modify setup and requirements and add pylintrc support [currently fai…

…ling]

.gitignore

@@ -1,7 +1,14 @@
+# editors
 .idea/
+.vscode/
+
+# Python
 *.egg-info/
 .pytest_cache/
 __pycache__
-data/
 .cache/
+venv*/
+
+# data
+data/
 *.mp4

.pylintrc

@@ -0,0 +1,14 @@
+[BASIC]
+
+#variable-rgx=[a-z0-9_]{1,30}$
+good-names=dx,dy,i,j,k,px,py,r,th,vx,vy,x,y,z
+
+[TYPECHECK]
+
+# List of members which are set dynamically and missed by pylint inference
+# system, and so shouldn't trigger E1101 when accessed. Python regular
+# expressions are accepted.
+generated-members=numpy.*,torch.*
+
+
+disable=missing-docstring,useless-object-inheritance

README.md

@@ -1,6 +1,6 @@
 # CrowdNav
 This repository contains the codes for our ICRA 2018 submission. For more details, please refer to the paper
-[Crowd-Robot Interaction: Crowd-aware Robot Navigation with Attention-based Deep Reinforcement Learning](https://arxiv.org/abs/1809.08835). 
+[Crowd-Robot Interaction: Crowd-aware Robot Navigation with Attention-based Deep Reinforcement Learning](https://arxiv.org/abs/1809.08835).
 
 If you find the codes or paper useful for your research, please cite our paper:
 ```
@@ -13,15 +13,15 @@ Eprint = {arXiv:1809.08835},
 ```
 
 ## Abstract
-Mobility in an effective and socially-compliant manner is an essential yet challenging task for robots operating in crowded spaces. 
-Recent works have shown the power of deep reinforcement learning techniques to learn socially cooperative policies. 
-However, their cooperation ability deteriorates as the crowd grows since they typically relax the problem as a one-way Human-Robot interaction problem. 
-In this work, we want to go beyond first-order Human-Robot interaction and more explicitly model Crowd-Robot Interaction (CRI). 
-We propose to (i) rethink pairwise interactions with a self-attention mechanism, and 
-(ii) jointly model Human-Robot as well as Human-Human interactions in the deep reinforcement learning framework. 
-Our model captures the Human-Human interactions occurring in dense crowds that indirectly affects the robot's anticipation capability. 
-Our proposed attentive pooling mechanism learns the collective importance of neighboring humans with respect to their future states. 
-Various experiments demonstrate that our model can anticipate human dynamics and navigate in crowds with time efficiency, 
+Mobility in an effective and socially-compliant manner is an essential yet challenging task for robots operating in crowded spaces.
+Recent works have shown the power of deep reinforcement learning techniques to learn socially cooperative policies.
+However, their cooperation ability deteriorates as the crowd grows since they typically relax the problem as a one-way Human-Robot interaction problem.
+In this work, we want to go beyond first-order Human-Robot interaction and more explicitly model Crowd-Robot Interaction (CRI).
+We propose to (i) rethink pairwise interactions with a self-attention mechanism, and
+(ii) jointly model Human-Robot as well as Human-Human interactions in the deep reinforcement learning framework.
+Our model captures the Human-Human interactions occurring in dense crowds that indirectly affects the robot's anticipation capability.
+Our proposed attentive pooling mechanism learns the collective importance of neighboring humans with respect to their future states.
+Various experiments demonstrate that our model can anticipate human dynamics and navigate in crowds with time efficiency,
 outperforming state-of-the-art methods.
 
 
@@ -34,22 +34,17 @@ outperforming state-of-the-art methods.
 ```
 pip install -e .
 ```
-3. Install all the dependencies listed in requirements.txt
-```
-pip install -r requirements.txt
-```
-
 
 ## Getting started
 This repository are organized in two parts: gym_crowd/ folder contains the simulation environment and
-crowd_nav/ folder contains codes for training and testing the policies. Details of the simulation framework can be found 
-[here](crowd_sim/README.md). Below are the instructions for training and testing policies, and they should be executed 
+crowd_nav/ folder contains codes for training and testing the policies. Details of the simulation framework can be found
+[here](crowd_sim/README.md). Below are the instructions for training and testing policies, and they should be executed
 inside the crowd_nav/ folder.
 
