diff --git a/.gitignore b/.gitignore
index 15fb17b133b4c707651b8411b4934d2734968688..a003eda10b9827f972a29ecccbe9b67c3cfa7de1 100644
--- a/.gitignore
+++ b/.gitignore
@@ -109,3 +109,4 @@ ENV/
images/test/
+test_save.dat
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index c722f3c334d412eb2796cff2d46e4e01f063e18d..885f2a32267882160943c6fe20f5b199c244de25 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -25,6 +25,7 @@ tests:
- apt update
- apt install -y libgl1-mesa-glx xvfb
- pip install tox
+ - apt install -y graphviz xdg-utils
- xvfb-run -s "-screen 0 800x600x24" tox
build_and_deploy_docs:
diff --git a/CONTRIBUTING.rst b/CONTRIBUTING.rst
index f552f7626cfffca9a4422cb9337c9cdc27d83867..7ae26bcc3d4e0f4a5cbbcfafd055e2c084a42345 100644
--- a/CONTRIBUTING.rst
+++ b/CONTRIBUTING.rst
@@ -79,7 +79,7 @@ Ready to contribute? Here's how to set up `flatland` for local development.
5. When you're done making changes, check that your changes pass flake8 and the
tests, including testing other Python versions with tox::
- $ flake8 flatland tests
+ $ flake8 flatland tests examples
$ python setup.py test or py.test
$ tox
@@ -125,4 +125,4 @@ $ bumpversion patch # possible: major / minor / patch
$ git push
$ git push --tags
-Travis will then deploy to PyPI if tests pass. (To be configured properly by Mohanty)
\ No newline at end of file
+Travis will then deploy to PyPI if tests pass. (To be configured properly by Mohanty)
diff --git a/MANIFEST.in b/MANIFEST.in
index 965b2dda7db7c49f68857dc3aea9af37e30a745e..30bf97cf0de32bad7e630a80e2b977eb85d48911 100644
--- a/MANIFEST.in
+++ b/MANIFEST.in
@@ -3,6 +3,7 @@ include CONTRIBUTING.rst
include HISTORY.rst
include LICENSE
include README.rst
+include requirements_dev.txt
recursive-include tests *
recursive-exclude * __pycache__
diff --git a/Makefile b/Makefile
index 42a8a430671f9464101ba0799bb6ac4e4f35be20..69ad1b42fd51ef9ec9420f5473dc8acef5468572 100644
--- a/Makefile
+++ b/Makefile
@@ -51,7 +51,7 @@ clean-test: ## remove test and coverage artifacts
rm -fr .pytest_cache
lint: ## check style with flake8
- flake8 flatland tests
+ flake8 flatland tests examples
test: ## run tests quickly with the default Python
echo "$$DISPLAY"
@@ -72,6 +72,7 @@ docs: ## generate Sphinx HTML documentation, including API docs
sphinx-apidoc -o docs/ flatland
$(MAKE) -C docs clean
$(MAKE) -C docs html
+ pydeps --no-config --noshow flatland -o docs/_build/html/flatland.svg
$(BROWSER) docs/_build/html/index.html
servedocs: docs ## compile the docs watching for changes
diff --git a/README.rst b/README.rst
index b4d3193c1b4fb55327216164c602d75cc92e3fdb..11b35adc127b2b5b8b22c55fd0f494871aee56cd 100644
--- a/README.rst
+++ b/README.rst
@@ -92,6 +92,10 @@ flatland
========
TODO: explain the interface here
+Module Dependencies
+===================
+.. image:: flatland.svg
+
Authors
--------
@@ -102,6 +106,7 @@ Authors
* Erik Nygren <erik.nygren@sbb.ch>
* Adrian Egli <adrian.egli@sbb.ch>
* Vaibhav Agrawal <theinfamouswayne@gmail.com>
+* Christian Eichenberger <christian.markus.eichenberger@sbb.ch>
<please fill yourself in>
diff --git a/examples/play_model.py b/examples/play_model.py
index bbbcfffdccf4493bc08da013027f4fdc31ca5aa7..34c6aadfeefd44771fd335e2957e1fbd0b2f740f 100644
--- a/examples/play_model.py
+++ b/examples/play_model.py
@@ -1,12 +1,14 @@
-from flatland.envs.rail_env import RailEnv
-from flatland.envs.generators import complex_rail_generator
-from flatland.utils.rendertools import RenderTool
-# from flatland.baselines.dueling_double_dqn import Agent
-from collections import deque
# import torch
import random
-import numpy as np
import time
+# from flatland.baselines.dueling_double_dqn import Agent
+from collections import deque
+
+import numpy as np
+
+from flatland.envs.generators import complex_rail_generator
+from flatland.envs.rail_env import RailEnv
+from flatland.utils.rendertools import RenderTool
class Player(object):
@@ -25,7 +27,7 @@ class Player(object):
self.done_window = deque(maxlen=100)
self.scores = []
self.dones_list = []
