fastplotlib · kushalkolar · Mar 5, 2024 · Mar 5, 2024
@@ -0,0 +1,3 @@
+This example shows how to use a zmq publisher-subscriber pattern to perform a computation in one process and visualize results in another process. First, run all cells in `multiprocessing_zmq_plot.ipynb`, and then run cells in `multiprocessing_zmq_compute.ipynb`. The raw bytes for the numpy array are sent using zmq in the compute notebook and received in the plot notebook and displayed.
+
+For more information on zmq see: https://zeromq.org/languages/python/
@@ -0,0 +1,73 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "ca2817a3-869c-4cc6-901b-c34509518175",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import zmq"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dd0b9780-b507-4ea2-af09-134abd76f45b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "context = zmq.Context()\n",
+    "\n",
+    "# create publisher\n",
+    "socket = context.socket(zmq.PUB)\n",
+    "socket.bind(\"tcp://127.0.0.1:5555\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4729bfe8-3474-4bc9-a489-a57d02e5a287",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for i in range(2_000):\n",
+    "    # make some data, make note of the dtype\n",
+    "    data = np.random.rand(512, 512).astype(np.float32)\n",
+    "\n",
+    "    # sent bytes over the socket\n",
+    "    socket.send(data.tobytes())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d47dab72-1061-439f-bf6e-a88b9ee8e5aa",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
@@ -0,0 +1,114 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "491e6050-64ae-4bfc-a480-5805cd684710",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import fastplotlib as fpl\n",
+    "import numpy as np\n",
+    "import zmq"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "97135f98-6810-49b6-a8de-d0e114720d6c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "context = zmq.Context()\n",
+    "\n",
+    "# create subscriber\n",
+    "sub = context.socket(zmq.SUB)\n",
+    "sub.setsockopt(zmq.SUBSCRIBE, b\"\")\n",
+    "\n",
+    "# keep only the most recent message\n",
+    "sub.setsockopt(zmq.CONFLATE, 1)\n",
+    "\n",
+    "# publisher address and port\n",
+    "sub.connect(\"tcp://127.0.0.1:5555\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2d4420f2-364a-445a-9658-63e9ffa586c3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def get_bytes():\n",
+    "    \"\"\"\n",
+    "    Gets the bytes from the publisher\n",
+    "    \"\"\"\n",
+    "    try:\n",
+    "        b = sub.recv(zmq.NOBLOCK)\n",
+    "    except zmq.Again:\n",
+    "        pass\n",
+    "    else:\n",
+    "        return b\n",
+    "    \n",
+    "    return None"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "42d20d77-b884-4379-80e4-e08738506eeb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plot = fpl.Plot()\n",
+    "\n",
+    "# initialize some data, must be of same dtype and shape as data sent by publisher\n",
+    "data = np.random.rand(512, 512).astype(np.float32)\n",
+    "plot.add_image(data, name=\"image\")\n",
+    "\n",
+    "def update_frame(p):\n",
+    "    # recieve bytes\n",
+    "    b = get_bytes()\n",
+    "    \n",
+    "    if b is not None:\n",
+    "        # numpy array from bytes, MUST specify dtype and make sure it matches what you sent\n",
+    "        a = np.frombuffer(b, dtype=np.float32).reshape(512, 512)\n",
+    "        \n",
+    "        # set graphic data\n",
+    "        p[\"image\"].data = a\n",
+    "\n",
+    "plot.add_animations(update_frame)\n",
+    "plot.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "0f8ac188-9359-4d3c-b8f1-384be84d1585",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,3 @@
		This example shows how to use a zmq publisher-subscriber pattern to perform a computation in one process and visualize results in another process. First, run all cells in `multiprocessing_zmq_plot.ipynb`, and then run cells in `multiprocessing_zmq_compute.ipynb`. The raw bytes for the numpy array are sent using zmq in the compute notebook and received in the plot notebook and displayed.

		For more information on zmq see: https://zeromq.org/languages/python/