nv-ingest 2025.5.21.dev20250521__py3-none-any.whl

nv_ingest/framework/orchestration/ray/util/pipeline/stage_builders.py

@@ -0,0 +1,609 @@
+# SPDX-FileCopyrightText: Copyright (c) 2024, NVIDIA CORPORATION & AFFILIATES.
+# All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+# TODO(Devin)
+# flake8: noqa
+import os
+
+import click
+import logging
+
+from nv_ingest.framework.orchestration.ray.stages.sinks.default_drain import DefaultDrainSink
+from nv_ingest.framework.orchestration.ray.stages.telemetry.otel_tracer import OpenTelemetryTracerStage
+from nv_ingest.framework.schemas.framework_otel_tracer_schema import OpenTelemetryTracerSchema
+from nv_ingest_api.internal.schemas.extract.extract_infographic_schema import InfographicExtractorSchema
+
+# Import our new pipeline class.
+from nv_ingest.framework.orchestration.ray.stages.extractors.audio_extractor import AudioExtractorStage
+from nv_ingest.framework.orchestration.ray.stages.extractors.chart_extractor import ChartExtractorStage
+from nv_ingest.framework.orchestration.ray.stages.extractors.docx_extractor import DocxExtractorStage
+from nv_ingest.framework.orchestration.ray.stages.extractors.image_extractor import ImageExtractorStage
+from nv_ingest.framework.orchestration.ray.stages.extractors.infographic_extractor import InfographicExtractorStage
+from nv_ingest.framework.orchestration.ray.stages.extractors.pdf_extractor import PDFExtractorStage
+from nv_ingest.framework.orchestration.ray.stages.extractors.pptx_extractor import PPTXExtractorStage
+from nv_ingest.framework.orchestration.ray.stages.extractors.table_extractor import TableExtractorStage
+
+from nv_ingest.framework.orchestration.ray.stages.injectors.metadata_injector import MetadataInjectionStage
+from nv_ingest.framework.orchestration.ray.stages.mutate.image_dedup import ImageDedupStage
+from nv_ingest.framework.orchestration.ray.stages.mutate.image_filter import ImageFilterStage
+from nv_ingest.framework.orchestration.ray.stages.sinks.message_broker_task_sink import (
+    MessageBrokerTaskSinkStage,
+    MessageBrokerTaskSinkConfig,
+)
+from nv_ingest.framework.orchestration.ray.stages.sources.message_broker_task_source import (
+    MessageBrokerTaskSourceStage,
+    MessageBrokerTaskSourceConfig,
+    start_simple_message_broker,
+)
+from nv_ingest.framework.orchestration.ray.stages.storage.image_storage import ImageStorageStage
+from nv_ingest.framework.orchestration.ray.stages.storage.store_embeddings import EmbeddingStorageStage
+from nv_ingest.framework.orchestration.ray.stages.transforms.image_caption import ImageCaptionTransformStage
+from nv_ingest.framework.orchestration.ray.stages.transforms.text_embed import TextEmbeddingTransformStage
+from nv_ingest.framework.orchestration.ray.stages.transforms.text_splitter import TextSplitterStage
+from nv_ingest.framework.schemas.framework_metadata_injector_schema import MetadataInjectorSchema
+from nv_ingest_api.internal.schemas.extract.extract_audio_schema import AudioExtractorSchema
+from nv_ingest_api.internal.schemas.extract.extract_chart_schema import ChartExtractorSchema
+from nv_ingest_api.internal.schemas.extract.extract_docx_schema import DocxExtractorSchema
+from nv_ingest_api.internal.schemas.extract.extract_image_schema import ImageConfigSchema
+from nv_ingest_api.internal.schemas.extract.extract_pdf_schema import PDFExtractorSchema
+from nv_ingest_api.internal.schemas.extract.extract_pptx_schema import PPTXExtractorSchema
+from nv_ingest_api.internal.schemas.extract.extract_table_schema import TableExtractorSchema
+from nv_ingest_api.internal.schemas.mutate.mutate_image_dedup_schema import ImageDedupSchema
+from nv_ingest_api.internal.schemas.store.store_embedding_schema import EmbeddingStorageSchema
+from nv_ingest_api.internal.schemas.store.store_image_schema import ImageStorageModuleSchema
+from nv_ingest_api.internal.schemas.transform.transform_image_caption_schema import ImageCaptionExtractionSchema
