add pycroscopy notebook

mpcontribs-portal/notebooks/contribs.materialsproject.org/pycroscopy.ipynb

@@ -0,0 +1,192 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "protective-banana",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from mpcontribs.client import Client, Attachments\n",
+    "import atomai as aoi\n",
+    "import numpy as np\n",
+    "import matplotlib.pyplot as plt\n",
+    "import torch\n",
+    "from atomai.utils import graphx\n",
+    "%matplotlib inline"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "rural-frontier",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "data_dir = \"/Users/patrick/GoogleDriveLBNL/My Drive/MaterialsProject/gitrepos/mpcontribs-data/pycroscopy\"\n",
+    "imgdata_path = f\"{data_dir}/Gr_SiCr.npy\"\n",
+    "imgdata = np.load(imgdata_path)\n",
+    "model_path = f\"{data_dir}/G_MD.tar\"\n",
+    "model = aoi.load_model(model_path)\n",
+    "# model as dict\n",
+    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
+    "model_dict = torch.load(model_path, map_location=device)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "horizontal-armenia",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "figsize = (8, 8)\n",
+    "# plt.figure(figsize=figsize)\n",
+    "# plt.imshow(imgdata, cmap=\"gray\")\n",
+    "# img_path = imgdata_path.replace(\".npy\", \".png\")\n",
+    "# plt.savefig(img_path, bbox_inches='tight')\n",
+    "# # TODO add img_path as attachment"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "downtown-budget",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "nn_out, coords = model.predict(imgdata)\n",
+    "# model.predict(imgdata, resize=(new_height, new_width))\n",
+    "\n",
+    "map_dict = {0: \"C\", 1: \"Si\"}  # classes to chemical elements\n",
+    "px2ang = 0.104 # pixel-to-angstrom conversion\n",
+    "coord = coords[0] # take the first (and the only one) frame\n",
+    "clusters = graphx.find_cycle_clusters(coord, cycles=[5,7], map_dict=map_dict, px2ang=px2ang)\n",
+    "fig, ax = plt.subplots(1, 1, figsize=figsize)\n",
+    "ax.imshow(imgdata, cmap='gray', origin='lower')\n",
+    "\n",
+    "for i, cl in enumerate(clusters):\n",
+    "    ax.scatter(cl[:, 1], cl[:, 0], s=2, color='red')\n",
+    "    xt = int(np.mean(cl[:, 1]))\n",
+    "    yt = int(np.mean(cl[:, 0]))\n",
+    "    ax.annotate(str(i+1), (xt, yt), size=10, color='white')\n",
+    "    \n",
+    "img_path_clusters = imgdata_path.replace(\".npy\", \"_clusters.png\")\n",
+    "plt.savefig(img_path_clusters, bbox_inches='tight')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "premium-intro",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "clusters_mod = []\n",
+    "#adding a column for C atom as class 0\n",
+    "pad_ = 1\n",
+    "for i in range(len(clusters)):\n",
+    "    clusters[i] = np.pad(clusters[i], (0, pad_), 'constant')\n",
+    "    clusters[i] = clusters[i][:-1]\n",
+    "    clusters_mod.append(clusters[i])\n",
+    "    \n",
+    "#we can also save all the defects per image frame\n",
+    "defect_num = 15\n",
+    "coords_def_15 = {0: clusters_mod[defect_num]}\n",
+    "plt.scatter(coords_def_15[0][:,1], coords_def_15[0][:,0])\n",
+    "\n",
+    "img_path_defects = imgdata_path.replace(\".npy\", \"_defects.png\")\n",
+    "plt.savefig(img_path_defects, bbox_inches='tight')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "important-glucose",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# client = Client()\n",
+    "# client.create_project(\n",
+    "#     name=\"pycroscopy\",\n",
+    "#     title=\"PyCroscopy\",\n",
+    "#     authors=\"A. Ghosh, S. Kalinin\",\n",
+    "#     description=\"Scientific Analysis of nanoscience Data\",\n",
+    "#     url=\"https://pycroscopy.github.io/pycroscopy/about.html\"\n",
+    "# )"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "common-class",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "client = Client(project=\"pycroscopy\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "pursuant-facility",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "imgdata_list = list(imgdata.tolist())\n",
+    "model_dict[\"weights\"] = {\n",
+    "    k: v.tolist()\n",
+    "    for k, v in model_dict[\"weights\"].items()\n",
+    "}"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "corporate-reputation",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "contributions = [{\n",
+    "    \"identifier\": \"mp-7576\", # CrSi on MP\n",
+    "    \"data\": {\"clusters\": len(clusters)},\n",
+    "    \"attachments\": Attachments.from_list([\n",
+    "        img_path_clusters, img_path_defects, #imgdata_list, model_dict,\n",
+    "    ])\n",
+    "}]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "referenced-glenn",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "client.delete_contributions()\n",
+    "client.submit_contributions(contributions)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "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": "3.9.1"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}