airbyte-cdk 6.34.0.dev2__py3-none-any.whl → 6.34.1.dev0__py3-none-any.whl

airbyte_cdk/sources/declarative/concurrent_declarative_source.py

@@ -3,7 +3,7 @@
 #
 
 import logging
-from typing import Any, Generic, Iterator, List, Mapping, MutableMapping, Optional, Tuple
+from typing import Any, Generic, Iterator, List, Mapping, Optional, Tuple
 
 from airbyte_cdk.models import (
     AirbyteCatalog,
@@ -19,7 +19,6 @@ from airbyte_cdk.sources.declarative.extractors import RecordSelector
 from airbyte_cdk.sources.declarative.extractors.record_filter import (
     ClientSideIncrementalRecordFilterDecorator,
 )
-from airbyte_cdk.sources.declarative.incremental import ConcurrentPerPartitionCursor
 from airbyte_cdk.sources.declarative.incremental.datetime_based_cursor import DatetimeBasedCursor
 from airbyte_cdk.sources.declarative.incremental.per_partition_with_global import (
     PerPartitionWithGlobalCursor,
@@ -47,6 +46,7 @@ from airbyte_cdk.sources.declarative.types import ConnectionDefinition
 from airbyte_cdk.sources.source import TState
 from airbyte_cdk.sources.streams import Stream
 from airbyte_cdk.sources.streams.concurrent.abstract_stream import AbstractStream
+from airbyte_cdk.sources.streams.concurrent.abstract_stream_facade import AbstractStreamFacade
 from airbyte_cdk.sources.streams.concurrent.availability_strategy import (
     AlwaysAvailableAvailabilityStrategy,
 )
@@ -121,6 +121,12 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
             message_repository=self.message_repository,
         )
 
+    # TODO: Remove this. This property is necessary to safely migrate Stripe during the transition state.
+    @property
+    def is_partially_declarative(self) -> bool:
+        """This flag used to avoid unexpected AbstractStreamFacade processing as concurrent streams."""
+        return False
+
     def read(
         self,
         logger: logging.Logger,
@@ -224,7 +230,6 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                     stream_state = self._connector_state_manager.get_stream_state(
                         stream_name=declarative_stream.name, namespace=declarative_stream.namespace
                     )
-                    stream_state = self._migrate_state(declarative_stream, stream_state)
 
                     retriever = self._get_retriever(declarative_stream, stream_state)
 
@@ -233,7 +238,7 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                     ):
                         cursor = declarative_stream.retriever.stream_slicer.stream_slicer
 
-                        if not isinstance(cursor, ConcurrentCursor | ConcurrentPerPartitionCursor):
+                        if not isinstance(cursor, ConcurrentCursor):
                             # This should never happen since we instantiate ConcurrentCursor in
                             # model_to_component_factory.py
                             raise ValueError(
@@ -332,8 +337,6 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                     stream_state = self._connector_state_manager.get_stream_state(
                         stream_name=declarative_stream.name, namespace=declarative_stream.namespace
                     )
-                    stream_state = self._migrate_state(declarative_stream, stream_state)
-
                     partition_router = declarative_stream.retriever.stream_slicer._partition_router
 
                     perpartition_cursor = (
@@ -375,6 +378,14 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                     )
                 else:
                     synchronous_streams.append(declarative_stream)
+            # TODO: Remove this. This check is necessary to safely migrate Stripe during the transition state.
+            # Condition below needs to ensure that concurrent support is not lost for sources that already support
+            # it before migration, but now are only partially migrated to declarative implementation (e.g., Stripe).
+            elif (
+                isinstance(declarative_stream, AbstractStreamFacade)
+                and self.is_partially_declarative
+            ):
+                concurrent_streams.append(declarative_stream.get_underlying_stream())
             else:
                 synchronous_streams.append(declarative_stream)
 
