indexify 0.2.23__tar.gz → 0.2.24__tar.gz

{indexify-0.2.23 → indexify-0.2.24}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: indexify
-Version: 0.2.23
+Version: 0.2.24
 Summary: Python Client for Indexify
 Home-page: https://github.com/tensorlakeai/indexify
 License: Apache 2.0
@@ -18,6 +18,7 @@ Requires-Dist: cloudpickle (>=3.1.0,<4.0.0)
 Requires-Dist: docker (>=7.1.0,<8.0.0)
 Requires-Dist: httpx-sse (>=0.4.0,<0.5.0)
 Requires-Dist: httpx[http2] (>=0,<1)
+Requires-Dist: jsonpickle (>=3.3.0,<4.0.0)
 Requires-Dist: msgpack (>=1.1.0,<2.0.0)
 Requires-Dist: nanoid (>=2.0.0,<3.0.0)
 Requires-Dist: pydantic (>=2.9.2,<3.0.0)

{indexify-0.2.23 → indexify-0.2.24}/indexify/executor/agent.py

@@ -18,7 +18,6 @@ from rich.theme import Theme
 from indexify.functions_sdk.data_objects import (
     FunctionWorkerOutput,
     IndexifyData,
-    RouterOutput,
 )
 from indexify.functions_sdk.graph_definition import ComputeGraphMetadata
 from indexify.http_client import IndexifyClient
@@ -315,7 +314,7 @@ class ExtractorAgent:
                             task=async_task.task,
                             task_outcome="failure",
                             outputs=[],
-                            errors=str(async_task.exception()),
+                            stderr=str(async_task.exception()),
                         )
                         self._task_store.complete(outcome=completed_task)
                         continue
@@ -332,7 +331,6 @@ class ExtractorAgent:
                             task_outcome=task_outcome,
                             outputs=outputs.fn_outputs,
                             router_output=outputs.router_output,
-                            errors=outputs.exception,
                             stdout=outputs.stdout,
                             stderr=outputs.stderr,
                             reducer=outputs.reducer,

{indexify-0.2.23 → indexify-0.2.24}/indexify/executor/function_worker.py

@@ -40,7 +40,6 @@ class FunctionOutput(BaseModel):
     router_output: Optional[RouterOutput]
     reducer: bool = False
     success: bool = True
-    exception: Optional[str] = None
     stdout: str = ""
     stderr: str = ""
 
@@ -111,7 +110,6 @@ class FunctionWorker:
             # TODO - bring back running in a separate process
         except Exception as e:
             return FunctionWorkerOutput(
-                exception=str(e),
                 stdout=e.stdout,
                 stderr=e.stderr,
                 reducer=e.is_reducer,
@@ -121,7 +119,6 @@ class FunctionWorker:
         return FunctionWorkerOutput(
             fn_outputs=result.fn_outputs,
             router_output=result.router_output,
-            exception=result.exception,
             stdout=result.stdout,
             stderr=result.stderr,
             reducer=result.reducer,
@@ -152,7 +149,6 @@ def _run_function(
     router_output = None
     fn_output = None
     has_failed = False
-    exception_msg = None
     print(
         f"[bold] function_worker: [/bold] invoking function {fn_name} in graph {graph_name}"
     )
@@ -180,7 +176,6 @@ def _run_function(
                 if router_call_result.traceback_msg is not None:
                     print(router_call_result.traceback_msg, file=sys.stderr)
                     has_failed = True
-                    exception_msg = router_call_result.traceback_msg
             else:
                 fn_call_result: FunctionCallResult = fn.invoke_fn_ser(
                     fn_name, input, init_value
@@ -190,11 +185,9 @@ def _run_function(
                 if fn_call_result.traceback_msg is not None:
                     print(fn_call_result.traceback_msg, file=sys.stderr)
                     has_failed = True
-                    exception_msg = fn_call_result.traceback_msg
-        except Exception as e:
+        except Exception:
             print(traceback.format_exc(), file=sys.stderr)
             has_failed = True
-            exception_msg = str(e)
 
     # WARNING - IF THIS FAILS, WE WILL NOT BE ABLE TO RECOVER
     # ANY LOGS
@@ -202,7 +195,6 @@ def _run_function(
         return FunctionOutput(
             fn_outputs=None,
             router_output=None,
-            exception=exception_msg,
             stdout=stdout_capture.getvalue(),
             stderr=stderr_capture.getvalue(),
             reducer=is_reducer,

