MAJOR Refactor opentraj toolkit

LICENSE.txt

@@ -0,0 +1,19 @@
+The MIT License (MIT)
+Copyright (c) 2020-2021 CrowdBot
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in the
+Software without restriction, including without limitation the rights to use,
+copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the
+Software, and to permit persons to whom the Software is furnished to do so,
+subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
+PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

docs/trajnet_graph.py

@@ -0,0 +1,58 @@
+# Author: Javad Amirian
+# Email: [email protected]
+
+import datetime
+import pandas as pd
+import matplotlib.pyplot as plt
+import matplotlib
+from adjustText import adjust_text
+import matplotlib.dates as mdates
+# import seaborn as sns
+# matplotlib.use('TkAgg')
+
+# Read the table
+table_file = '/home/cyrus/Dropbox/OpenTraj-paper/presentation/trajnet-leaders.xls'
+data = pd.read_excel(table_file, header=0)
+ADEs = data["ADE"]
+FDEs = data["FDE"]
+date_objs = []
+for date in data["Date"]:
+    year_month = date.split('-')
+    date_obj = datetime.datetime(year=int(year_month[0]), month=int(year_month[1]), day=1)
+    date_objs.append(date_obj)
+dates = matplotlib.dates.date2num(date_objs)
+
+# Plots
+fig, ax = plt.subplots(nrows=1, ncols=1)
+
+# matplotlib.pyplot.plot_date(dates, FDEs, 'bx', label="FDE")
+# name_texts = [plt.text(dates[i], FDEs[i], pred_name, horizontalalignment='center')
+#               for i, pred_name in enumerate(data["Predictor"])]
+
+matplotlib.pyplot.plot_date(dates, ADEs, 'bo',
+                            label="Average Displacement Error")
+name_texts = [plt.text(dates[i], ADEs[i], pred_name,
+                       color="magenta" if "DCNet" in pred_name else "black")
+              for i, pred_name in enumerate(data["Predictor"])]
+plt.ylim([0.21, 0.53])
+arrows = adjust_text(name_texts, arrowprops=dict(arrowstyle='->', color='red'))
+
+plt.grid(axis='both', linestyle='--')
+plt.ylabel("Prediction Error", fontweight='bold')
+
+
+plt.xlim([min(dates)-500, max(dates) + 300])
+plt.xticks([datetime.datetime(1995, 1, 1),
+            datetime.datetime(2016, 1, 1),
+            datetime.datetime(2018, 1, 1),
+            datetime.datetime(2019, 1, 1),
+            datetime.datetime(2020, 1, 1),
+            datetime.datetime(2021, 1, 1)], rotation=45)
+myFmt = mdates.DateFormatter('%Y')
+ax.xaxis.set_major_formatter(myFmt)
+plt.xlabel("Origin of the Idea", labelpad=-15, fontweight='bold')
+
+ax.set_facecolor((0.90, 0.97, 0.82))
+plt.legend(loc="lower left")
+plt.title("Trajnet Benchmark (World Plane Human-Human Dataset)")
+plt.show()

toolkit/benchmarking/indicators/global_multimodality.py → opentraj/toolkit/indicators/global_multimodality.py

@@ -443,7 +443,7 @@ def run(trajlets, opentraj_root, output_dir):
                                                  'multimodality indicators.')
     parser.add_argument('--opentraj_root', '--root')
     parser.add_argument('--output_dir', '--output')
-    parser.add_argument('--execution', '--exec',
+    parser.add_argument('--execution', '--exe',
                         default='normal',
                         choices=['normal', 'parallelized'],
                         help='pick a execution (default: "vae")')