- self.action_prob = [0]*4
+ self.action_prob = [0] * 4
# Removing refs to a real agent for now.
# self.agent = Agent(self.state_size, self.action_size, "FC", 0)
@@ -35,7 +37,7 @@ class Player(object):
self.iFrame = 0
self.tStart = time.time()
-
+
# Reset environment
# self.obs = self.env.reset()
self.env.obs_builder.reset()
@@ -70,7 +72,7 @@ class Player(object):
# Environment step - pass the agent actions to the environment,
# retrieve the response - observations, rewards, dones
next_obs, all_rewards, done, _ = self.env.step(self.action_dict)
-
+
for handle in env.get_agent_handles():
norm = max(1, max_lt(next_obs[handle], np.inf))
next_obs[handle] = np.clip(np.array(next_obs[handle]) / norm, -1, 1)
@@ -79,8 +81,8 @@ class Player(object):
if False:
for handle in self.env.get_agent_handles():
self.agent.step(self.obs[handle], self.action_dict[handle],
- all_rewards[handle], next_obs[handle], done[handle],
- train=False)
+ all_rewards[handle], next_obs[handle], done[handle],
+ train=False)
self.score += all_rewards[handle]
self.iFrame += 1
@@ -96,7 +98,7 @@ def max_lt(seq, val):
None is returned if seq was empty or all items in seq were >= val.
"""
- idx = len(seq)-1
+ idx = len(seq) - 1
while idx >= 0:
if seq[idx] < val and seq[idx] >= 0:
return seq[idx]
@@ -135,16 +137,16 @@ def main(render=True, delay=0.0, n_trials=3, n_steps=50, sGL="QT"):
oPlayer.step()
if render:
env_renderer.renderEnv(show=True, frames=True, iEpisode=trials, iStep=step,
- action_dict=oPlayer.action_dict)
+ action_dict=oPlayer.action_dict)
# time.sleep(10)
if delay > 0:
time.sleep(delay)
-
+
def main_old(render=True, delay=0.0):
''' DEPRECATED main which drives agent directly
Please use the new main() which creates a Player object which is also used by the Editor.
- Please fix any bugs in main() and Player rather than here.
+ Please fix any bugs in main() and Player rather than here.
Will delete this one shortly.
'''
@@ -169,7 +171,7 @@ def main_old(render=True, delay=0.0):
done_window = deque(maxlen=100)
scores = []
dones_list = []
- action_prob = [0]*4
+ action_prob = [0] * 4
# Real Agent
# state_size = 105
@@ -183,7 +185,7 @@ def main_old(render=True, delay=0.0):
None is returned if seq was empty or all items in seq were >= val.