+from nv_ingest_api.internal.schemas.transform.transform_image_filter_schema import ImageFilterSchema
+from nv_ingest_api.internal.schemas.transform.transform_text_embedding_schema import TextEmbeddingSchema
+from nv_ingest_api.internal.schemas.transform.transform_text_splitter_schema import TextSplitterSchema
+from nv_ingest_api.util.system.hardware_info import SystemResourceProbe
+
+logger = logging.getLogger(__name__)
+
+_system_resource_probe = SystemResourceProbe()
+
+
+def validate_positive(ctx, param, value):
+    if value <= 0:
+        raise click.BadParameter("must be a positive integer")
+    return value
+
+
+def get_message_provider_config():
+    message_provider_host = os.environ.get("MESSAGE_CLIENT_HOST", "localhost")
+    message_provider_port = os.environ.get("MESSAGE_CLIENT_PORT", "6379")
+
+    logger.info(f"MESSAGE_CLIENT_HOST: {message_provider_host}")
+    logger.info(f"MESSAGE_CLIENT_PORT: {message_provider_port}")
+
+    return message_provider_host, message_provider_port
+
+
+def get_caption_classifier_service():
+    triton_service_caption_classifier = os.environ.get(
+        "CAPTION_CLASSIFIER_GRPC_TRITON",
+        "",
+    )
+    triton_service_caption_classifier_name = os.environ.get(
+        "CAPTION_CLASSIFIER_MODEL_NAME",
+        "",
+    )
+
+    logger.info(f"CAPTION_CLASSIFIER_GRPC_TRITON: {triton_service_caption_classifier}")
+
+    return triton_service_caption_classifier, triton_service_caption_classifier_name
+
+
+def get_nim_service(env_var_prefix):
+    prefix = env_var_prefix.upper()
+    grpc_endpoint = os.environ.get(
+        f"{prefix}_GRPC_ENDPOINT",
+        "",
+    )
+    http_endpoint = os.environ.get(
+        f"{prefix}_HTTP_ENDPOINT",
+        "",
+    )
+    auth_token = os.environ.get(
+        "NVIDIA_BUILD_API_KEY",
+        "",
+    ) or os.environ.get(
+        "NGC_API_KEY",
+        "",
+    )
+
+    infer_protocol = os.environ.get(
+        f"{prefix}_INFER_PROTOCOL",
+        "http" if http_endpoint else "grpc" if grpc_endpoint else "",
+    )
+
+    logger.info(f"{prefix}_GRPC_ENDPOINT: {grpc_endpoint}")
+    logger.info(f"{prefix}_HTTP_ENDPOINT: {http_endpoint}")
+    logger.info(f"{prefix}_INFER_PROTOCOL: {infer_protocol}")
+
+    return grpc_endpoint, http_endpoint, auth_token, infer_protocol
+
+
+def get_audio_retrieval_service(env_var_prefix):
+    prefix = env_var_prefix.upper()
+    grpc_endpoint = os.environ.get(
+        "AUDIO_GRPC_ENDPOINT",
+        "",
+    )
+    http_endpoint = os.environ.get(
+        "AUDIO_HTTP_ENDPOINT",
+        "",
+    )
+    auth_token = os.environ.get(
+        "NVIDIA_BUILD_API_KEY",
+        "",
+    ) or os.environ.get(
+        "NGC_API_KEY",
+        "",
+    )
+    infer_protocol = os.environ.get(
+        "AUDIO_INFER_PROTOCOL",
+        "http" if http_endpoint else "grpc" if grpc_endpoint else "",
+    )
+    function_id = os.environ.get(
+        "AUDIO_FUNCTION_ID",
+        "",
+    )
+
+    logger.info(f"{prefix}_GRPC_ENDPOINT: {grpc_endpoint}")
+    logger.info(f"{prefix}_HTTP_ENDPOINT: {http_endpoint}")
+    logger.info(f"{prefix}_INFER_PROTOCOL: {infer_protocol}")
+    logger.info(f"{prefix}_FUNCTION_ID: {function_id}")
+
+    return grpc_endpoint, http_endpoint, auth_token, infer_protocol, function_id
+
+
+def add_metadata_injector_stage(pipeline, default_cpu_count, stage_name="metadata_injector"):
+    _ = default_cpu_count  # Placeholder for future use
+    config = MetadataInjectorSchema()
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=MetadataInjectionStage,
+        config=config,
+        min_replicas=0,
+        max_replicas=1,
+    )
+
+    return stage_name
+
+
+def add_pdf_extractor_stage(pipeline, default_cpu_count, stage_name="pdf_extractor"):
+    yolox_grpc, yolox_http, yolox_auth, yolox_protocol = get_nim_service("yolox")
+    nemoretriever_parse_grpc, nemoretriever_parse_http, nemoretriever_parse_auth, nemoretriever_parse_protocol = (
