Granular deps

dimos/stream/video_operators.py

@@ -21,7 +21,6 @@
 import cv2
 import numpy as np
 from reactivex import Observable, Observer, create, operators as ops
-import zmq
 
 if TYPE_CHECKING:
     from dimos.stream.frame_processor import FrameProcessor
@@ -215,27 +214,6 @@ def with_optical_flow(
             ops.map(frame_processor.visualize_flow),
         )
 
-    @staticmethod
-    def with_zmq_socket(
-        socket: zmq.Socket,  # type: ignore[type-arg]
-        scheduler: Any | None = None,
-    ) -> Callable[[Observable], Observable]:  # type: ignore[type-arg]
-        def send_frame(frame, socket) -> None:  # type: ignore[no-untyped-def]
-            _, img_encoded = cv2.imencode(".jpg", frame)
-            socket.send(img_encoded.tobytes())
-            # print(f"Frame received: {frame.shape}")
-
-        # Use a default scheduler if none is provided
-        if scheduler is None:
-            from reactivex.scheduler import ThreadPoolScheduler
-
-            scheduler = ThreadPoolScheduler(1)  # Single-threaded pool for isolation
-
-        return lambda source: source.pipe(
-            ops.observe_on(scheduler),  # Ensure this part runs on its own thread
-            ops.do_action(lambda frame: send_frame(frame, socket)),
-        )
-
     @staticmethod
     def encode_image() -> Callable[[Observable], Observable]:  # type: ignore[type-arg]
         """

docker/python/Dockerfile

@@ -49,4 +49,4 @@ COPY . /app/
 
 # Install dependencies with UV (10-100x faster than pip)
 RUN uv pip install --upgrade 'pip>=24' 'setuptools>=70' 'wheel' 'packaging>=24' && \
-    uv pip install '.[cpu,sim,drone]'
+    uv pip install '.[misc,cpu,sim,drone,unitree,web,perception,visualization]'

pyproject.toml

@@ -29,88 +29,75 @@ requires-python = ">=3.10"
 readme = "README.md"
 
 dependencies = [
-    # Core requirements
-    "opencv-python",
-    "numba>=0.60.0",      # Python 3.12 support
-    "llvmlite>=0.42.0",   # Required by numba 0.60+
-    "python-dotenv",
-    "openai",
-    "anthropic>=0.19.0",
-    "cerebras-cloud-sdk",
-    "moondream",
+    # Transport Protocols
+    "dimos-lcm",
+    "PyTurboJPEG==1.8.2",
+
+    # Core
     "numpy>=1.26.4",
-    "rerun-sdk>=0.20.0",
+    "scipy>=1.15.1",
+    "reactivex",
+    "asyncio==3.4.3",
+    "sortedcontainers==2.4.0",
+    "pydantic",
+    "python-dotenv",
+
+    # Multiprocess
+    "dask[complete]==2025.5.1",
+    "plum-dispatch==2.5.7",
+
+    # Logging
+    "structlog>=25.5.0,<26",
     "colorlog==6.9.0",
+
+    # Core Msgs
+    "opencv-python",
+    "open3d",
+
+    # CLI
+    "pydantic-settings>=2.11.0,<3",
+    "textual==3.7.1",
+    "terminaltexteffects==0.12.2",
+    "typer>=0.19.2,<1",
+    "plotext==5.3.2",
+
+    # Used for calculating the occupancy map.
+    "numba>=0.60.0",    # First version supporting Python 3.12
+    "llvmlite>=0.42.0", # Required by numba 0.60+
+]
+
+
+[project.scripts]
+lcmspy = "dimos.utils.cli.lcmspy.run_lcmspy:main"
+foxglove-bridge = "dimos.utils.cli.foxglove_bridge.run_foxglove_bridge:main"
+skillspy = "dimos.utils.cli.skillspy.skillspy:main"
+agentspy = "dimos.utils.cli.agentspy.agentspy:main"
+humancli = "dimos.utils.cli.human.humanclianim:main"
+dimos = "dimos.robot.cli.dimos:main"
+
+[project.optional-dependencies]
+misc = [
+    # Core requirements
+    "cerebras-cloud-sdk",
     "yapf==0.40.2",
     "typeguard",
     "empy==3.3.4",
     "catkin_pkg",
     "lark",
-    "plum-dispatch==2.5.7",
-    "ffmpeg-python",
     "tiktoken>=0.8.0",
-    "Flask>=2.2",
     "python-multipart==0.0.20",
-    "reactivex",
-    "asyncio==3.4.3",
-    "unitree-webrtc-connect-leshy>=2.0.7",
     "tensorzero==2025.7.5",
-    "structlog>=25.5.0,<26",
-
-    # Web Extensions
-    "fastapi>=0.115.6",
-    "sse-starlette>=2.2.1",
-    "uvicorn>=0.34.0",
-
-    # MCP Server
-    "mcp>=1.0.0",
-
-    # Agents
-    "langchain>=1,<2",
-    "langchain-chroma>=1,<2",
-    "langchain-core>=1,<2",
-    "langchain-openai>=1,<2",
-    "langchain-text-splitters>=1,<2",
-    "langchain-huggingface>=1,<2",
-    "langchain-ollama>=1,<2",
-    "bitsandbytes>=0.48.2,<1.0; sys_platform == 'linux'",
-    "ollama>=0.6.0",
-
-    # Class Extraction
-    "pydantic",
 