toolkit/loaders/loader_crowds.py → opentraj/toolkit/loaders/loader_crowds.py

@@ -138,8 +138,13 @@ def load_crowds(path, **kwargs):
     zara_01_vsp = os.path.join(OPENTRAJ_ROOT, 'datasets/UCY/zara01/annotation.vsp')
     zara_hmg_file = os.path.join(OPENTRAJ_ROOT, 'datasets/UCY/zara01/H.txt')
     zara_01_ds = load_crowds(zara_01_vsp, use_kalman=False, homog_file=zara_hmg_file)
-    trajs = zara_01_ds.get_trajectories()
-    trajs = [g for _, g in trajs]
+    # trajs = zara_01_ds.get_trajectories()
+    # trajs = [g for _, g in trajs]
+    trajs = zara_01_ds.data.groupby(["scene_id", "agent_id"])
+    trajs = [(scene_id, agent_id, tr) for (scene_id, agent_id), tr in trajs]
+    for scene_id, agent_id, traj in trajs:
+        cur_locs.append(traj[traj["timestamp"] <= cur_time].iloc)
+
     samples = zara_01_ds.get_entries()
     plt.scatter(samples["pos_x"], samples["pos_y"])
     plt.show()

toolkit/benchmarking/load_all_datasets.py → opentraj/toolkit/test/load_all_datasets.py

@@ -96,7 +96,7 @@ def get_trajlets(opentraj_root, dataset_names, to_numpy=True):
 # ================================================================
 # ===================== Load the datasets ========================
 # ================================================================
-def get_datasets(opentraj_root, dataset_names):
+def get_datasets(opentraj_root, dataset_names, use_kalman='default'):
     datasets = {}
 