+        get_nim_service("nemoretriever_parse")
+    )
+    model_name = os.environ.get("NEMORETRIEVER_PARSE_MODEL_NAME", "nvidia/nemoretriever-parse")
+
+    extractor_config = PDFExtractorSchema(
+        **{
+            "pdfium_config": {
+                "auth_token": yolox_auth,  # All auth tokens are the same for the moment
+                "yolox_endpoints": (yolox_grpc, yolox_http),
+                "yolox_infer_protocol": yolox_protocol,
+            },
+            "nemoretriever_parse_config": {
+                "auth_token": nemoretriever_parse_auth,
+                "nemoretriever_parse_endpoints": (nemoretriever_parse_grpc, nemoretriever_parse_http),
+                "nemoretriever_parse_infer_protocol": nemoretriever_parse_protocol,
+                "nemoretriever_parse_model_name": model_name,
+                "yolox_endpoints": (yolox_grpc, yolox_http),
+                "yolox_infer_protocol": yolox_protocol,
+            },
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=PDFExtractorStage,
+        config=extractor_config,
+        min_replicas=0,
+        max_replicas=int(max(1, (default_cpu_count // 3))),  # 33% of available CPU cores
+    )
+
+    return stage_name
+
+
+def add_table_extractor_stage(pipeline, default_cpu_count, stage_name="table_extractor"):
+    yolox_table_structure_grpc, yolox_table_structure_http, yolox_auth, yolox_table_structure_protocol = (
+        get_nim_service("yolox_table_structure")
+    )
+    paddle_grpc, paddle_http, paddle_auth, paddle_protocol = get_nim_service("paddle")
+
+    table_extractor_config = TableExtractorSchema(
+        **{
+            "endpoint_config": {
+                "yolox_endpoints": (yolox_table_structure_grpc, yolox_table_structure_http),
+                "yolox_infer_protocol": yolox_table_structure_protocol,
+                "paddle_endpoints": (paddle_grpc, paddle_http),
+                "paddle_infer_protocol": paddle_protocol,
+                "auth_token": yolox_auth,
+            }
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=TableExtractorStage,
+        config=table_extractor_config,
+        min_replicas=0,
+        max_replicas=int(max(1, (default_cpu_count // 7))),  # 14% of available CPU cores
+    )
+
+    return stage_name
+
+
+def add_chart_extractor_stage(pipeline, default_cpu_count, stage_name="chart_extractor"):
+    yolox_graphic_elements_grpc, yolox_graphic_elements_http, yolox_auth, yolox_graphic_elements_protocol = (
+        get_nim_service("yolox_graphic_elements")
+    )
+    paddle_grpc, paddle_http, paddle_auth, paddle_protocol = get_nim_service("paddle")
+
+    chart_extractor_config = ChartExtractorSchema(
+        **{
+            "endpoint_config": {
+                "yolox_endpoints": (yolox_graphic_elements_grpc, yolox_graphic_elements_http),
+                "yolox_infer_protocol": yolox_graphic_elements_protocol,
+                "paddle_endpoints": (paddle_grpc, paddle_http),
+                "paddle_infer_protocol": paddle_protocol,
+                "auth_token": yolox_auth,
+            }
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=ChartExtractorStage,
+        config=chart_extractor_config,
+        min_replicas=0,
+        max_replicas=int(max(1, (default_cpu_count // 7))),  # 14% of available CPU cores
+    )
+
+    return stage_name
+
+
+def add_infographic_extractor_stage(pipeline, default_cpu_count, stage_name="infographic_extractor"):
+    paddle_grpc, paddle_http, paddle_auth, paddle_protocol = get_nim_service("paddle")
+
+    infographic_content_extractor_config = InfographicExtractorSchema(
+        **{
+            "endpoint_config": {
+                "paddle_endpoints": (paddle_grpc, paddle_http),
+                "paddle_infer_protocol": paddle_protocol,
+                "auth_token": paddle_auth,
+            }
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=InfographicExtractorStage,
+        config=infographic_content_extractor_config,