     # Developer Specific
     "ipykernel",
 
-    # Unitree webrtc streaming
-    "pycryptodome",
-    "sounddevice",
-    "pyaudio",
-    "requests",
-    "wasmtime",
-
-    # Image
-    "PyTurboJPEG==1.8.2",
-
-    # Audio
-    "openai-whisper",
-    "soundfile",
-
-    # Hugging Face
-    "transformers[torch]==4.49.0",
-
     # Vector Embedding
     "sentence_transformers",
 
     # Perception Dependencies
-    "ultralytics>=8.3.70",
-    "filterpy>=1.4.5",
-    "scipy>=1.15.1",
     "scikit-learn",
-    "Pillow",
     "clip",
     "timm>=1.0.15",
-    "lap>=0.5.12",
     "opencv-contrib-python==4.10.0.84",
 
     # embedding models
@@ -120,41 +107,65 @@ dependencies = [
     "tensorboard==2.20.0",
 
     # Mapping
-    "open3d",
     "googlemaps>=4.10.0",
 
     # Inference
     "onnx",
     "einops==0.8.1",
-    # Multiprocess
-    "dask[complete]==2025.5.1",
-
-    # LCM / DimOS utilities
-    "dimos-lcm==0.1.0",
-
-    # CLI
-    "pydantic-settings>=2.11.0,<3",
-    "typer>=0.19.2,<1",
-    "plotext==5.3.2",
 
     # Teleop
     "pygame>=2.6.1",
     # Hardware SDKs
     "xarm-python-sdk>=1.17.0",
+]
 
-    "numba>=0.60.0",    # First version supporting Python 3.12
-    "llvmlite>=0.42.0", # Required by numba 0.59+
+visualization = [
+    "rerun-sdk>=0.20.0",
 ]
 
-[project.scripts]
-lcmspy = "dimos.utils.cli.lcmspy.run_lcmspy:main"
-foxglove-bridge = "dimos.utils.cli.foxglove_bridge.run_foxglove_bridge:main"
-skillspy = "dimos.utils.cli.skillspy.skillspy:main"
-agentspy = "dimos.utils.cli.agentspy.agentspy:main"
-humancli = "dimos.utils.cli.human.humanclianim:main"
-dimos = "dimos.robot.cli.dimos:main"
+agents = [
+    "langchain>=1,<2",
+    "langchain-chroma>=1,<2",
+    "langchain-core>=1,<2",
+    "langchain-openai>=1,<2",
+    "langchain-text-splitters>=1,<2",
+    "langchain-huggingface>=1,<2",
+    "langchain-ollama>=1,<2",
+    "bitsandbytes>=0.48.2,<1.0; sys_platform == 'linux'",
+    "ollama>=0.6.0",
+    "anthropic>=0.19.0",
+
+    # Audio
+    "openai",
+    "openai-whisper",
+    "sounddevice",
+
+    # MCP Server
+    "mcp>=1.0.0",
+]
+
+web = [
+    "fastapi>=0.115.6",
+    "sse-starlette>=2.2.1",
+    "uvicorn>=0.34.0",
+    "ffmpeg-python",
+    "soundfile",
+]
+
+perception = [
+    "ultralytics>=8.3.70",
+    "filterpy>=1.4.5",
+    "Pillow",
+    "lap>=0.5.12",
+    "transformers[torch]==4.49.0",
+    "moondream",
+]
+
+unitree = [
+    "dimos[agents,web,perception,visualization]",
+    "unitree-webrtc-connect-leshy>=2.0.7"
+]
 
-[project.optional-dependencies]
 manipulation = [
     # Contact Graspnet Dependencies
     "h5py>=3.7.0",
@@ -215,7 +226,6 @@ dev = [
     "pytest-env==1.1.5",
     "pytest-timeout==2.4.0",
     "coverage>=7.0",  # Required for numba compatibility (coverage.types)
-    "textual==3.7.1",
     "requests-mock==1.12.1",
     "terminaltexteffects==0.12.2",
     "watchdog>=3.0.0",