"""
- idx = len(seq)-1
+ idx = len(seq) - 1
while idx >= 0:
if seq[idx] < val and seq[idx] >= 0:
return seq[idx]
@@ -196,7 +198,8 @@ def main_old(render=True, delay=0.0):
# Reset environment
obs = env.reset()
- env_renderer.set_new_rail()
+ if render:
+ env_renderer.set_new_rail()
for a in range(env.get_num_agents()):
norm = max(1, max_lt(obs[a], np.inf))
@@ -252,25 +255,25 @@ def main_old(render=True, delay=0.0):
print(('\rTraining {} Agents.\tEpisode {}\tAverage Score: {:.0f}\tDones: {:.2f}%' +
'\tEpsilon: {:.2f} \t Action Probabilities: \t {}').format(
- env.get_num_agents(),
- trials,
- np.mean(scores_window),
- 100 * np.mean(done_window),
- eps, action_prob/np.sum(action_prob)),
+ env.get_num_agents(),
+ trials,
+ np.mean(scores_window),
+ 100 * np.mean(done_window),
+ eps, action_prob / np.sum(action_prob)),
end=" ")
if trials % 100 == 0:
tNow = time.time()
rFps = iFrame / (tNow - tStart)
print(('\rTraining {} Agents.\tEpisode {}\tAverage Score: {:.0f}\tDones: {:.2f}%' +
'\tEpsilon: {:.2f} fps: {:.2f} \t Action Probabilities: \t {}').format(
- env.get_num_agents(),
- trials,
- np.mean(scores_window),
- 100 * np.mean(done_window),
- eps, rFps, action_prob / np.sum(action_prob)))
+ env.get_num_agents(),
+ trials,
+ np.mean(scores_window),
+ 100 * np.mean(done_window),
+ eps, rFps, action_prob / np.sum(action_prob)))
# torch.save(agent.qnetwork_local.state_dict(),
# '../flatland/baselines/Nets/avoid_checkpoint' + str(trials) + '.pth')
- action_prob = [1]*4
+ action_prob = [1] * 4
if __name__ == "__main__":
diff --git a/examples/qt2.py b/examples/qt2.py
index 6074106523c7fbe503f79b6bc7604f055dc27b35..ee3ea0cd123a3e6d0b1fc970eced219ed8503203 100644
--- a/examples/qt2.py
+++ b/examples/qt2.py
@@ -1,9 +1,8 @@
-
-
import sys
+
from PyQt5 import QtSvg
-from PyQt5.QtWidgets import QApplication, QLabel, QMainWindow, QGridLayout, QWidget
from PyQt5.QtCore import Qt, QByteArray
+from PyQt5.QtWidgets import QApplication, QLabel, QMainWindow, QGridLayout, QWidget
from flatland.utils import svg
@@ -75,4 +74,3 @@ window = MainWindow()
window.show()
app.exec_()
-
diff --git a/examples/temporary_example.py b/examples/temporary_example.py
index 1f3504f221d59d0205974bb135c2237364a22e07..862369411056d87d411c3e173bd479e9a7e93e01 100644
--- a/examples/temporary_example.py
+++ b/examples/temporary_example.py
@@ -1,11 +1,9 @@
import random
-import numpy as np
-import matplotlib.pyplot as plt
-from flatland.envs.rail_env import *
-from flatland.envs.generators import *
-from flatland.envs.observations import TreeObsForRailEnv
-from flatland.utils.rendertools import *
+from flatland.envs.generators import random_rail_generator
+from flatland.envs.rail_env import RailEnv
+from flatland.utils.rendertools import RenderTool
+import numpy as np
random.seed(0)
np.random.seed(0)
@@ -94,7 +92,7 @@ env = RailEnv(width=7,
# print(env.obs_builder.distance_map[0, :, :, i])
# Print the observation vector for agent 0
-obs, all_rewards, done, _ = env.step({0:0})
+obs, all_rewards, done, _ = env.step({0: 0})
for i in range(env.get_num_agents()):
env.obs_builder.util_print_obs_subtree(tree=obs[i], num_features_per_node=5)
@@ -113,6 +111,7 @@ for step in range(100):
while i < len(cmds):
if cmds[i] == 'q':
import sys
+
sys.exit()
elif cmds[i] == 's':
obs, all_rewards, done, _ = env.step(action_dict)
@@ -120,9 +119,9 @@ for step in range(100):
print("Rewards: ", all_rewards, " [done=", done, "]")
else:
agent_id = int(cmds[i])
- action = int(cmds[i+1])
+ action = int(cmds[i + 1])
action_dict[agent_id] = action
- i = i+1
+ i = i + 1
i += 1
env_renderer.renderEnv(show=True)
diff --git a/examples/training_navigation.py b/examples/training_navigation.py
index cabb655e3eb2bc2d12d908559b0102b072163052..85f9531b8820139e5559081feee4a93c4e01ac6c 100644
--- a/examples/training_navigation.py
+++ b/examples/training_navigation.py
@@ -1,11 +1,15 @@
-from flatland.envs.rail_env import *