+        min_replicas=0,
+        max_replicas=int(max(1, (default_cpu_count // 14))),  # 7% of available CPU cores
+    )
+
+    return stage_name
+
+
+def add_image_extractor_stage(pipeline, default_cpu_count, stage_name="image_extractor"):
+    yolox_grpc, yolox_http, yolox_auth, yolox_protocol = get_nim_service("yolox")
+
+    image_extractor_config = ImageConfigSchema(
+        **{
+            "yolox_endpoints": (yolox_grpc, yolox_http),
+            "yolox_infer_protocol": yolox_protocol,
+            "auth_token": yolox_auth,  # All auth tokens are the same for the moment
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=ImageExtractorStage,
+        config=image_extractor_config,
+        min_replicas=0,
+        max_replicas=int(max(1, (default_cpu_count // 14))),  # 7% of available CPU cores
+    )
+
+    return stage_name
+
+
+def add_docx_extractor_stage(pipeline, default_cpu_count, stage_name="docx_extractor"):
+    yolox_grpc, yolox_http, yolox_auth, yolox_protocol = get_nim_service("yolox")
+
+    docx_extractor_config = {
+        "docx_extraction_config": {
+            "yolox_endpoints": (yolox_grpc, yolox_http),
+            "yolox_infer_protocol": yolox_protocol,
+            "auth_token": yolox_auth,
+        }
+    }
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=DocxExtractorStage,
+        config=DocxExtractorSchema(**docx_extractor_config),
+        min_replicas=0,
+        max_replicas=int(max(1, (default_cpu_count // 14))),  # 7% of available CPU cores
+    )
+
+    return stage_name
+
+
+def add_pptx_extractor_stage(pipeline, default_cpu_count, stage_name="pptx_extractor"):
+    yolox_grpc, yolox_http, yolox_auth, yolox_protocol = get_nim_service("yolox")
+
+    pptx_extractor_config = {
+        "pptx_extraction_config": {
+            "yolox_endpoints": (yolox_grpc, yolox_http),
+            "yolox_infer_protocol": yolox_protocol,
+            "auth_token": yolox_auth,
+        }
+    }
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=PPTXExtractorStage,
+        config=PPTXExtractorSchema(**pptx_extractor_config),
+        min_replicas=0,
+        max_replicas=int(max(1, (default_cpu_count // 14))),  # 7% of available CPU cores
+    )
+
+    return stage_name
+
+
+def add_audio_extractor_stage(pipeline, default_cpu_count, stage_name="audio_extractor"):
+    audio_grpc, audio_http, audio_auth, audio_infer_protocol, audio_function_id = get_audio_retrieval_service("audio")
+
+    audio_extractor_config = AudioExtractorSchema(
+        **{
+            "audio_extraction_config": {
+                "audio_endpoints": (audio_grpc, audio_http),
+                "audio_infer_protocol": audio_infer_protocol,
+                "function_id": audio_function_id,
+                "auth_token": audio_auth,
+                # All auth tokens are the same for the moment
+            }
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=AudioExtractorStage,
+        config=audio_extractor_config,
+        min_replicas=0,
+        max_replicas=1,  # Audio extraction is a heavy IO bound operation with minimal CPU usage
+    )
+
+    return stage_name
+
+
+def add_otel_tracer_stage(pipeline, default_cpu_count, stage_name="otel_tracer"):
+    _ = default_cpu_count  # Placeholder for future use
+    otel_endpoint = os.getenv("OTEL_EXPORTER_OTLP_ENDPOINT", "http://localhost:4317")
+
+    otel_tracer_config = OpenTelemetryTracerSchema(
+        **{
+            "otel_endpoint": otel_endpoint,
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=OpenTelemetryTracerStage,
+        config=otel_tracer_config,
+        min_replicas=0,
+        max_replicas=2,
+    )
+
+    return stage_name
+
+
+def add_image_dedup_stage(pipeline, default_cpu_count, stage_name="image_dedup"):
+    config = ImageDedupSchema()
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=ImageDedupStage,
+        config=config,
+        min_replicas=0,
+        max_replicas=1,