     # Make a temp dir to store and load trajdatasets (no postprocess anymore)

opentraj/toolkit/ui/ui_dash.py

@@ -0,0 +1,115 @@
+# Author: Javad Amirian
+# Email: [email protected]
+
+# Run this app with `python app.py` and
+import os
+import numpy as np
+import pandas as pd
+
+import dash
+import dash_core_components as dcc
+import dash_html_components as html
+import dash_bootstrap_components as dbc
+import plotly.express as px
+import plotly.graph_objects as go
+from dash.dependencies import Input, Output
+
+from toolkit.loaders.loader_crowds import load_crowds
+from toolkit.benchmarking.load_all_datasets import all_dataset_names, get_datasets, get_trajlets
+
+# Application
+# ===========================================
+# see https://plotly.com/python/px-arguments/ for more options
+external_stylesheets = ['https://codepen.io/chriddyp/pen/bWLwgP.css']
+app = dash.Dash(__name__, external_stylesheets=external_stylesheets)
+# ----------------------
+
+
+# default values
+# ----------------------
+traj_dataset = pd.DataFrame({})
+cur_time = 0
+time_horizon = [-4.8, 4.8]
+traj_fig = go.Figure()
+graph_height = 800
+time_horizon_marks = [0.4, 0.8, 1.6, 3.2, 4.8, 9.6]
+time_horizon_marks = [-x for x in time_horizon_marks[::-1]] + [0] + time_horizon_marks
+cur_time_slider = dcc.Slider(id='cur_time', min=0, max=1000, step=0.1, value=0, updatemode='drag')
+
+# Biuld the layout
+# dbc.Spinner()
+app.layout = html.Div(children=[
+    html.H2('OpenTraj'),
+    html.H5('Benchmarking Human Trajectory Prediction Datasets'),
+    html.Hr(),
+    dbc.Container(['OpenTraj Root: ',
+                   dcc.Input(id='opentraj_root', type='text', spellCheck='false',
+                             placeholder='Enter path to the root directory of OpenTraj',
+                             # value='Enter the path to the root directory of OpenTraj',
+                             value='/home/cyrus/workspace2/OpenTraj',
+                             inputMode="url", style={'width': 360})],
+                  fluid=True, id='opentraj_root_container'),
+    dcc.Dropdown(id='select_dataset', placeholder='Select Dataset...',
+                 options=[{'label': ds_name, 'value': ds_name} for ds_name in all_dataset_names], multi=False),
+    dcc.Graph(id='traj_graph', figure=traj_fig, style={'height': graph_height}),
+    cur_time_slider,
+    dcc.RangeSlider(id='time_horizon', min=time_horizon_marks[0], max=time_horizon_marks[-1], step=0.1,
+                    marks=dict(
+                        zip(time_horizon_marks, ['%s%.1fs' % (["", "+"][x > 0.001], x) for x in time_horizon_marks])),
+                    value=[-4.8, 4.8], updatemode='drag')
+])
+
+
+@app.callback(Output('cur_time', 'max'),
+              Input('opentraj_root', 'value'),
+              Input('select_dataset', 'value'))
+def load_dataset(opentraj_root, dataset_name):
+    if os.path.exists(opentraj_root) \
+            and os.path.exists(os.path.join(opentraj_root, 'datasets')) and dataset_name:
+        global traj_dataset, traj_fig, cur_time
+        cur_time = 0
+        traj_datasets = get_datasets(opentraj_root, [dataset_name])
+        if dataset_name in traj_datasets:
+            traj_dataset = traj_datasets[dataset_name].data
+        return np.max(traj_dataset["timestamp"])
+    return 0
+
+
+@app.callback(Output('traj_graph', 'figure'),
+              Input('cur_time', 'value'),
+              Input('time_horizon', 'value'),
+              Input('opentraj_root', 'value'),
+              Input('select_dataset', 'value')
+              )
+def update_graph(cur_time_, time_horizon_, _, __):
+    global cur_time, time_horizon
+    cur_time, time_horizon = cur_time_, time_horizon_
+    fig = go.Figure()
+    if len(traj_dataset):
+        print("Scatter Trajs...")
+        selected_data = traj_dataset[(traj_dataset["timestamp"] >= cur_time + time_horizon[0])
+                                     & (traj_dataset["timestamp"] <= cur_time + time_horizon[1])]
+        print("len selected data = ", len(selected_data))
+        trajs = selected_data.groupby(["scene_id", "agent_id"])
+        trajs = [(scene_id, agent_id, tr) for (scene_id, agent_id), tr in trajs]
+
+        cur_poss = []
+        for scene_id, agent_id, traj in trajs:
+            fig.add_scatter(x=traj["pos_x"], y=traj["pos_y"], name="%s[%d]" % (scene_id, agent_id),
+                            mode="lines+markers")
+            if len(traj):
+                # print(agent_id)
+                # print(traj[traj["timestamp"] <= cur_time])
+                cur_poss.append(traj[traj["timestamp"] <= cur_time].iloc[-1][["pos_x", "pos_y"]].to_numpy())
+        cur_poss = np.stack(cur_poss)
+        fig.add_scatter(x=cur_poss[:, 0], y=cur_poss[:, 1], mode="markers", showlegend=False)
+        fig['layout']['xaxis'].update(title='', range=[np.min(traj_dataset["pos_x"]), np.max(traj_dataset["pos_x"])],
+                                      dtick=2.5, autorange=False)
+        fig['layout']['yaxis'].update(title='', range=[np.min(traj_dataset["pos_y"]), np.max(traj_dataset["pos_y"])],
+                                      dtick=2.5, autorange=False)
+        fig.update_yaxes(scaleanchor="x", scaleratio=1)
+    return fig
+
+
+if __name__ == '__main__':
+    app.run_server(debug=True, port=8051)

toolkit/ui/play.py → opentraj/toolkit/ui/ui_pyqt.py

@@ -285,4 +285,4 @@ def play(self, traj_dataset, Hinv, media_file):
 
     play = Play(args.gui_mode)
     play.play(traj_dataset, Hinv, media_file)
-    # qtui.app.exec()
+    # qtui.app.exe()

setup.py

@@ -0,0 +1,17 @@
+from setuptools import setup
+
+ setup(
+   name='opentraj',
+   version='0.1.0',
+   author='Javad Amirian',
+   author_email='[email protected]',
+   packages=['package_name', 'package_name.test'],
+   scripts=['bin/script1','bin/script2'],
+   url='https://github.com/crowdbotp/OpenTraj',
+   license='MIT',
+   description='Tools for working with trajectory datasets',
+   long_description=open('README.txt').read(),
+   install_requires=[
+       "pytest",
+   ],
+)