-from flatland.envs.generators import *
-from flatland.envs.observations import TreeObsForRailEnv
-from flatland.utils.rendertools import *
-from flatland.baselines.dueling_double_dqn import Agent
-from collections import deque
-import torch, random
+import random
import time
+from collections import deque
+
+import numpy as np
+import torch
+
+from flatland.baselines.dueling_double_dqn import Agent
+from flatland.envs.generators import complex_rail_generator
+from flatland.envs.rail_env import RailEnv
+from flatland.utils.rendertools import RenderTool
+
random.seed(1)
np.random.seed(1)
@@ -190,25 +194,34 @@ for trials in range(1, n_trials + 1):
dones_list.append((np.mean(done_window)))
print(
- '\rTraining {} Agents.\tEpisode {}\tAverage Score: {:.0f}\tDones: {:.2f}%\tEpsilon: {:.2f} \t Action Probabilities: \t {}'.format(
+ '\rTraining {} Agents.\t' +
+ 'Episode {}\t' +
+ 'Average Score: {:.0f}\t' +
+ 'Dones: {:.2f}%\t' +
+ 'Epsilon: {:.2f} \t ' +
+ 'Action Probabilities: \t ' +
+ '{}'.format(
env.get_num_agents(),
trials,
- np.mean(
- scores_window),
- 100 * np.mean(
- done_window),
+ np.mean(scores_window),
+ 100 * np.mean(done_window),
eps, action_prob / np.sum(action_prob)),
end=" ")
if trials % 100 == 0:
print(
- '\rTraining {} Agents.\tEpisode {}\tAverage Score: {:.0f}\tDones: {:.2f}%\tEpsilon: {:.2f} \t Action Probabilities: \t {}'.format(
+ '\rTraining {} Agents.\t' +
+ 'Episode {}\t' +
+ 'Average Score: {:.0f}\t' +
+ 'Dones: {:.2f}%\t' +
+ 'Epsilon: {:.2f} \t ' +
+ 'Action Probabilities: \t ' +
+ '{}'.format(
env.get_num_agents(),
trials,
- np.mean(
- scores_window),
- 100 * np.mean(
- done_window),
- eps, action_prob / np.sum(action_prob)))
+ np.mean(scores_window),
+ 100 * np.mean(done_window),
+ eps,
+ action_prob / np.sum(action_prob)))
torch.save(agent.qnetwork_local.state_dict(),
'../flatland/baselines/Nets/avoid_checkpoint' + str(trials) + '.pth')
action_prob = [1] * 4
diff --git a/flatland/envs/observations.py b/flatland/envs/observations.py
index 651d83520ee1f492ea71ba6ddf82cfa5f9093964..fa77b58a53ef63767c2690edc782b6699f6c8459 100644
--- a/flatland/envs/observations.py
+++ b/flatland/envs/observations.py
@@ -492,8 +492,11 @@ class GlobalObsForRailEnv(ObservationBuilder):
self.rail_obs = np.zeros((self.env.height, self.env.width, 16))
for i in range(self.rail_obs.shape[0]):
for j in range(self.rail_obs.shape[1]):
- self.rail_obs[i, j] = np.array(
- list(f'{self.env.rail.get_transitions((i, j)):016b}')).astype(int)
+ bitlist = [int(digit) for digit in bin(self.env.rail.get_transitions((i, j)))[2:]]
+ bitlist = [0] * (16 - len(bitlist)) + bitlist
+ self.rail_obs[i, j] = np.array(bitlist)
+ # self.rail_obs[i, j] = np.array(
+ # list(f'{self.env.rail.get_transitions((i, j)):016b}')).astype(int)
# self.targets = np.zeros(self.env.height, self.env.width)
# for target_pos in self.env.agents_target:
diff --git a/make_docs.py b/make_docs.py
index 7ccbdb736b9b53743b58cfd985538705c5e79f08..8cc1124a6fe624fe5afff416450a0a5d30d654ca 100644
--- a/make_docs.py
+++ b/make_docs.py
@@ -25,5 +25,6 @@ os.environ["SPHINXPROJ"] = "flatland"
os.chdir('docs')
subprocess.call(['python', '-msphinx', '-M', 'clean', '.', '_build'])
subprocess.call(['python', '-msphinx', '-M', 'html', '.', '_build'])
+subprocess.call(['python', '-mpydeps', '../flatland', '-o', '_build/html/flatland.svg'])
browser('_build/html/index.html')
diff --git a/requirements_dev.txt b/requirements_dev.txt
index 08fcc9d17a409f4df71b0b67e037dde5ae042f8a..4b288cee88f4231f330c51f70b1cd9c9f9d05389 100644
--- a/requirements_dev.txt
+++ b/requirements_dev.txt
@@ -2,6 +2,7 @@ bumpversion==0.5.3
wheel==0.32.1
watchdog==0.9.0
flake8==3.5.0
+pydeps==1.7.2
tox==3.5.2
coverage==4.5.1
Sphinx==1.8.1
diff --git a/setup.py b/setup.py
index 96afbe85dc99aa47533936d2a001b9d74717ec58..d14070351559b5e59ed677751b70c20891dbc2f4 100644
--- a/setup.py
+++ b/setup.py
@@ -2,7 +2,7 @@
# -*- coding: utf-8 -*-
"""The setup script."""