{indexify-0.2.23 → indexify-0.2.24}/indexify/executor/task_reporter.py

@@ -37,17 +37,6 @@ class TaskReporter:
                 ("node_outputs", (nanoid.generate(), io.BytesIO(output_bytes)))
             )
 
-        if completed_task.errors:
-            print(
-                f"[bold]task-reporter[/bold] uploading error of size: {len(completed_task.errors)}"
-            )
-            fn_outputs.append(
-                (
-                    "exception_msg",
-                    (nanoid.generate(), io.BytesIO(completed_task.errors.encode())),
-                )
-            )
-
         if completed_task.stdout:
             print(
                 f"[bold]task-reporter[/bold] uploading stdout of size: {len(completed_task.stdout)}"

{indexify-0.2.23 → indexify-0.2.24}/indexify/executor/task_store.py

@@ -14,7 +14,6 @@ class CompletedTask(BaseModel):
     task_outcome: Literal["success", "failure"]
     outputs: Optional[List[IndexifyData]] = None
     router_output: Optional[RouterOutput] = None
-    errors: Optional[str] = None
     stdout: Optional[str] = None
     stderr: Optional[str] = None
     reducer: bool = False

{indexify-0.2.23 → indexify-0.2.24}/indexify/functions_sdk/data_objects.py

@@ -1,4 +1,4 @@
-from typing import Any, Dict, List, Optional, Union
+from typing import Any, Dict, List, Optional, Union, Literal
 
 from pydantic import BaseModel, Json
 
@@ -17,13 +17,12 @@ class RouterOutput(BaseModel):
 class IndexifyData(BaseModel):
     id: Optional[str] = None
     payload: bytes
-    payload_encoding: str = "cloudpickle"
+    encoder: Literal["cloudpickle", "json"] = "cloudpickle"
 
 
 class FunctionWorkerOutput(BaseModel):
     fn_outputs: Optional[List[IndexifyData]]
     router_output: Optional[RouterOutput]
-    exception: Optional[str]
     stdout: Optional[str]
     stderr: Optional[str]
     reducer: bool = False

{indexify-0.2.23 → indexify-0.2.24}/indexify/functions_sdk/graph.py

@@ -160,6 +160,7 @@ class Graph:
             reducer=start_node.accumulate is not None,
             image_name=start_node.image._image_name,
             image_information=start_node.image.to_image_information(),
+            payload_encoder=start_node.encoder
         )
         metadata_edges = self.edges.copy()
         metadata_nodes = {}
@@ -171,7 +172,7 @@ class Graph:
                         description=node.description or "",
                         source_fn=node_name,
                         target_fns=self.routers[node_name],
-                        payload_encoder=node.payload_encoder,
+                        payload_encoder=node.encoder,
                         image_name=node.image._image_name,
                         image_information=node.image.to_image_information(),
                     )
@@ -185,6 +186,7 @@ class Graph:
                         reducer=node.accumulate is not None,
                         image_name=node.image._image_name,
                         image_information=node.image.to_image_information(),
+                        encoder=node.encoder,
                     )
                 )
 
@@ -202,16 +204,16 @@ class Graph:
 
     def run(self, block_until_done: bool = False, **kwargs) -> str:
         start_node = self.nodes[self._start_node]
-        serializer = get_serializer(start_node.payload_encoder)
-        input = IndexifyData(id=generate(), payload=serializer.serialize(kwargs))
+        serializer = get_serializer(start_node.encoder)
+        input = IndexifyData(id=generate(), payload=serializer.serialize(kwargs), encoder=start_node.encoder)
         print(f"[bold] Invoking {self._start_node}[/bold]")
         outputs = defaultdict(list)
         self._accumulator_values[input.id] = {}
         for k, v in self.accumulator_zero_values.items():
             node = self.nodes[k]
-            serializer = get_serializer(node.payload_encoder)
+            serializer = get_serializer(node.encoder)
             self._accumulator_values[input.id] = {
-                k: IndexifyData(payload=serializer.serialize(v))
+                k: IndexifyData(payload=serializer.serialize(v), encoder=node.encoder)
             }
         self._results[input.id] = outputs
         enable_cache = kwargs.get("enable_cache", True)
@@ -289,7 +291,7 @@ class Graph:
             raise ValueError(f"no results found for fn {fn_name} on graph {self.name}")
         fn = self.nodes[fn_name]
         fn_model = self.get_function(fn_name).get_output_model()
-        serializer = get_serializer(fn.payload_encoder)
+        serializer = get_serializer(fn.encoder)
         outputs = []
         for result in results[fn_name]:
             payload_dict = serializer.deserialize(result.payload)

