plt.style.use('ggplot')
import numpy as np
from scipy.stats import ttest_ind
from dataAnalysis import *
from machinePolicy.onlineVIWithObstacle import RunVI
gridSize = 15
noise = 0.067
noiseActionSpace = [(0, -1), (0, 1), (-1, 0), (1, 0), (1, 1), (1, -1),
(-1, -1), (-1, 1)]
gamma = 0.9
goalReward = [30]
actionSpace = [(0, -1), (0, 1), (-1, 0), (1, 0)]
runVI = RunVI(gridSize, actionSpace, noiseActionSpace, noise, gamma,
goalReward)
softmaxBeta = 2.5
softmaxPolicy = SoftmaxPolicy(softmaxBeta)
initPrior = [0.5, 0.5]
intentionInfernce = IntentionInfernce(initPrior, softmaxPolicy, runVI)
inferThreshold = .9
calFirstIntentionFromPosterior = CalFirstIntentionFromPosterior(inferThreshold)
if __name__ == '__main__':
resultsPath = os.path.join(os.path.join(DIRNAME, '..'), 'results')
statsList = []
stdList = []
statData = []
# participants = ['human', 'softmaxBeta0.1', 'softmaxBeta0.5', 'softmaxBeta1', 'softmaxBeta2.5', 'softmaxBeta5']
# participants = ['human', 'noise0.0673_softmaxBeta2.5']
def main():
picturePath = os.path.abspath(os.path.join(os.path.join(os.getcwd(), os.pardir), 'pictures'))
resultsPath = os.path.abspath(os.path.join(os.path.join(os.getcwd(), os.pardir), 'results'))
machinePolicyPath = os.path.abspath(os.path.join(os.path.join(os.getcwd(), os.pardir), 'machinePolicy'))
dataPath = os.path.abspath(os.path.join(os.path.join(os.getcwd(), os.pardir), 'conditionData'))
df = pd.read_csv(os.path.join(dataPath, 'DesignConditionForAvoidCommitmentZone.csv'))
df['intentionedDisToTargetMin'] = df.apply(lambda x: x['minDis'] - x['avoidCommitmentZone'], axis=1)
gridSize = 15
bounds = [0, 0, gridSize - 1, gridSize - 1]
screenWidth = 600
screenHeight = 600
fullScreen = False
renderOn = False
initializeScreen = InitializeScreen(screenWidth, screenHeight, fullScreen)
screen = initializeScreen()
# pg.mouse.set_visible(False)
leaveEdgeSpace = 2
lineWidth = 1
backgroundColor = [205, 255, 204]
lineColor = [0, 0, 0]
targetColor = [255, 50, 50]
playerColor = [50, 50, 255]
targetRadius = 10
playerRadius = 10
textColorTuple = (255, 50, 50)
drawBackground = DrawBackground(screen, gridSize, leaveEdgeSpace, backgroundColor, lineColor, lineWidth, textColorTuple)
drawText = DrawText(screen, drawBackground)
drawNewState = DrawNewState(screen, drawBackground, targetColor, playerColor, targetRadius, playerRadius)
drawImage = DrawImage(screen)
# condition maps
condition = namedtuple('condition', 'name decisionSteps initAgent avoidCommitPoint crossPoint targetDisToCrossPoint fixedObstacles')
map1ObsStep0a = condition(name='expCondition1', decisionSteps=0, initAgent=(0, 2), avoidCommitPoint=(1, 2), crossPoint=(3, 3), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(1, 1), (1, 3), (3, 1)])
map2ObsStep0a = condition(name='expCondition2', decisionSteps=0, initAgent=(0, 2), avoidCommitPoint=(1, 2), crossPoint=(4, 4), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(1, 1), (1, 3), (3, 1), (1, 4), (4, 1)])
map1ObsStep0b = condition(name='expCondition1', decisionSteps=0, initAgent=(2, 0), avoidCommitPoint=(2, 1), crossPoint=(3, 3), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(1, 1), (1, 3), (3, 1)])
map2ObsStep0b = condition(name='expCondition2', decisionSteps=0, initAgent=(2, 0), avoidCommitPoint=(2, 1), crossPoint=(4, 4), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(1, 1), (1, 3), (3, 1), (1, 4), (4, 1)])