@@ -478,21 +489,10 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
             # Also a temporary hack. In the legacy Stream implementation, as part of the read,
             # set_initial_state() is called to instantiate incoming state on the cursor. Although we no
             # longer rely on the legacy low-code cursor for concurrent checkpointing, low-code components
-            # like StopConditionPaginationStrategyDecorator still rely on a DatetimeBasedCursor that is
-            # properly initialized with state.
+            # like StopConditionPaginationStrategyDecorator and ClientSideIncrementalRecordFilterDecorator
+            # still rely on a DatetimeBasedCursor that is properly initialized with state.
             if retriever.cursor:
                 retriever.cursor.set_initial_state(stream_state=stream_state)
-
-            # Similar to above, the ClientSideIncrementalRecordFilterDecorator cursor is a separate instance
-            # from the one initialized on the SimpleRetriever, so it also must also have state initialized
-            # for semi-incremental streams using is_client_side_incremental to filter properly
-            if isinstance(retriever.record_selector, RecordSelector) and isinstance(
-                retriever.record_selector.record_filter, ClientSideIncrementalRecordFilterDecorator
-            ):
-                retriever.record_selector.record_filter._cursor.set_initial_state(
-                    stream_state=stream_state
-                )  # type: ignore  # After non-concurrent cursors are deprecated we can remove these cursor workarounds
-
             # We zero it out here, but since this is a cursor reference, the state is still properly
             # instantiated for the other components that reference it
             retriever.cursor = None
@@ -524,14 +524,3 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                 if stream.stream.name not in concurrent_stream_names
             ]
         )
-
-    @staticmethod
-    def _migrate_state(
-        declarative_stream: DeclarativeStream, stream_state: MutableMapping[str, Any]
-    ) -> MutableMapping[str, Any]:
-        for state_migration in declarative_stream.state_migrations:
-            if state_migration.should_migrate(stream_state):
-                # The state variable is expected to be mutable but the migrate method returns an immutable mapping.
-                stream_state = dict(state_migration.migrate(stream_state))
-
-        return stream_state