-
+import os
from setuptools import setup, find_packages
with open('README.rst') as readme_file:
@@ -11,11 +11,19 @@ with open('README.rst') as readme_file:
with open('HISTORY.rst') as history_file:
history = history_file.read()
-requirements = ['Click>=6.0', ]
-
-setup_requirements = ['pytest-runner', ]
-
-test_requirements = ['pytest', ]
+# Gather requirements from requirements_dev.txt
+# TODO : We could potentially split up the test/dev dependencies later
+install_reqs = []
+requirements_path = 'requirements_dev.txt'
+with open(requirements_path, 'r') as f:
+ install_reqs += [
+ s for s in [
+ line.strip(' \n') for line in f
+ ] if not s.startswith('#') and s != ''
+ ]
+requirements = install_reqs
+setup_requirements = install_reqs
+test_requirements = install_reqs
setup(
author="S.P. Mohanty",
diff --git a/tests/test_rendertools.py b/tests/test_rendertools.py
index 7d3ddf63226b55217edd28bf25aed0307377433b..8204a305328df746a772d034f3c763c848cceb93 100644
--- a/tests/test_rendertools.py
+++ b/tests/test_rendertools.py
@@ -4,14 +4,14 @@
Tests for `flatland` package.
"""
-from flatland.envs.rail_env import RailEnv, random_rail_generator
-import numpy as np
import sys
import matplotlib.pyplot as plt
+import numpy as np
import flatland.utils.rendertools as rt
from flatland.envs.observations import TreeObsForRailEnv
+from flatland.envs.rail_env import RailEnv, random_rail_generator
def checkFrozenImage(oRT, sFileImage, resave=False):
diff --git a/tox.ini b/tox.ini
index e97f8ab022c69cf7ff7aa65d782dddafbecfc90f..6dd011aadeb2e7ba802ff692278aa763fb665f10 100644
--- a/tox.ini
+++ b/tox.ini
@@ -8,12 +8,12 @@ python =
[flake8]
max-line-length = 120
-ignore = E121 E126 E123 E128 E133 E226 E241 E242 E704 W291 W293 W391 W503 W504 W505
+ignore = E121 E126 E123 E128 E133 E226 E241 E242 E704 W291 W293 W391 W503 W504 W505
[testenv:flake8]
basepython = python
deps = flake8
-commands = flake8 flatland
+commands = flake8 flatland tests examples
[testenv:docs]
basepython = python
@@ -23,12 +23,15 @@ commands = make docs
[testenv:coverage]
basepython = python
whitelist_externals = make
-commands = make coverage
+commands =
+ pip install -U pip
+ pip install -r requirements_dev.txt
+ make coverage
[testenv]
whitelist_externals = xvfb-run
- sh
- pip
+ sh
+ pip
setenv =
PYTHONPATH = {toxinidir}
deps =
@@ -38,6 +41,7 @@ deps =
; -r{toxinidir}/requirements.txt
commands =
pip install -U pip
+ pip install -r requirements_dev.txt
sh -c 'echo DISPLAY: $DISPLAY'
xvfb-run -a py.test --basetemp={envtmpdir}