map1ObsStep1a = condition(name='expCondition1', decisionSteps=1, initAgent=(1, 0), avoidCommitPoint=(2, 1), crossPoint=(3, 3), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(1, 1), (1, 3), (3, 1)])
map2ObsStep1a = condition(name='expCondition2', decisionSteps=1, initAgent=(1, 0), avoidCommitPoint=(2, 1), crossPoint=(4, 4), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(1, 1), (1, 3), (3, 1), (1, 4), (4, 1)])
map1ObsStep1b = condition(name='expCondition1', decisionSteps=1, initAgent=(0, 1), avoidCommitPoint=(1, 2), crossPoint=(3, 3), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(1, 1), (1, 3), (3, 1)])
map2ObsStep1b = condition(name='expCondition2', decisionSteps=1, initAgent=(0, 1), avoidCommitPoint=(1, 2), crossPoint=(4, 4), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(1, 1), (1, 3), (3, 1), (1, 4), (4, 1)])
map1ObsStep2 = condition(name='expCondition1', decisionSteps=2, initAgent=(0, 0), avoidCommitPoint=(2, 2), crossPoint=(3, 3), targetDisToCrossPoint=[4, 5, 6], fixedObstacles=[(1, 1), (1, 3), (3, 1)])
map2ObsStep2 = condition(name='expCondition2', decisionSteps=2, initAgent=(0, 0), avoidCommitPoint=(2, 2), crossPoint=(4, 4), targetDisToCrossPoint=[4, 5, 6], fixedObstacles=[(1, 1), (1, 3), (3, 1), (1, 4), (4, 1)])
map1ObsStep4 = condition(name='expCondition1', decisionSteps=4, initAgent=(0, 0), avoidCommitPoint=(3, 3), crossPoint=(4, 4), targetDisToCrossPoint=[4, 5, 6], fixedObstacles=[(2, 2), (2, 0), (2, 4), (0, 2), (4, 2)])
map2ObsStep4 = condition(name='expCondition2', decisionSteps=4, initAgent=(0, 0), avoidCommitPoint=(3, 3), crossPoint=(5, 5), targetDisToCrossPoint=[4, 5, 6], fixedObstacles=[(2, 2), (2, 0), (2, 4), (0, 2), (4, 2), (2, 5), (5, 2)])
map1ObsStep6 = condition(name='expCondition1', decisionSteps=6, initAgent=(0, 0), avoidCommitPoint=(4, 4), crossPoint=(5, 5), targetDisToCrossPoint=[4, 5, 6], fixedObstacles=[(3, 3), (4, 0), (3, 1), (3, 5), (5, 3), (1, 3), (0, 4)])
map2ObsStep6 = condition(name='expCondition2', decisionSteps=6, initAgent=(0, 0), avoidCommitPoint=(4, 4), crossPoint=(6, 6), targetDisToCrossPoint=[4, 5, 6], fixedObstacles=[(3, 3), (4, 0), (3, 1), (3, 5), (5, 3), (1, 3), (0, 4), (3, 6), (6, 3), ])
specialCondition = condition(name='specialCondition', decisionSteps=0, initAgent=(0, 0), avoidCommitPoint=[-1, -1], crossPoint=(5, 5), targetDisToCrossPoint=[5], fixedObstacles=[(3, 0), (0, 3), (1, 4), (1, 6), (4, 1), (6, 1), (6, 2), (2, 6), (5, 3), (3, 5)])
controlCondition1 = condition(name='controlCondition', decisionSteps=0, initAgent=(0, 0), avoidCommitPoint=[-1, -1], crossPoint=(5, 5), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(3, 0), (0, 3), (1, 4), (1, 6), (4, 1), (6, 1), (6, 2), (2, 6), (5, 3), (3, 5)])
controlCondition2 = condition(name='controlCondition', decisionSteps=0, initAgent=(0, 0), avoidCommitPoint=[-1, -1], crossPoint=(5, 5), targetDisToCrossPoint=[5, 6, 7], fixedObstacles=[(3, 0), (0, 3), (1, 4), (5, 3), (4, 1), (3, 5), (2, 2), (6, 1), (6, 2), (1, 6), (2, 6), (5, 3)])
numOfObstacles = 18
controlDiffList = [0, 1, 2]
minSteps = 8
maxsteps = 13
minDistanceBetweenTargets = 4
minDistanceBetweenGrids = max(controlDiffList) + 1
maxDistanceBetweenGrids = calculateMaxDistanceOfGrid(bounds) - minDistanceBetweenGrids
randomWorld = RandomWorld(bounds, minDistanceBetweenGrids, maxDistanceBetweenGrids, numOfObstacles)
