feat(low-code cdk): add dynamic schema loader

airbyte_cdk/sources/declarative/declarative_component_schema.yaml

@@ -1676,6 +1676,92 @@ definitions:
       $parameters:
         type: object
         additionalProperties: true
+  TypesMap:
+    title: Types Map
+    description: (This component is experimental. Use at your own risk.) Represents a mapping between a current type and its corresponding target type.
+    type: object
+    required:
+      - target_type
+      - current_type
+    properties:
+      target_type:
+        anyOf:
+          - type: string
+          - type: array
+            items:
+              type: string
+      current_type:
+        anyOf:
+          - type: string
+          - type: array
+            items:
+              type: string
+  SchemaTypeIdentifier:
+    title: Schema Type Identifier
+    description: (This component is experimental. Use at your own risk.) Identifies schema details for dynamic schema extraction and processing.
+    type: object
+    required:
+      - key_pointer
+    properties:
+      type:
+        type: string
+        enum: [SchemaTypeIdentifier]
+      schema_pointer:
+        title: Schema Path
+        description: List of nested fields defining the schema field path to extract. Defaults to [].
+        type: array
+        default: []
+        items:
+          - type: string
+        interpolation_content:
+          - config
+      key_pointer:
+        title: Key Path
+        description: List of potentially nested fields describing the full path of the field key to extract.
+        type: array
+        items:
+          - type: string
+        interpolation_content:
+          - config
+      type_pointer:
+        title: Type Path
+        description: List of potentially nested fields describing the full path of the field type to extract.
+        type: array
+        items:
+          - type: string
+        interpolation_content:
+          - config
+      types_mapping:
+        type: array
+        items:
+          - "$ref": "#/definitions/TypesMap"
+      $parameters:
+        type: object
+        additionalProperties: true
+  DynamicSchemaLoader:
+    title: Dynamic Schema Loader
+    description: (This component is experimental. Use at your own risk.) Loads a schema by extracting data from retrieved records.
+    type: object
+    required:
+      - type
+      - retriever
+      - schema_type_identifier
+    properties:
+      type:
+        type: string
+        enum: [DynamicSchemaLoader]
+      retriever:
+        title: Retriever
+        description: Component used to coordinate how records are extracted across stream slices and request pages.
+        anyOf:
+          - "$ref": "#/definitions/AsyncRetriever"
+          - "$ref": "#/definitions/CustomRetriever"
+          - "$ref": "#/definitions/SimpleRetriever"
+      schema_type_identifier:
+        "$ref": "#/definitions/SchemaTypeIdentifier"
+      $parameters:
+        type: object
+        additionalProperties: true
   InlineSchemaLoader:
     title: Inline Schema Loader
     description: Loads a schema that is defined directly in the manifest file.
@@ -2057,17 +2143,18 @@ definitions:
               The DeclarativeOAuth Specific URL templated string to obtain the `access_token`, `refresh_token` etc.
               The placeholders are replaced during the processing to provide neccessary values.
             examples:
-              - access_token_url: https://auth.host.com/oauth2/token?{client_id_key}={{client_id_key}}&{client_secret_key}={{client_secret_key}}&{auth_code_key}={{auth_code_key}}&{redirect_uri_key}={urlEncoder:{{redirect_uri_key}}} 
+              - access_token_url: https://auth.host.com/oauth2/token?{client_id_key}={{client_id_key}}&{client_secret_key}={{client_secret_key}}&{auth_code_key}={{auth_code_key}}&{redirect_uri_key}={urlEncoder:{{redirect_uri_key}}}
           access_token_headers:
             title: (Optional) DeclarativeOAuth Access Token Headers
             type: object
             additionalProperties: true
