonDecoder"
           - "$ref": "#/definitions/CompositeRawDecoder"
+          - "$ref": "#/definitions/ZipfileDecoder"
       $parameters:
         type: object
         additionalProperties: true

airbyte_cdk/sources/declarative/decoders/__init__.py

@@ -2,7 +2,12 @@
 # Copyright (c) 2023 Airbyte, Inc., all rights reserved.
 #
 
-from airbyte_cdk.sources.declarative.decoders.composite_raw_decoder import CompositeRawDecoder
+from airbyte_cdk.sources.declarative.decoders.composite_raw_decoder import (
+    CompositeRawDecoder,
+    GzipParser,
+    JsonParser,
+    Parser,
+)
 from airbyte_cdk.sources.declarative.decoders.decoder import Decoder
 from airbyte_cdk.sources.declarative.decoders.json_decoder import (
     GzipJsonDecoder,
@@ -15,15 +20,18 @@
     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/sources/declarative/decoders/zipfile_decoder.py

@@ -0,0 +1,47 @@
+#
+# Copyright (c) 2024 Airbyte, Inc., all rights reserved.
+#
+
+import logging
+import zipfile
+from dataclasses import InitVar, dataclass
+from io import BufferedReader, BytesIO
+from typing import Any, Generator, Mapping, MutableMapping, Optional
+
+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 True
+
+    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 filename in zip_file.namelist():
+            with zip_file.open(filename) as file:
+                yield from self.parser.parse(BytesIO(file.read()))
+                zip_file.close()

airbyte_cdk/sources/declarative/models/declarative_component_schema.py

@@ -1222,9 +1222,6 @@ class LegacySessionTokenAuthenticator(BaseModel):
 
 
 class JsonParser(BaseModel):
-    class Config:
-        extra = Extra.allow
-
     type: Literal["JsonParser"]
     encoding: Optional[str] = "utf-8"
 
@@ -1660,6 +1657,18 @@ 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]
@@ -2070,6 +2079,7 @@ class SimpleRetriever(BaseModel):
             XmlDecoder,
             GzipJsonDecoder,
             CompositeRawDecoder,
+            ZipfileDecoder,
         ]
     ] = Field(
         None,

airbyte_cdk/sources/declarative/parsers/model_to_component_factory.py

@@ -66,6 +66,7 @@
     JsonlDecoder,
     PaginationDecoderDecorator,
     XmlDecoder,
+    ZipfileDecoder,
 )
 from airbyte_cdk.sources.declarative.decoders.composite_raw_decoder import (
     CompositeRawDecoder,
@@ -356,6 +357,9 @@
 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,
@@ -571,6 +575,7 @@ def _init_mappings(self) -> None:
             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
@@ -1796,6 +1801,12 @@ def create_gzipjson_decoder(
     ) -> 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:

unit_tests/sources/declarative/decoders/test_zipfile_decoder.py

@@ -0,0 +1,43 @@
+#
+# Copyright (c) 2024 Airbyte, Inc., all rights reserved.
+#
+import gzip
+import json
+import zipfile
+from io import BytesIO
+from typing import Union
+
+import pytest
+import requests
+
+from airbyte_cdk.sources.declarative.decoders import ZipfileDecoder
+from airbyte_cdk.sources.declarative.decoders.parsers import JsonParser
+
+
+def create_zip_from_dict(data: Union[dict, list]):
+    zip_buffer = BytesIO()
+    with zipfile.ZipFile(zip_buffer, mode="w") as zip_file:
+        zip_file.writestr("data.json", data)
+    zip_buffer.seek(0)
+    return zip_buffer.getvalue()
+
+
+@pytest.mark.parametrize(
+    "json_data",
+    [
+        {"test": "test"},
+        [{"id": 1}, {"id": 2}],
+    ],
+)
+def test_zipfile_decoder_with_valid_response(requests_mock, json_data):
+    zipfile_decoder = ZipfileDecoder(parameters={}, parser=JsonParser)
+    compressed_data = gzip.compress(json.dumps(json_data).encode())
+    zipped_data = create_zip_from_dict(compressed_data)
+    requests_mock.register_uri("GET", "https://airbyte.io/", content=zipped_data)
+    response = requests.get("https://airbyte.io/")
+
+    if isinstance(json_data, list):
+        for i, actual in enumerate(zipfile_decoder.decode(response=response)):
+            assert actual == json_data[i]
+    else:
+        assert next(zipfile_decoder.decode(response=response)) == json_data