Remove references to plotly-orca and scripts making use of it.

2023-12-14 14:26:54 +00:00 · 2023-12-14 14:26:54 +00:00 · 4774200e4b
--- a/README.md
+++ b/README.md
@ -52,7 +52,7 @@ The following script summarizes the dataset statistics:
 python3 scripts/summarize_dataset.py --data_path path/to/save/dataset/orbit_benchmark_<FRAME_SIZE> 
 # to aggregate stats across train, validation, and test collectors, add --combine_modes
 ```
-These should match the values in Table 2 (`combine_modes=True`) and Table A.2 (`combine_modes=False`) in the dataset paper. The Jupyter notebook `scripts/plot_dataset.ipynb` can be used to plot bar charts summarizing the dataset (uses Plotly). These should match Figure 2 (`combine_modes=True`) and Figure A.3/A.4 (`combine_modes=False`) in the dataset paper.
+These should match the values in Table 2 (`combine_modes=True`) and Table A.2 (`combine_modes=False`) in the dataset paper.

 # Training & testing models on ORBIT

@ -245,7 +245,7 @@ bash scripts/download_unfiltered_dataset.sh folder/to/save/dataset
 ```
 Alternatively, the train/validation/test/other ZIPs can be manually downloaded [here](https://city.figshare.com/articles/dataset/_/14294597). Use `scripts/merge_and_split_benchmark_users.py` to merge the other folder (see script for usage details).

-To summarize and plot the unfiltered dataset, use `scripts/summarize_dataset.py` (with `--no_modes` rather than `--combine_modes`) and `scripts/plot_dataset.ipynb` (with `no_modes=True`) similar to above.
+To summarize and plot the unfiltered dataset, use `scripts/summarize_dataset.py` (with `--no_modes` rather than `--combine_modes`) similar to above.