             description: |-
               The DeclarativeOAuth Specific optional headers to inject while exchanging the `auth_code` to `access_token` during `completeOAuthFlow` step.
             examples:
-              - access_token_headers: {
-                  "Authorization": "Basic {base64Encoder:{client_id}:{client_secret}}"
-                }
+              - access_token_headers:
+                  {
+                    "Authorization": "Basic {base64Encoder:{client_id}:{client_secret}}",
+                  }
           access_token_params:
             title: (Optional) DeclarativeOAuth Access Token Query Params (Json Encoded)
             type: object
@@ -2076,18 +2163,19 @@ definitions:
               The DeclarativeOAuth Specific optional query parameters to inject while exchanging the `auth_code` to `access_token` during `completeOAuthFlow` step.
               When this property is provided, the query params will be encoded as `Json` and included in the outgoing API request.
             examples:
-              - access_token_params: {
-                  "{auth_code_key}": "{{auth_code_key}}",
-                  "{client_id_key}": "{{client_id_key}}",
-                  "{client_secret_key}": "{{client_secret_key}}"
-                }
+              - access_token_params:
+                  {
+                    "{auth_code_key}": "{{auth_code_key}}",
+                    "{client_id_key}": "{{client_id_key}}",
+                    "{client_secret_key}": "{{client_secret_key}}",
+                  }
           extract_output:
             title: DeclarativeOAuth Extract Output
             type: array
             items:
               type: string
             description: |-
-              The DeclarativeOAuth Specific list of strings to indicate which keys should be extracted and returned back to the input config.                
+              The DeclarativeOAuth Specific list of strings to indicate which keys should be extracted and returned back to the input config.
             examples:
               - extract_output: ["access_token", "refresh_token", "other_field"]
           state:
@@ -2099,17 +2187,14 @@ definitions:
               - max
             description: |-
               The DeclarativeOAuth Specific object to provide the criteria of how the `state` query param should be constructed,
-              including length and complexity.                
+              including length and complexity.
             properties:
               min:
                 type: integer
               max:
                 type: integer
             examples:
-              - state: {
-                  "min": 7,
-                  "max": 128,
-                }
+              - state: { "min": 7, "max": 128 }
           client_id_key:
             title: (Optional) DeclarativeOAuth Client ID Key Override
             type: string
@@ -2135,14 +2220,14 @@ definitions:
             title: (Optional) DeclarativeOAuth State Key Override
             type: string
             description: |-
-              The DeclarativeOAuth Specific optional override to provide the custom `state` key name, if required by data-provider.                
+              The DeclarativeOAuth Specific optional override to provide the custom `state` key name, if required by data-provider.
             examples:
               - state_key: "my_custom_state_key_key_name"
           auth_code_key:
             title: (Optional) DeclarativeOAuth Auth Code Key Override
             type: string
             description: |-
-              The DeclarativeOAuth Specific optional override to provide the custom `code` key name to something like `auth_code` or `custom_auth_code`, if required by data-provider.                
+              The DeclarativeOAuth Specific optional override to provide the custom `code` key name to something like `auth_code` or `custom_auth_code`, if required by data-provider.
             examples:
               - auth_code_key: "my_custom_auth_code_key_name"
           redirect_uri_key:

airbyte_cdk/sources/declarative/models/declarative_component_schema.py

@@ -650,6 +650,32 @@ class HttpResponseFilter(BaseModel):
     parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
 
 
+class TypesMap(BaseModel):
+    target_type: Union[str, List[str]]
+    current_type: Union[str, List[str]]
+
+
+class SchemaTypeIdentifier(BaseModel):
+    type: Optional[Literal["SchemaTypeIdentifier"]] = None
+    schema_pointer: Optional[List[str]] = Field(
+        [],
+        description="List of nested fields defining the schema field path to extract. Defaults to [].",
+        title="Schema Path",
+    )
+    key_pointer: List[str] = Field(
+        ...,
+        description="List of potentially nested fields describing the full path of the field key to extract.",
+        title="Key Path",
+    )
+    type_pointer: Optional[List[str]] = Field(
+        None,
+        description="List of potentially nested fields describing the full path of the field type to extract.",
+        title="Type Path",
+    )
+    types_mapping: Optional[List[TypesMap]] = None
+    parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
+
+
 class InlineSchemaLoader(BaseModel):
     type: Literal["InlineSchemaLoader"]
     schema_: Optional[Dict[str, Any]] = Field(
@@ -822,13 +848,13 @@ class Config:
     )
     extract_output: List[str] = Field(
         ...,
-        description="The DeclarativeOAuth Specific list of strings to indicate which keys should be extracted and returned back to the input config.                ",
+        description="The DeclarativeOAuth Specific list of strings to indicate which keys should be extracted and returned back to the input config.",
         examples=[{"extract_output": ["access_token", "refresh_token", "other_field"]}],
         title="DeclarativeOAuth Extract Output",
     )
     state: Optional[State] = Field(
         None,
-        description="The DeclarativeOAuth Specific object to provide the criteria of how the `state` query param should be constructed,\nincluding length and complexity.                ",
+        description="The DeclarativeOAuth Specific object to provide the criteria of how the `state` query param should be constructed,\nincluding length and complexity.",
         examples=[{"state": {"min": 7, "max": 128}}],
         title="(Optional) DeclarativeOAuth Configurable State Query Param",
     )
@@ -852,13 +878,13 @@ class Config:
     )
     state_key: Optional[str] = Field(
         None,
-        description="The DeclarativeOAuth Specific optional override to provide the custom `state` key name, if required by data-provider.                ",
+        description="The DeclarativeOAuth Specific optional override to provide the custom `state` key name, if required by data-provider.",
         examples=[{"state_key": "my_custom_state_key_key_name"}],
         title="(Optional) DeclarativeOAuth State Key Override",
     )
     auth_code_key: Optional[str] = Field(
         None,
-        description="The DeclarativeOAuth Specific optional override to provide the custom `code` key name to something like `auth_code` or `custom_auth_code`, if required by data-provider.                ",
+        description="The DeclarativeOAuth Specific optional override to provide the custom `code` key name to something like `auth_code` or `custom_auth_code`, if required by data-provider.",
         examples=[{"auth_code_key": "my_custom_auth_code_key_name"}],
         title="(Optional) DeclarativeOAuth Auth Code Key Override",
     )
@@ -1705,6 +1731,17 @@ class HttpRequester(BaseModel):
     parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
 
 
+class DynamicSchemaLoader(BaseModel):
+    type: Literal["DynamicSchemaLoader"]
+    retriever: Union[AsyncRetriever, CustomRetriever, SimpleRetriever] = Field(
+        ...,
+        description="Component used to coordinate how records are extracted across stream slices and request pages.",
+        title="Retriever",
+    )
+    schema_type_identifier: SchemaTypeIdentifier
+    parameters: Optional[Dict[str, Any]] = Field(None, alias="$parameters")
+
+
 class ParentStreamConfig(BaseModel):
     type: Literal["ParentStreamConfig"]
     parent_key: str = Field(
@@ -1888,5 +1925,6 @@ class SubstreamPartitionRouter(BaseModel):
 SelectiveAuthenticator.update_forward_refs()
 DeclarativeStream.update_forward_refs()
 SessionTokenAuthenticator.update_forward_refs()
+DynamicSchemaLoader.update_forward_refs()
 SimpleRetriever.update_forward_refs()
 AsyncRetriever.update_forward_refs()

airbyte_cdk/sources/declarative/parsers/manifest_component_transformer.py

@@ -58,6 +58,10 @@
     "AddFields.fields": "AddedFieldDefinition",
     # CustomPartitionRouter
     "CustomPartitionRouter.parent_stream_configs": "ParentStreamConfig",
+    # DynamicSchemaLoader
+    "DynamicSchemaLoader.retriever": "SimpleRetriever",
+    # SchemaTypeIdentifier
+    "SchemaTypeIdentifier.types_map": "TypesMap",
 }
 
 # We retain a separate registry for custom components to automatically insert the type if it is missing. This is intended to