{indexify-0.2.23 → indexify-0.2.24}/indexify/functions_sdk/graph_definition.py

@@ -48,4 +48,4 @@ class ComputeGraphMetadata(BaseModel):
     runtime_information: RuntimeInformation
 
     def get_input_payload_serializer(self):
-        return get_serializer(self.start_node.compute_fn.payload_encoder)
+        return get_serializer(self.start_node.compute_fn.encoder)

{indexify-0.2.23 → indexify-0.2.24}/indexify/functions_sdk/indexify_functions.py

@@ -22,7 +22,7 @@ from typing_extensions import get_type_hints
 
 from .data_objects import IndexifyData
 from .image import DEFAULT_IMAGE_3_10, Image
-from .object_serializer import CloudPickleSerializer, get_serializer
+from .object_serializer import get_serializer
 
 
 class GraphInvocationContext(BaseModel):
@@ -119,7 +119,7 @@ class IndexifyFunction:
     image: Optional[Image] = DEFAULT_IMAGE_3_10
     placement_constraints: List[PlacementConstraints] = []
     accumulate: Optional[Type[Any]] = None
-    payload_encoder: Optional[str] = "cloudpickle"
+    encoder: Optional[str] = "cloudpickle"
 
     def run(self, *args, **kwargs) -> Union[List[Any], Any]:
         pass
@@ -131,7 +131,7 @@ class IndexifyFunction:
 
     @classmethod
     def deserialize_output(cls, output: IndexifyData) -> Any:
-        serializer = get_serializer(cls.payload_encoder)
+        serializer = get_serializer(cls.encoder)
         return serializer.deserialize(output.payload)
 
 
@@ -140,7 +140,7 @@ class IndexifyRouter:
     description: str = ""
     image: Optional[Image] = DEFAULT_IMAGE_3_10
     placement_constraints: List[PlacementConstraints] = []
-    payload_encoder: Optional[str] = "cloudpickle"
+    encoder: Optional[str] = "cloudpickle"
 
     def run(self, *args, **kwargs) -> Optional[List[IndexifyFunction]]:
         pass
@@ -151,7 +151,7 @@ def indexify_router(
     description: Optional[str] = "",
     image: Optional[Image] = DEFAULT_IMAGE_3_10,
     placement_constraints: List[PlacementConstraints] = [],
-    payload_encoder: Optional[str] = "cloudpickle",
+    encoder: Optional[str] = "cloudpickle",
 ):
     def construct(fn):
         args = locals().copy()
@@ -174,7 +174,7 @@ def indexify_router(
                 setattr(IndexifyRo, key, value)
 
         IndexifyRo.image = image
-        IndexifyRo.payload_encoder = payload_encoder
+        IndexifyRo.encoder = encoder
         return IndexifyRo
 
     return construct
@@ -185,7 +185,7 @@ def indexify_function(
     description: Optional[str] = "",
     image: Optional[Image] = DEFAULT_IMAGE_3_10,
     accumulate: Optional[Type[BaseModel]] = None,
-    payload_encoder: Optional[str] = "cloudpickle",
+    encoder: Optional[str] = "cloudpickle",
     placement_constraints: List[PlacementConstraints] = [],
 ):
     def construct(fn):
@@ -207,10 +207,9 @@ def indexify_function(
         for key, value in args.items():
             if key != "fn" and key != "self":
                 setattr(IndexifyFn, key, value)
-
         IndexifyFn.image = image
         IndexifyFn.accumulate = accumulate
-        IndexifyFn.payload_encoder = payload_encoder
+        IndexifyFn.encoder = encoder
         return IndexifyFn
 
     return construct
@@ -303,7 +302,7 @@ class IndexifyFunctionWrapper:
         self, name: str, input: IndexifyData, acc: Optional[Any] = None
     ) -> FunctionCallResult:
         input = self.deserialize_input(name, input)
-        serializer = get_serializer(self.indexify_function.payload_encoder)
+        serializer = get_serializer(self.indexify_function.encoder)
         if acc is not None:
             acc = self.indexify_function.accumulate.model_validate(
                 serializer.deserialize(acc.payload)
@@ -314,7 +313,7 @@ class IndexifyFunctionWrapper:
             )
         outputs, err = self.run_fn(input, acc=acc)
         ser_outputs = [
-            IndexifyData(payload=serializer.serialize(output)) for output in outputs
+            IndexifyData(payload=serializer.serialize(output), encoder=self.indexify_function.encoder) for output in outputs
         ]
         return FunctionCallResult(ser_outputs=ser_outputs, traceback_msg=err)
 