 
 1. Train a policy.
 ```
-python train.py --policy sarl
+python -m crowd_nav.train --policy sarl
 ```
 2. Test policies with 500 test cases.
 ```
@@ -67,7 +62,7 @@ python utils/plot.py data/output/output.log
 ```
 
 ## Simulation Videos
-CADRL             | LSTM-RL  
+CADRL             | LSTM-RL
 :-------------------------:|:-------------------------:
 <img src="https://i.imgur.com/vrWsxPM.gif" width="400" />|<img src="https://i.imgur.com/6gjT0nG.gif" width="400" />
 SARL             |  OM-SARL

crowd_nav/test.py

@@ -5,9 +5,9 @@
 import numpy as np
 import gym
 import os
-from crowd_sim.envs.utils.robot import Robot
 from crowd_nav.utils.explorer import Explorer
 from crowd_nav.policy.policy_factory import policy_factory
+from crowd_sim.envs.utils.robot import Robot
 from crowd_sim.envs.policy.orca import ORCA
 
 

crowd_sim/__init__.py

@@ -3,4 +3,4 @@
 register(
     id='CrowdSim-v0',
     entry_point='crowd_sim.envs:CrowdSim',
-)
+)

crowd_sim/envs/__init__.py

@@ -1 +1 @@
-from crowd_sim.envs.crowd_sim import CrowdSim
+from .crowd_sim import CrowdSim

crowd_sim/envs/crowd_sim.py

@@ -306,11 +306,11 @@ def reset(self, phase='test', test_case=None):
 
         # get current observation
         if self.robot.sensor == 'coordinates':
-            ob = [human.get_observable_state() for human in self.humans]
+            obs = [human.get_observable_state() for human in self.humans]
         elif self.robot.sensor == 'RGB':
             raise NotImplemented
 
-        return ob
+        return obs
 
     def onestep_lookahead(self, action):
         return self.step(action, update=False)
@@ -323,9 +323,9 @@ def step(self, action, update=True):
         human_actions = []
         for human in self.humans:
             # observation for humans is always coordinates
-            ob = [other_human.get_observable_state() for other_human in self.humans if other_human != human]
+            obs = [other_human.get_observable_state() for other_human in self.humans if other_human != human]
             if self.robot.visible:
-                ob += [self.robot.get_observable_state()]
+                obs += [self.robot.get_observable_state()]
             human_actions.append(human.act(ob))
 
         # collision detection
@@ -409,19 +409,19 @@ def step(self, action, update=True):
 
             # compute the observation
             if self.robot.sensor == 'coordinates':
-                ob = [human.get_observable_state() for human in self.humans]
+                obs = [human.get_observable_state() for human in self.humans]
             elif self.robot.sensor == 'RGB':
                 raise NotImplemented
         else:
             if self.robot.sensor == 'coordinates':
-                ob = [human.get_next_observable_state(action) for human, action in zip(self.humans, human_actions)]
+                obs = [human.get_next_observable_state(action) for human, action in zip(self.humans, human_actions)]
             elif self.robot.sensor == 'RGB':
                 raise NotImplemented
 
-        return ob, reward, done, info
+        return obs, reward, done, info
 
     def render(self, mode='human', output_file=None):
-        import matplotlib.animation as animation
+        from matplotlib import animation
         import matplotlib.pyplot as plt
         plt.rcParams['animation.ffmpeg_path'] = '/usr/bin/ffmpeg'
 

setup.py

@@ -1,12 +1,32 @@
 from setuptools import setup
 
 
-setup(name='crowd_sim',
-      version='0.0.1',
-      install_requires=['gym', 'numpy', 'matplotlib']
+setup(
+    name='crowdnav',
+    version='0.0.1',
+    packages=[
+        'crowd_nav',
+        'crowd_nav.configs',
+        'crowd_nav.policy',
+        'crowd_nav.utils',
+        'crowd_sim',
+        'crowd_sim.envs',
+        'crowd_sim.envs.policy',
+        'crowd_sim.envs.utils',
+    ],
+    install_requires=[
+        'gitpython',
+        'gym',
+        'matplotlib',
+        'numpy',
+        'scipy',
+        'torch',
+        'torchvision',
+    ],
+    extras_require={
+        'test': [
+            'pylint',
+            'pytest',
+        ],
+    },
 )
-
-setup(name='crowd_nav',
-      version='0.0.1',
-      install_requires=['gym', 'numpy', 'torch', 'matplotlib']
-)