 # Citations

--- a/environment.yml
+++ b/environment.yml
@ -7,7 +7,6 @@ dependencies:
  - python=3.7
  - pip=22.2.2
  - plotly=4.8.1
-  - plotly-orca=1.3.1
  - pytorch=1.13.1
  - pytorch-cuda=11.6
  - torchvision=0.14.1
--- a/scripts/plot_dataset.ipynb
+++ b/scripts/plot_dataset.ipynb
@ -1,198 +0,0 @@
-{
- "cells": [
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "# Copyright (c) Microsoft Corporation.\n",
-    "# Licensed under the MIT license.\n",
-    "\n",
-    "import os\n",
-    "import plotly\n",
-    "import plotly.graph_objects as go\n",
-    "\n",
-    "# produce plots for ORBIT benchmark dataset, combining train/validation/test users\n",
-    "DATASET_DIR = \"/path/to/benchmark/root\"\n",
-    "PHASE = \"benchmark\"\n",
-    "COMBINE_MODES = True\n",
-    "PLOT_FOLDER = \"../plots\"\n",
-    "\n",
-    "# produce plots for just <train,validation,test> users in ORBIT benchmark dataset\n",
-    "# DATASET_DIR = os.path.join(\"/path/to/benchmark/root\", \"<train,validation,test>\")\n",
-    "# PHASE = \"benchmark_<train,validation,test>\"\n",
-    "# COMBINE_MODES = False\n",
-    "# PLOT_FOLDER = \"../plots\"\n",
-    "\n",
-    "# produce plots for ORBIT unfiltered dataset, combining all users\n",
-    "# DATASET_DIR = \"/path/to/unfiltered/root\"\n",
-    "# PHASE = \"unfiltered\"\n",
-    "# COMBINE_MODES = False # unfiltered dataset does not have modes\n",
-    "# PLOT_FOLDER = \"../plots\""
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "tags": []
-   },
-   "outputs": [],
-   "source": [
-    "def get_videos_per_object_by_user(path):\n",
-    "    user2objs, user2numvids, user2vids, vid2type = {}, {}, {}, {}\n",
-    "    modes = ['train', 'validation', 'test'] if COMBINE_MODES else ['']\n",
-    "    for mode in modes:\n",
-    "        mode_dir = os.path.join(path, mode)\n",
-    "        for user in os.listdir(mode_dir):\n",
-    "            user_dir = os.path.join(mode_dir, user)\n",
-    "            if os.path.isdir(user_dir):\n",
-    "                user2objs[user] = []\n",
-    "                user2numvids[user] = {'clean': [], 'clutter': [], 'clutter-pan': []}\n",
-    "                user2vids[user] = []\n",
-    "                for obj in os.listdir(user_dir):\n",
-    "                    obj_dir = os.path.join(user_dir, obj)\n",
-    "                    user2objs[user].append(obj)\n",
-    "                    for video_type in ['clean', 'clutter', 'clutter-pan']:\n",
-    "                        type_dir = os.path.join(obj_dir, video_type)\n",
-    "                        if os.path.exists(type_dir):\n",
-    "                            user2numvids[user][video_type].append( len(os.listdir(type_dir)) )\n",
-    "                            for video in os.listdir(type_dir):\n",
-    "                                user2vids[user].append(video)\n",
-    "                                vid2type[video] = video_type\n",
-    "                        else:\n",
-    "                            user2numvids[user][video_type].append(0)\n",
-    "\n",
-    "    return user2objs, user2numvids, user2vids, vid2type\n",
-    "\n",
-    "os.makedirs(PLOT_FOLDER, exist_ok=True)\n",
-    "user2objs, user2numvids, user2vids, vid2type = get_videos_per_object_by_user(DATASET_DIR)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "tags": []
-   },
-   "outputs": [],
-   "source": [
-    "def plot_objects_per_user_histogram(user2objs):\n",
-    "    x, y, user_ids = [],[],[]\n",
-    "    for i, k in enumerate(sorted(user2objs.items(), key=lambda x: len(x[1]), reverse=True)):\n",
-    "        user_id, user_objs = k\n",
-    "        x.append(i)\n",
-    "        y.append(len(user_objs))\n",
-    "        user_ids.append(user_id)\n",
-    "    fig = go.Figure(data=[go.Bar(x=x, y=y)])\n",
-    "    title=\"objects_per_user\"\n",
-    "    layout = go.Layout(\n",
-    "        height=1500,\n",
-    "        width=3000,\n",
-    "        font=dict(size=40),\n",
-    "    \txaxis = go.layout.XAxis(\n",
-    "       \t\ttitle = \"Collector\",\n",
-    "            tickangle = 90,\n",
-    "            ticks='',\n",
-    "            ticktext = user_ids,\n",
-    "            tickvals = x,\n",
-    "            tickfont = dict(size=40)),\n",
-    "        yaxis = go.layout.YAxis(\n",
-    "    \t\ttitle = \"Number of objects\",\n",
-    "            tickfont=dict(size=40))\n",
-    "    )\n",
-    "    fig.update_layout(layout)\n",
-    "    plotly.offline.iplot(fig)\n",
-    "    fig.write_html( os.path.join(PLOT_FOLDER, \"{:}_{:}.html\".format(title, PHASE)) )\n",
-    "    fig.write_image( os.path.join(PLOT_FOLDER, \"{:}_{:}.pdf\".format(title, PHASE)) )\n",
-    "\n",
-    "plot_objects_per_user_histogram(user2objs)\n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "tags": []
-   },
-   "outputs": [],
-   "source": [
-    "def plot_videos_by_object_per_user_histogram(user2objs, user2numvids):\n",
-    "    x, user_ids = [],[]\n",
-    "    MAX_OBJS = 15\n",
-    "    COLORS=[(99, 110, 250, 1), (239, 85, 59, 1), (0, 204, 150, 1), (171, 99, 250, 1), (255, 161, 90, 1), (25, 211, 243, 1), (255, 102, 146, 1), (182, 232, 128, 1), (255, 151, 255, 1), (254, 203, 82, 1), (44, 160, 44, 1), (23, 190, 207, 1), (140, 86, 75, 1), (114, 183, 178, 1), (188, 189, 34, 1), (254, 203, 82, 1) ]\n",
-    "    num_clean_vids_per_obj = [ [] for i in range(MAX_OBJS) ]\n",
-    "    num_cluttered_vids_per_obj = [ [] for i in range(MAX_OBJS) ]\n",
-    "    trace_names = [ \"obj {:}\".format(i+1) for i in range(MAX_OBJS)]\n",
-    "    for i, k in enumerate(sorted(user2objs.items(), key=lambda x: len(x[1]), reverse=True)):\n",
-    "        user_id, _ = k\n",
-    "        x.append(i)\n",
-    "        num_objs = len(user2objs[user_id])\n",
-    "        for j in range(MAX_OBJS):\n",
-    "            if j < num_objs:\n",
-    "                num_clean_vids_per_obj[j].append( user2numvids[user_id]['clean'][j] )\n",
-    "                num_cluttered_vids_per_obj[j].append( user2numvids[user_id]['clutter'][j] )\n",
-    "        user_ids.append(user_id)\n",
-    "    fig = go.Figure()\n",
-    "    for j in range(MAX_OBJS):\n",
-    "        lighter_color = list(COLORS[j])\n",
-    "        lighter_color[3] = 0.65\n",
-    "        fig.add_trace(go.Bar(x=x, y=num_clean_vids_per_obj[j], name=trace_names[j] + \" clean\", marker_color=\"rgba{:}\".format(tuple(lighter_color))))\n",
-    "        fig.add_trace(go.Bar(x=x, y=num_cluttered_vids_per_obj[j], name=trace_names[j] + \" clutter\", marker_color=\"rgba{:}\".format(COLORS[j])))\n",
-    "    title=\"videos_by_object_per_user\"\n",
-    "    layout = go.Layout(\n",
-    "        height=1500,\n",
-    "        width=3000,\n",
-    "        font=dict(size=40),\n",
-    "    \txaxis = go.layout.XAxis(\n",
-    "       \t\ttitle = \"Collector\",\n",
-    "            tickangle = 90,\n",
-    "            ticks='',\n",
-    "            ticktext = user_ids,\n",
-    "            tickvals = x,\n",
-    "            tickfont=dict(size=40)),\n",
-    "        yaxis = go.layout.YAxis(\n",
-    "    \t\ttitle = \"Number of videos by object\",\n",
-    "            tickfont=dict(size=40)),\n",
-    "        barmode = 'stack',\n",
-    "        showlegend=True,\n",
-    "        legend = dict(font=dict(size=28), orientation='h', yanchor=\"bottom\", y=1, xanchor=\"left\", x=0)\n",
-    "    )\n",
-    "    fig.update_layout(layout)\n",
-    "    plotly.offline.iplot(fig)\n",
-    "    fig.write_html( os.path.join(PLOT_FOLDER, \"{:}_{:}.html\".format(title, PHASE)) )\n",
-    "    fig.write_image( os.path.join(PLOT_FOLDER, \"{:}_{:}.pdf\".format(title, PHASE)) )\n",
-    "\n",
-    "plot_videos_by_object_per_user_histogram(user2objs, user2numvids)"
-   ]
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "Python 3.8.13 ('orbit-dataset-v3')",
-   "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": "3.8.13"
-  },
-  "orig_nbformat": 2,
-  "vscode": {
-   "interpreter": {
-    "hash": "5708594860c198df6886bc22825155b8182a7ae4388351ec2cbfd6b7c109203f"
-   }
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 2
-}