@@ -324,32 +323,10 @@ class IndexifyFunctionWrapper:
         return RouterCallResult(edges=edges, traceback_msg=err)
 
     def deserialize_input(self, compute_fn: str, indexify_data: IndexifyData) -> Any:
-        if self.indexify_function.payload_encoder == "cloudpickle":
-            return CloudPickleSerializer.deserialize(indexify_data.payload)
-        payload = msgpack.unpackb(indexify_data.payload)
-        signature = inspect.signature(self.indexify_function.run)
-        arg_types = {}
-        for name, param in signature.parameters.items():
-            if (
-                param.annotation != inspect.Parameter.empty
-                and param.annotation != getattr(compute_fn, "accumulate", None)
-            ):
-                arg_types[name] = param.annotation
-        if len(arg_types) > 1:
-            raise ValueError(
-                f"Compute function {compute_fn} has multiple arguments, but only one is supported"
-            )
-        elif len(arg_types) == 0:
-            raise ValueError(f"Compute function {compute_fn} has no arguments")
-        arg_name, arg_type = next(iter(arg_types.items()))
-        if arg_type is None:
-            raise ValueError(f"Argument {arg_name} has no type annotation")
-        if is_pydantic_model_from_annotation(arg_type):
-            if len(payload.keys()) == 1 and isinstance(list(payload.values())[0], dict):
-                payload = list(payload.values())[0]
-            return arg_type.model_validate(payload)
-        return payload
-
+        encoder = indexify_data.encoder
+        payload = indexify_data.payload
+        serializer = get_serializer(encoder)
+        return serializer.deserialize(payload)
 
 def get_ctx() -> GraphInvocationContext:
     frame = inspect.currentframe()

{indexify-0.2.23 → indexify-0.2.24}/indexify/functions_sdk/object_serializer.py

@@ -1,5 +1,6 @@
 from typing import Any, List
 
+import jsonpickle
 import cloudpickle
 import msgpack
 from pydantic import BaseModel
@@ -12,8 +13,26 @@ def get_serializer(serializer_type: str) -> Any:
         return CloudPickleSerializer()
     elif serializer_type == "msgpack":
         return MsgPackSerializer()
-    else:
-        raise ValueError(f"Unknown serializer type: {serializer_type}")
+    elif serializer_type == "json":
+        return JsonSerializer()
+    raise ValueError(f"Unknown serializer type: {serializer_type}")
+
+class JsonSerializer:
+    @staticmethod
+    def serialize(data: Any) -> str:
+        return jsonpickle.encode(data)
+
+    @staticmethod
+    def deserialize(data: str) -> Any:
+        return jsonpickle.decode(data)
+
+    @staticmethod
+    def serialize_list(data: List[Any]) -> str:
+        return jsonpickle.encode(data)
+
+    @staticmethod
+    def deserialize_list(data: str) -> List[Any]:
+        return jsonpickle.decode(data)
 
 
 class CloudPickleSerializer:

{indexify-0.2.23 → indexify-0.2.24}/indexify/functions_sdk/pipeline.py

@@ -28,3 +28,6 @@ class Pipeline:
     def run(self, **kwargs):
         invocation_id = self._graph.run(**kwargs)
         return invocation_id
+
+    def output(self, invocation_id: str, function_name: str):
+        return self._graph.output(invocation_id, function_name)

{indexify-0.2.23 → indexify-0.2.24}/indexify/http_client.py

@@ -14,8 +14,7 @@ from indexify.error import ApiException, GraphStillProcessing
 from indexify.functions_sdk.data_objects import IndexifyData
 from indexify.functions_sdk.graph import ComputeGraphMetadata, Graph
 from indexify.functions_sdk.indexify_functions import IndexifyFunction
-from indexify.settings import DEFAULT_SERVICE_URL, DEFAULT_SERVICE_URL_HTTPS
-
+from indexify.settings import DEFAULT_SERVICE_URL
 