airbyte_cdk/sources/declarative/declarative_component_schema.yaml

@@ -40,8 +40,6 @@ properties:
     "$ref": "#/definitions/Spec"
   concurrency_level:
     "$ref": "#/definitions/ConcurrencyLevel"
-  api_budget:
-    "$ref": "#/definitions/HTTPAPIBudget"
   metadata:
     type: object
     description: For internal Airbyte use only - DO NOT modify manually. Used by consumers of declarative manifests for storing related metadata.
@@ -796,7 +794,7 @@ definitions:
         description: This option is used to adjust the upper and lower boundaries of each datetime window to beginning and end of the provided target period (day, week, month)
         type: object
         required:
-          - target
+        - target
         properties:
           target:
             title: Target
@@ -1367,170 +1365,6 @@ definitions:
       $parameters:
         type: object
         additional_properties: true
-  HTTPAPIBudget:
-    title: HTTP API Budget
-    description: >
-      Defines how many requests can be made to the API in a given time frame. `HTTPAPIBudget` extracts the remaining
-      call count and the reset time from HTTP response headers using the header names provided by
-      `ratelimit_remaining_header` and `ratelimit_reset_header`. Only requests using `HttpRequester`
-      are rate-limited; custom components that bypass `HttpRequester` are not covered by this budget.
-    type: object
-    required:
-      - type
-      - policies
-    properties:
-      type:
-        type: string
-        enum: [HTTPAPIBudget]
-      policies:
-        title: Policies
-        description: List of call rate policies that define how many calls are allowed.
-        type: array
-        items:
-          anyOf:
-            - "$ref": "#/definitions/FixedWindowCallRatePolicy"
-            - "$ref": "#/definitions/MovingWindowCallRatePolicy"
-            - "$ref": "#/definitions/UnlimitedCallRatePolicy"
-      ratelimit_reset_header:
-        title: Rate Limit Reset Header
-        description: The HTTP response header name that indicates when the rate limit resets.
-        type: string
-        default: "ratelimit-reset"
-      ratelimit_remaining_header:
-        title: Rate Limit Remaining Header
-        description: The HTTP response header name that indicates the number of remaining allowed calls.
-        type: string
-        default: "ratelimit-remaining"
-      status_codes_for_ratelimit_hit:
-        title: Status Codes for Rate Limit Hit
-        description: List of HTTP status codes that indicate a rate limit has been hit.
-        type: array
-        items:
-          type: integer
-        default: [429]
-    additionalProperties: true
-  FixedWindowCallRatePolicy:
-    title: Fixed Window Call Rate Policy
-    description: A policy that allows a fixed number of calls within a specific time window.
-    type: object
-    required:
-      - type
-      - period
-      - call_limit
-      - matchers
-    properties:
-      type:
-        type: string
-        enum: [FixedWindowCallRatePolicy]
-      period:
-        title: Period
-        description: The time interval for the rate limit window.
-        type: string
-      call_limit:
-        title: Call Limit
-        description: The maximum number of calls allowed within the period.
-        type: integer
-      matchers:
-        title: Matchers
-        description: List of matchers that define which requests this policy applies to.
-        type: array
-        items:
-          "$ref": "#/definitions/HttpRequestRegexMatcher"
-    additionalProperties: true
-  MovingWindowCallRatePolicy:
-    title: Moving Window Call Rate Policy
-    description: A policy that allows a fixed number of calls within a moving time window.
-    type: object
-    required:
-      - type
-      - rates
-      - matchers
-    properties:
-      type:
-        type: string
-        enum: [MovingWindowCallRatePolicy]
-      rates:
-        title: Rates
-        description: List of rates that define the call limits for different time intervals.
-        type: array
-        items:
-          "$ref": "#/definitions/Rate"
-      matchers:
-        title: Matchers
-        description: List of matchers that define which requests this policy applies to.
-        type: array
-        items:
-          "$ref": "#/definitions/HttpRequestRegexMatcher"
-    additionalProperties: true
-  UnlimitedCallRatePolicy:
-    title: Unlimited Call Rate Policy
-    description: A policy that allows unlimited calls for specific requests.
-    type: object
-    required:
-      - type
-      - matchers
-    properties:
-      type:
-        type: string
-        enum: [UnlimitedCallRatePolicy]
-      matchers:
-        title: Matchers
-        description: List of matchers that define which requests this policy applies to.
-        type: array
-        items:
-          "$ref": "#/definitions/HttpRequestRegexMatcher"
-    additionalProperties: true
-  Rate:
-    title: Rate
-    description: Defines a rate limit with a specific number of calls allowed within a time interval.
-    type: object
-    required:
-      - limit
-      - interval
-    properties:
-      limit:
-        title: Limit
-        description: The maximum number of calls allowed within the interval.
-        type: integer
-      interval:
-        title: Interval
-        description: The time interval for the rate limit.
-        type: string
-        examples:
-          - "PT1H"
-          - "P1D"
-    additionalProperties: true
-  HttpRequestRegexMatcher:
-    title: HTTP Request Matcher
-    description: >
-      Matches HTTP requests based on method, base URL, URL path pattern, query parameters, and headers.
-      Use `url_base` to specify the scheme and host (without trailing slash) and
-      `url_path_pattern` to apply a regex to the request path.
-    type: object
-    properties:
-      method:
-        title: Method
-        description: The HTTP method to match (e.g., GET, POST).
-        type: string
-      url_base:
-        title: URL Base
-        description: The base URL (scheme and host, e.g. "https://api.example.com") to match.
-        type: string
-      url_path_pattern:
-        title: URL Path Pattern
-        description: A regular expression pattern to match the URL path.
-        type: string
-      params:
-        title: Parameters
-        description: The query parameters to match.
-        type: object
-        additionalProperties: true
-      headers:
-        title: Headers
-        description: The headers to match.
-        type: object
-        additionalProperties: true
-    additionalProperties: true
   DefaultErrorHandler:
     title: Default Error Handler
     description: Component defining how to handle errors. Default behavior includes only retrying server errors (HTTP 5XX) and too many requests (HTTP 429) with an exponential backoff.
@@ -1715,6 +1549,7 @@ definitions:
         anyOf:
           - "$ref": "#/definitions/JsonDecoder"