+    )
+
+    return stage_name
+
+
+def add_image_filter_stage(pipeline, default_cpu_count, stage_name="image_filter"):
+    config = ImageFilterSchema()
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=ImageFilterStage,
+        config=config,
+        min_replicas=0,
+        max_replicas=1,
+    )
+
+    return stage_name
+
+
+def add_text_splitter_stage(pipeline, default_cpu_count, stage_name="text_splitter"):
+    _ = default_cpu_count
+
+    config = TextSplitterSchema()
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=TextSplitterStage,
+        config=config,
+        min_replicas=0,
+        max_replicas=int(max(1, (default_cpu_count // 14))),  # 7% of available CPU cores
+    )
+
+    return stage_name
+
+
+def add_image_caption_stage(pipeline, default_cpu_count, stage_name="image_caption"):
+    auth_token = os.environ.get(
+        "NVIDIA_BUILD_API_KEY",
+        "",
+    ) or os.environ.get(
+        "NGC_API_KEY",
+        "",
+    )
+
+    endpoint_url = os.environ.get("VLM_CAPTION_ENDPOINT", "localhost:5000")
+    model_name = os.environ.get("VLM_CAPTION_MODEL_NAME", "meta/llama-3.2-11b-vision-instruct")
+
+    config = ImageCaptionExtractionSchema(
+        **{
+            "api_key": auth_token,
+            "endpoint_url": endpoint_url,
+            "image_caption_model_name": model_name,
+            "prompt": "Caption the content of this image:",
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=ImageCaptionTransformStage,
+        config=config,
+        min_replicas=0,
+        max_replicas=1,
+    )
+
+    return stage_name
+
+
+def add_text_embedding_stage(pipeline, default_cpu_count, stage_name="text_embedding"):
+    api_key = os.environ.get(
+        "NVIDIA_BUILD_API_KEY",
+        "",
+    ) or os.environ.get(
+        "NGC_API_KEY",
+        "",
+    )
+    embedding_nim_endpoint = os.getenv("EMBEDDING_NIM_ENDPOINT", "http://embedding:8000/v1")
+    embedding_model = os.getenv("EMBEDDING_NIM_MODEL_NAME", "nvidia/llama-3.2-nv-embedqa-1b-v2")
+
+    config = TextEmbeddingSchema(
+        **{
+            "api_key": api_key,
+            "embedding_nim_endpoint": embedding_nim_endpoint,
+            "embedding_model": embedding_model,
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=TextEmbeddingTransformStage,
+        config=config,
+        min_replicas=0,
+        max_replicas=int(max(1, (default_cpu_count // 14))),  # 7% of available CPU cores
+    )
+
+    return stage_name
+
+
+def add_embedding_storage_stage(pipeline, default_cpu_count, stage_name="embedding_storage"):
+    config = EmbeddingStorageSchema()
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=EmbeddingStorageStage,
+        config=config,
+        min_replicas=0,
+        max_replicas=1,
+    )
+
+    return stage_name
+
+
+def add_image_storage_stage(pipeline, default_cpu_count, stage_name="image_storage"):
+    config = ImageStorageModuleSchema()
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=ImageStorageStage,
+        config=config,
+        min_replicas=0,
+        max_replicas=1,
+    )
+
+    return stage_name
+
+
+def add_default_drain_stage(pipeline, default_cpu_count, stage_name="pipeline_drain"):
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=DefaultDrainSink,
+        config=None,
+        min_replicas=1,
+        max_replicas=1,
+    )
+
+    return stage_name
+
+
+def add_message_broker_response_stage(pipeline, default_cpu_count, stage_name="broker_response"):
+    task_broker_host = os.environ.get("MESSAGE_CLIENT_HOST", "localhost")
+    task_broker_port = os.environ.get("MESSAGE_CLIENT_PORT", "6379")
+    client_type = os.environ.get("MESSAGE_CLIENT_TYPE", "redis")
+
+    sink_config = MessageBrokerTaskSinkConfig(
+        **{
+            "broker_client": {
+                "host": task_broker_host,
+                "port": task_broker_port,
+                "client_type": client_type,
+            },
+        }
+    )