 class InvocationEventPayload(BaseModel):
     invocation_id: str
@@ -47,6 +46,7 @@ class IndexifyClient:
         service_url: str = DEFAULT_SERVICE_URL,
         config_path: Optional[str] = None,
         namespace: str = "default",
+        api_key: Optional[str] = None,
         **kwargs,
     ):
         if os.environ.get("INDEXIFY_URL"):
@@ -74,6 +74,10 @@ class IndexifyClient:
         self._timeout = kwargs.get("timeout")
         self._graphs: Dict[str, Graph] = {}
         self._fns: Dict[str, IndexifyFunction] = {}
+        self._api_key = api_key
+        if not self._api_key:
+            print("API key not provided. Trying to fetch from environment TENSORLAKE_API_KEY variable")
+            self._api_key = os.getenv("TENSORLAKE_API_KEY")
 
     def _request(self, method: str, **kwargs) -> httpx.Response:
         try:
@@ -87,10 +91,9 @@ class IndexifyClient:
                 raise ApiException(response.text)
         except httpx.ConnectError:
             message = (
-                f"Make sure the server is running and accesible at {self._service_url}"
+                f"Make sure the server is running and accessible at {self._service_url}"
             )
             ex = ApiException(message=message)
-            print(ex)
             raise ex
         return response
 
@@ -100,7 +103,7 @@ class IndexifyClient:
         cert_path: str,
         key_path: str,
         ca_bundle_path: Optional[str] = None,
-        service_url: str = DEFAULT_SERVICE_URL_HTTPS,
+        service_url: str = DEFAULT_SERVICE_URL,
         *args,
         **kwargs,
     ) -> "IndexifyClient":
@@ -140,17 +143,25 @@ class IndexifyClient:
             verify=verify_option,
         )
         return client
+    
+    def _add_api_key(self, kwargs):
+        if self._api_key:
+            kwargs["headers"] = {"Authorization": f"Bearer {self._api_key}"}
 
     def _get(self, endpoint: str, **kwargs) -> httpx.Response:
+        self._add_api_key(kwargs)
         return self._request("GET", url=f"{self._service_url}/{endpoint}", **kwargs)
 
     def _post(self, endpoint: str, **kwargs) -> httpx.Response:
+        self._add_api_key(kwargs)
         return self._request("POST", url=f"{self._service_url}/{endpoint}", **kwargs)
 
     def _put(self, endpoint: str, **kwargs) -> httpx.Response:
+        self._add_api_key(kwargs)
         return self._request("PUT", url=f"{self._service_url}/{endpoint}", **kwargs)
 
     def _delete(self, endpoint: str, **kwargs) -> httpx.Response:
+        self._add_api_key(kwargs)
         return self._request("DELETE", url=f"{self._service_url}/{endpoint}", **kwargs)
 
     def _close(self):
@@ -259,14 +270,14 @@ class IndexifyClient:
     ) -> str:
         ser_input = cloudpickle.dumps(kwargs)
         params = {"block_until_finish": block_until_done}
+        kwargs = {"headers": {"Content-Type": "application/cbor"}, "data": ser_input, "params":params}
+        self._add_api_key(kwargs)
         with httpx.Client() as client:
             with connect_sse(
                 client,
                 "POST",
                 f"{self.service_url}/namespaces/{self.namespace}/compute_graphs/{graph}/invoke_object",
-                headers={"Content-Type": "application/cbor"},
-                data=ser_input,
-                params=params,
+                **kwargs,
             ) as event_source:
                 if not event_source.response.is_success:
                     resp = event_source.response.read().decode("utf-8")

indexify-0.2.24/indexify/settings.py

@@ -0,0 +1 @@
+DEFAULT_SERVICE_URL = "https://api.tensorlake.ai"

{indexify-0.2.23 → indexify-0.2.24}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "indexify"
-version = "0.2.23"
+version = "0.2.24"
 description = "Python Client for Indexify"
 authors = ["Tensorlake Inc. <support@tensorlake.ai>"]
 license = "Apache 2.0"
@@ -23,6 +23,7 @@ docker = "^7.1.0"
 msgpack= "^1.1.0"
 typer = "^0.12.5"
 httpx-sse = "^0.4.0"
+jsonpickle = "^3.3.0"
 
 [tool.poetry.dev-dependencies]
 black = "^22.3.0"

indexify-0.2.23/indexify/settings.py

@@ -1,2 +0,0 @@
-DEFAULT_SERVICE_URL = "http://localhost:8900"
-DEFAULT_SERVICE_URL_HTTPS = "https://localhost:8900"