           - "$ref": "#/definitions/XmlDecoder"
+          - "$ref": "#/definitions/CompositeRawDecoder"
       $parameters:
         type: object
         additionalProperties: true
@@ -2298,6 +2133,23 @@ definitions:
       $parameters:
         type: object
         additionalProperties: true
+  GzipJsonDecoder:
+    title: GzipJson Decoder
+    description: Use this if the response is Gzip compressed Json.
+    type: object
+    additionalProperties: true
+    required:
+      - type
+    properties:
+      type:
+        type: string
+        enum: [GzipJsonDecoder]
+      encoding:
+        type: string
+        default: utf-8
+      $parameters:
+        type: object
+        additionalProperties: true
   ZipfileDecoder:
     title: Zipfile Decoder
     description: Decoder for response data that is returned as zipfile(s).
@@ -2305,19 +2157,19 @@ definitions:
     additionalProperties: true
     required:
       - type
-      - decoder
+      - parser
     properties:
       type:
         type: string
         enum: [ZipfileDecoder]
-      decoder:
+      parser:
         title: Parser
         description: Parser to parse the decompressed data from the zipfile(s).
         anyOf:
-          - "$ref": "#/definitions/CsvDecoder"
-          - "$ref": "#/definitions/GzipDecoder"
-          - "$ref": "#/definitions/JsonDecoder"
-          - "$ref": "#/definitions/JsonlDecoder"
+          - "$ref": "#/definitions/GzipParser"
+          - "$ref": "#/definitions/JsonParser"
+          - "$ref": "#/definitions/JsonLineParser"
+          - "$ref": "#/definitions/CsvParser"
   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.
@@ -2995,35 +2847,25 @@ definitions:
         enum: [RequestPath]
   RequestOption:
     title: Request Option
-    description: Specifies the key field or path and where in the request a component's value should be injected.
+    description: Specifies the key field and where in the request a component's value should be injected.
     type: object
     required:
       - type
+      - field_name
       - inject_into
     properties:
       type:
         type: string
         enum: [RequestOption]
       field_name:
-        title: Field Name
-        description: Configures which key should be used in the location that the descriptor is being injected into. We hope to eventually deprecate this field in favor of `field_path` for all request_options, but must currently maintain it for backwards compatibility in the Builder.
+        title: Request Option
+        description: Configures which key should be used in the location that the descriptor is being injected into
         type: string
         examples:
           - segment_id
         interpolation_context:
           - config
           - parameters
-      field_path:
-        title: Field Path
-        description: Configures a path to be used for nested structures in JSON body requests (e.g. GraphQL queries)
-        type: array
-        items:
-          type: string
-        examples:
-          - ["data", "viewer", "id"]
-        interpolation_context:
-          - config
-          - parameters
       inject_into:
         title: Inject Into
         description: Configures where the descriptor should be set on the HTTP requests. Note that request parameters that are already encoded in the URL path will not be duplicated.
@@ -3150,39 +2992,79 @@ definitions:
         description: Component decoding the response so records can be extracted.
         anyOf:
           - "$ref": "#/definitions/CustomDecoder"
-          - "$ref": "#/definitions/CsvDecoder"
-          - "$ref": "#/definitions/GzipDecoder"
           - "$ref": "#/definitions/JsonDecoder"
           - "$ref": "#/definitions/JsonlDecoder"
           - "$ref": "#/definitions/IterableDecoder"
           - "$ref": "#/definitions/XmlDecoder"
+          - "$ref": "#/definitions/GzipJsonDecoder"
+          - "$ref": "#/definitions/CompositeRawDecoder"
           - "$ref": "#/definitions/ZipfileDecoder"
       $parameters:
         type: object
         additionalProperties: true
-  GzipDecoder:
+  CompositeRawDecoder:
+    description: "(This is experimental, use at your own risk)"
     type: object
     required:
       - type
-      - decoder
+      - parser
     properties:
       type:
         type: string
-        enum: [GzipDecoder]
-      decoder:
+        enum: [CompositeRawDecoder]
+      parser:
         anyOf:
-          - "$ref": "#/definitions/CsvDecoder"
-          - "$ref": "#/definitions/GzipDecoder"
-          - "$ref": "#/definitions/JsonDecoder"
-          - "$ref": "#/definitions/JsonlDecoder"
-  CsvDecoder:
+          - "$ref": "#/definitions/GzipParser"
+          - "$ref": "#/definitions/JsonParser"
+          - "$ref": "#/definitions/JsonLineParser"
+          - "$ref": "#/definitions/CsvParser"
+  #    PARSERS
+  GzipParser:
+    type: object
+    required:
+      - type
+      - inner_parser
+    properties:
+      type:
+        type: string
+        enum: [GzipParser]
+      inner_parser:
+        anyOf:
+          - "$ref": "#/definitions/JsonLineParser"
+          - "$ref": "#/definitions/CsvParser"
+          - "$ref": "#/definitions/JsonParser"
+  JsonParser:
+    title: JsonParser
+    description: Parser used for parsing str, bytes, or bytearray data and returning data in a dictionary format.
+    type: object
+    required:
+      - type
+    properties:
+      type:
+        type: string
+        enum: [JsonParser]
+      encoding:
+        type: string
+        default: utf-8
+  JsonLineParser:
+    type: object
+    required:
+      - type
+    properties:
+      type:
+        type: string
+        enum: [JsonLineParser]
+      encoding:
+        type: string
+        default: utf-8
+  CsvParser:
     type: object
     required:
       - type
     properties:
       type:
         type: string
-        enum: [CsvDecoder]
+        enum: [CsvParser]
       encoding:
         type: string
         default: utf-8
@@ -3310,24 +3192,24 @@ definitions:
         description: Component decoding the response so records can be extracted.
         anyOf:
           - "$ref": "#/definitions/CustomDecoder"
-          - "$ref": "#/definitions/CsvDecoder"
-          - "$ref": "#/definitions/GzipDecoder"
           - "$ref": "#/definitions/JsonDecoder"
           - "$ref": "#/definitions/JsonlDecoder"
           - "$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.
         anyOf:
           - "$ref": "#/definitions/CustomDecoder"
-          - "$ref": "#/definitions/CsvDecoder"
-          - "$ref": "#/definitions/GzipDecoder"
           - "$ref": "#/definitions/JsonDecoder"
           - "$ref": "#/definitions/JsonlDecoder"
           - "$ref": "#/definitions/IterableDecoder"
           - "$ref": "#/definitions/XmlDecoder"
+          - "$ref": "#/definitions/GzipJsonDecoder"
+          - "$ref": "#/definitions/CompositeRawDecoder"
           - "$ref": "#/definitions/ZipfileDecoder"
       $parameters:
         type: object

