airbyte-cdk 6.18.0.dev0__tar.gz → 6.18.0.dev2__tar.gz

{airbyte_cdk-6.18.0.dev0 → airbyte_cdk-6.18.0.dev2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: airbyte-cdk
-Version: 6.18.0.dev0
+Version: 6.18.0.dev2
 Summary: A framework for writing Airbyte Connectors.
 License: MIT
 Keywords: airbyte,connector-development-kit,cdk

{airbyte_cdk-6.18.0.dev0 → airbyte_cdk-6.18.0.dev2}/airbyte_cdk/sources/declarative/declarative_component_schema.yaml

@@ -2014,20 +2014,6 @@ definitions:
       $parameters:
         type: object
         additionalProperties: true
-  JsonParser:
-    title: JsonParser
-    description: Parser used for parsing str, bytes, or bytearray data and returning data in a dictionary format.
-    type: object
-    additionalProperties: true
-    required:
-      - type
-    properties:
-      type:
-        type: string
-        enum: [JsonParser]
-      encoding:
-        type: string
-        default: utf-8
   ListPartitionRouter:
     title: List Partition Router
     description: A Partition router that specifies a list of attributes where each attribute describes a portion of the complete data set for a stream. During a sync, each value is iterated over and can be used as input to outbound API requests.
@@ -2873,6 +2859,7 @@ definitions:
       parser:
         anyOf:
           - "$ref": "#/definitions/GzipParser"
+          - "$ref": "#/definitions/JsonParser"
           - "$ref": "#/definitions/JsonLineParser"
           - "$ref": "#/definitions/CsvParser"
   #    PARSERS
@@ -2889,6 +2876,21 @@ definitions:
         anyOf:
           - "$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
+    additionalProperties: true
+    required:
+      - type
+    properties:
+      type:
+        type: string
+        enum: [JsonParser]
+      encoding:
+        type: string
+        default: utf-8
   JsonLineParser:
     type: object
     required:

{airbyte_cdk-6.18.0.dev0 → airbyte_cdk-6.18.0.dev2}/airbyte_cdk/sources/declarative/decoders/composite_raw_decoder.py

@@ -7,6 +7,7 @@ from dataclasses import dataclass
 from io import BufferedIOBase, TextIOWrapper
 from typing import Any, Generator, MutableMapping, Optional
 
+import orjson
 import requests
 
 from airbyte_cdk.models import FailureType
@@ -46,30 +47,43 @@ class GzipParser(Parser):
 
 @dataclass
 class JsonParser(Parser):
-    """
-    Parser strategy for converting JSON-structure str, bytes, or bytearray data into MutableMapping[str, Any].
-    """
-
     encoding: str = "utf-8"
 
     def parse(self, data: BufferedIOBase) -> Generator[MutableMapping[str, Any], None, None]:
+        """
+        Attempts to deserialize data using orjson library. As an extra layer of safety we fallback on the json library to deserialize the data.
+        """
         raw_data = data.read()
-        try:
-            body_json = json.loads(raw_data.decode(self.encoding))
-        except json.JSONDecodeError as exc:
+
+        body_json = self._parse_orjson(raw_data) or self._parse_json(raw_data)
+
+        if body_json is None:
             raise AirbyteTracedException(
-                message="JSON data failed to be parsed. See logs for more inforation.",
-                internal_message=f"JSON data faild to be parsed: {exc=}",
+                message="Response JSON data failed to be parsed. See logs for more information.",
+                internal_message=f"Response JSON data failed to be parsed.",
                 failure_type=FailureType.system_error,
-                exception=exc,
             )
 
-        if not isinstance(body_json, list):
-            body_json = [body_json]
-        if len(body_json) == 0:
-            yield {}
-        else:
+        if isinstance(body_json, list):
             yield from body_json
+        else:
+            yield from [body_json]
+
+    def _parse_orjson(self, raw_data: bytes) -> Optional[Any]:
+        try:
+            return orjson.loads(raw_data.decode(self.encoding))
+        except Exception as exc:
+            logger.warning(
+                f"Failed to parse JSON data using orjson library. Falling back to json library. {exc=}"
+            )
+            return None
+
+    def _parse_json(self, raw_data: bytes) -> Optional[Any]:
+        try:
+            return json.loads(raw_data.decode(self.encoding))
+        except Exception as exc:
+            logger.error(f"Failed to parse JSON data using json library. {exc=}")
+            return None
 
 
 @dataclass

{airbyte_cdk-6.18.0.dev0 → airbyte_cdk-6.18.0.dev2}/airbyte_cdk/sources/declarative/models/declarative_component_schema.py

@@ -805,14 +805,6 @@ class GzipJsonDecoder(BaseModel):
     parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
 
 
-class JsonParser(BaseModel):
-    class Config:
-        extra = Extra.allow
-
-    type: Literal["JsonParser"]
-    encoding: Optional[str] = "utf-8"
-
-
 class MinMaxDatetime(BaseModel):
     type: Literal["MinMaxDatetime"]
     datetime: str = Field(
@@ -1181,6 +1173,14 @@ class LegacySessionTokenAuthenticator(BaseModel):
     parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
 
 
+class JsonParser(BaseModel):
+    class Config:
+        extra = Extra.allow
+
+    type: Literal["JsonParser"]
+    encoding: Optional[str] = "utf-8"
+
+
 class JsonLineParser(BaseModel):
     type: Literal["JsonLineParser"]
     encoding: Optional[str] = "utf-8"
@@ -1579,7 +1579,7 @@ class RecordSelector(BaseModel):
 
 class GzipParser(BaseModel):
     type: Literal["GzipParser"]
-    inner_parser: Union[JsonLineParser, CsvParser]
+    inner_parser: Union[JsonLineParser, CsvParser, JsonParser]
 
 
 class Spec(BaseModel):
@@ -1614,7 +1614,7 @@ class CompositeErrorHandler(BaseModel):
 
 class CompositeRawDecoder(BaseModel):
     type: Literal["CompositeRawDecoder"]
-    parser: Union[GzipParser, JsonLineParser, CsvParser]
+    parser: Union[GzipParser, JsonParser, JsonLineParser, CsvParser]
 
 
 class DeclarativeSource1(BaseModel):

{airbyte_cdk-6.18.0.dev0 → airbyte_cdk-6.18.0.dev2}/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 = "6.18.0dev0"  # Version will be calculated dynamically.
+version = "6.18.0dev2"  # Version will be calculated dynamically.
 
 [tool.poetry-dynamic-versioning]
 enable = false