+
+    pipeline.add_stage(
+        name=stage_name,
+        stage_actor=MessageBrokerTaskSinkStage,
+        config=sink_config,
+        min_replicas=0,
+        max_replicas=2,
+    )
+
+    return stage_name
+
+
+def add_source_stage(pipeline, default_cpu_count, source_name="pipeline_source"):
+    _ = default_cpu_count  # Placeholder for future use
+    task_broker_host = os.environ.get("MESSAGE_CLIENT_HOST", "localhost")
+    task_broker_port = os.environ.get("MESSAGE_CLIENT_PORT", "6379")
+
+    client_type = os.environ.get("MESSAGE_CLIENT_TYPE", "redis")
+    task_queue_name = os.environ.get("MESSAGE_CLIENT_QUEUE", "ingest_task_queue")
+
+    source_config = MessageBrokerTaskSourceConfig(
+        **{
+            "broker_client": {
+                "host": task_broker_host,
+                "port": task_broker_port,
+                "client_type": client_type,
+            },
+            "task_queue": task_queue_name,
+            "poll_interval": "0.1",
+        }
+    )
+
+    pipeline.add_source(
+        name=source_name,
+        source_actor=MessageBrokerTaskSourceStage,
+        config=source_config,
+        min_replicas=1,
+        max_replicas=1,
+    )
+
+    if source_config.broker_client.client_type == "simple":
+        start_simple_message_broker(source_config.broker_client.model_dump())
+
+    return source_name

nv_ingest/framework/orchestration/ray/util/system_tools/__init__.py

@@ -0,0 +1,3 @@
+# SPDX-FileCopyrightText: Copyright (c) 2024-25, NVIDIA CORPORATION & AFFILIATES.
+# All rights reserved.
+# SPDX-License-Identifier: Apache-2.0

nv_ingest/framework/orchestration/ray/util/system_tools/memory.py

@@ -0,0 +1,59 @@
+# SPDX-FileCopyrightText: Copyright (c) 2024-25, NVIDIA CORPORATION & AFFILIATES.
+# All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+import time
+import uuid
+import psutil
+import ray
+
+
+def estimate_actor_memory_overhead(
+    actor_class, iterations=1, stabilization_threshold=1 * 1024 * 1024, wait_time=2, actor_args=None, actor_kwargs=None
+):
+    """
+    Estimate the additional system memory overhead when launching a Ray actor of the given actor_class.
+
+    Parameters:
+        actor_class: A Ray remote actor class.
+        iterations (int): Number of measurement iterations.
+        stabilization_threshold (int): Maximum difference (in bytes) between min and max measurements to
+        consider results stable.
+        wait_time (float): Seconds to wait after spawning or killing an actor for memory to stabilize.
+        actor_args (list): Positional arguments to pass to the actor's remote() call.
+        actor_kwargs (dict): Keyword arguments to pass to the actor's remote() call.
+
+    Returns:
+        float: Estimated average overhead in bytes for replicating the actor.
+    """
+    actor_args = actor_args if actor_args is not None else []
+    actor_kwargs = actor_kwargs if actor_kwargs is not None else {}
+
+    measurements = []
+
+    iterations = 0  # TODO
+    for i in range(iterations):
+        # Record baseline system memory usage.
+        baseline = psutil.virtual_memory().used
+
+        # Spin up a new actor with provided arguments.
+        actor = actor_class.options(name=f"mem_estimator_{uuid.uuid4()}").remote(*actor_args, **actor_kwargs)
+        # Allow time for the actor to start.
+        time.sleep(wait_time)
+
+        # Measure memory after actor has started.
+        after_spawn = psutil.virtual_memory().used
+        overhead = after_spawn - baseline
+        measurements.append(overhead)
+
+        # Kill the actor.
+        ray.kill(actor, no_restart=True)
+        # Allow time for system memory to be released.
+        time.sleep(wait_time)
+
+    if measurements:
+        _ = max(measurements) - min(measurements)
+        _ = sum(measurements) / len(measurements)
+
+    return 1_500_000_000
+    # return estimated_overhead Need to come up with a better way to estiamte actor overhead.