airbyte_cdk/sources/declarative/declarative_stream.py

@@ -138,9 +138,7 @@ class DeclarativeStream(Stream):
         """
         :param: stream_state We knowingly avoid using stream_state as we want cursors to manage their own state.
         """
-        if stream_slice is None or (
-            not isinstance(stream_slice, StreamSlice) and stream_slice == {}
-        ):
+        if stream_slice is None or stream_slice == {}:
             # As the parameter is Optional, many would just call `read_records(sync_mode)` during testing without specifying the field
             # As part of the declarative model without custom components, this should never happen as the CDK would wire up a
             # SinglePartitionRouter that would create this StreamSlice properly

airbyte_cdk/sources/declarative/decoders/__init__.py

@@ -10,8 +10,10 @@ from airbyte_cdk.sources.declarative.decoders.composite_raw_decoder import (
 )
 from airbyte_cdk.sources.declarative.decoders.decoder import Decoder
 from airbyte_cdk.sources.declarative.decoders.json_decoder import (
+    GzipJsonDecoder,
     IterableDecoder,
     JsonDecoder,
+    JsonlDecoder,
 )
 from airbyte_cdk.sources.declarative.decoders.noop_decoder import NoopDecoder
 from airbyte_cdk.sources.declarative.decoders.pagination_decoder_decorator import (
@@ -25,7 +27,9 @@ __all__ = [
     "CompositeRawDecoder",
     "JsonDecoder",
     "JsonParser",
+    "JsonlDecoder",
     "IterableDecoder",
+    "GzipJsonDecoder",
     "NoopDecoder",
     "PaginationDecoderDecorator",
     "XmlDecoder",

airbyte_cdk/sources/declarative/decoders/composite_raw_decoder.py

@@ -1,6 +1,5 @@
 import csv
 import gzip
-import io
 import json
 import logging
 from abc import ABC, abstractmethod
@@ -131,15 +130,11 @@ class CompositeRawDecoder(Decoder):
     """
 
     parser: Parser
-    stream_response: bool = True
 
     def is_stream_response(self) -> bool:
-        return self.stream_response
+        return True
 
     def decode(
         self, response: requests.Response
     ) -> Generator[MutableMapping[str, Any], None, None]:
-        if self.is_stream_response():
-            yield from self.parser.parse(data=response.raw)  # type: ignore[arg-type]
-        else:
-            yield from self.parser.parse(data=io.BytesIO(response.content))
+        yield from self.parser.parse(data=response.raw)  # type: ignore[arg-type]

airbyte_cdk/sources/declarative/decoders/json_decoder.py

@@ -10,24 +10,21 @@ from typing import Any, Generator, List, Mapping, MutableMapping, Optional
 import orjson
 import requests
 
-from airbyte_cdk.sources.declarative.decoders import CompositeRawDecoder, JsonParser
 from airbyte_cdk.sources.declarative.decoders.decoder import Decoder
 
 logger = logging.getLogger("airbyte")
 
 
+@dataclass
 class JsonDecoder(Decoder):
     """
     Decoder strategy that returns the json-encoded content of a response, if any.
-
-    Usually, we would try to instantiate the equivalent `CompositeRawDecoder(parser=JsonParser(), stream_response=False)` but there were specific historical behaviors related to the JsonDecoder that we didn't know if we could remove like the fallback on {} in case of errors.
     """
 
-    def __init__(self, parameters: Mapping[str, Any]):
-        self._decoder = CompositeRawDecoder(parser=JsonParser(), stream_response=False)
+    parameters: InitVar[Mapping[str, Any]]
 
     def is_stream_response(self) -> bool:
-        return self._decoder.is_stream_response()
+        return False
 
     def decode(
         self, response: requests.Response
@@ -35,16 +32,25 @@ class JsonDecoder(Decoder):
         """
         Given the response is an empty string or an emtpy list, the function will return a generator with an empty mapping.
         """
-        has_yielded = False
         try:
-            for element in self._decoder.decode(response):
-                yield element
-                has_yielded = True
-        except Exception:
+            body_json = response.json()
+            yield from self.parse_body_json(body_json)
+        except requests.exceptions.JSONDecodeError:
+            logger.warning(
+                f"Response cannot be parsed into json: {response.status_code=}, {response.text=}"
+            )
             yield {}
 
-        if not has_yielded:
+    @staticmethod
+    def parse_body_json(
+        body_json: MutableMapping[str, Any] | List[MutableMapping[str, Any]],
+    ) -> Generator[MutableMapping[str, Any], None, None]:
+        if not isinstance(body_json, list):
+            body_json = [body_json]
+        if len(body_json) == 0:
             yield {}
+        else:
+            yield from body_json
 
 
 @dataclass
@@ -63,3 +69,43 @@ class IterableDecoder(Decoder):
     ) -> Generator[MutableMapping[str, Any], None, None]:
         for line in response.iter_lines():
             yield {"record": line.decode()}
+
+
+@dataclass
+class JsonlDecoder(Decoder):
+    """
+    Decoder strategy that returns the json-encoded content of the response, if any.
+    """
+
+    parameters: InitVar[Mapping[str, Any]]
+
+    def is_stream_response(self) -> bool:
+        return True
+
+    def decode(
+        self, response: requests.Response
+    ) -> Generator[MutableMapping[str, Any], None, None]:
+        # TODO???: set delimiter? usually it is `\n` but maybe it would be useful to set optional?
+        #  https://github.com/airbytehq/airbyte-internal-issues/issues/8436
+        for record in response.iter_lines():
+            yield orjson.loads(record)
+
+
+@dataclass
+class GzipJsonDecoder(JsonDecoder):
+    encoding: Optional[str]
+
+    def __post_init__(self, parameters: Mapping[str, Any]) -> None:
+        if self.encoding:
+            try:
+                codecs.lookup(self.encoding)
+            except LookupError:
+                raise ValueError(
+                    f"Invalid encoding '{self.encoding}'. Please check provided encoding"
+                )
+
+    def decode(
+        self, response: requests.Response
+    ) -> Generator[MutableMapping[str, Any], None, None]:
+        raw_string = decompress(response.content).decode(encoding=self.encoding or "utf-8")
+        yield from self.parse_body_json(orjson.loads(raw_string))