randomCondition = namedtuple('condition', 'name creatMap decisionSteps minSteps maxsteps minDistanceBetweenTargets controlDiffList')
randomMaps = randomCondition(name='randomCondition', creatMap=randomWorld, decisionSteps=0, minSteps=minSteps, maxsteps=maxsteps, minDistanceBetweenTargets=minDistanceBetweenTargets, controlDiffList=controlDiffList)
conditionList = [map1ObsStep0a, map1ObsStep0b, map1ObsStep1a, map1ObsStep1b, controlCondition1] + [map1ObsStep2, map1ObsStep4, map1ObsStep6] * 2 + [map2ObsStep0a, map2ObsStep0b, map2ObsStep1a, map2ObsStep1b, controlCondition2] + [map2ObsStep2, map2ObsStep4, map2ObsStep6] * 2 + [randomMaps] * 2
targetDiffsList = [0, 1, 2, 'controlAvoid']
# conditionList = [map1ObsStep0a] + [randomMaps] * 2
n = 1
numBlocks = 3 * n
expDesignValues = [[condition, diff] for condition in conditionList for diff in targetDiffsList] * numBlocks
numExpTrial = len(expDesignValues)
# conditionList = [condition2]
specialDesign = [specialCondition, 0]
numTrialsPerNoiseBlock = 3
noiseCondition = list(permutations([1, 2, 0], numTrialsPerNoiseBlock)) + [(1, 1, 1)]
blockNumber = int(numExpTrial / numTrialsPerNoiseBlock)
noiseDesignValues = createNoiseDesignValue(noiseCondition, blockNumber)
rotatePoint = RotatePoint(gridSize)
isInBoundary = IsInBoundary([0, gridSize - 1], [0, gridSize - 1])
rotateAngles = [0, 90, 180, 270]
creatMap = CreatMap(rotateAngles, gridSize, rotatePoint, numOfObstacles)
checkBoundary = CheckBoundary([0, gridSize - 1], [0, gridSize - 1])
noiseActionSpace = [(0, -1), (0, 1), (-1, 0), (1, 0), (1, 1), (1, -1), (-1, -1), (-1, 1)]
normalNoise = AimActionWithNoise(noiseActionSpace, gridSize)
specialNoise = backToCrossPointNoise
# initPrior = [0.5, 0.5]
# inferGoalPosterior = InferGoalPosterior(goalPolicy)
softmaxBetaList = [2.5]
noise = 0.067
gamma = 0.9
goalReward = [50, 50]
actionSpace = [(0, -1), (0, 1), (-1, 0), (1, 0)]
runVI = RunVI(gridSize, actionSpace, noiseActionSpace, noise, gamma, goalReward)
for softmaxBeta in softmaxBetaList:
# for noise in noiseList:
for i in range(20):
print(i)
expDesignValues = [[condition, diff] for condition in conditionList for diff in targetDiffsList] * numBlocks
random.shuffle(expDesignValues)
expDesignValues.append(specialDesign)
modelController = ModelControllerOnline(softmaxBeta)
# modelController = AvoidCommitModel(softmaxBeta, actionSpace, checkBoundary)
controller = modelController
renderOn = 1
normalTrial = NormalTrial(renderOn, controller, drawNewState, drawText, normalNoise, checkBoundary)
specialTrial = SpecialTrial(renderOn, controller, drawNewState, drawText, specialNoise, checkBoundary)
experimentValues = co.OrderedDict()
experimentValues["name"] = "noise" + str(noise) + '_' + "softmaxBeta" + str(softmaxBeta) + '_' + str(i)
resultsDirPath = os.path.join(resultsPath, "noise" + str(noise) + '_' + "softmaxBeta" + str(softmaxBeta))
if not os.path.exists(resultsDirPath):
os.mkdir(resultsDirPath)
writerPath = os.path.join(resultsDirPath, experimentValues["name"] + '.csv')
writer = WriteDataFrameToCSV(writerPath)
experiment = ObstacleModelSimulation(creatMap, normalTrial, specialTrial, writer, experimentValues, drawImage, resultsPath, runVI)
experiment(noiseDesignValues, expDesignValues)
def main():
picturePath = os.path.abspath(
os.path.join(os.path.join(os.getcwd(), os.pardir), 'pictures'))
resultsPath = os.path.abspath(
os.path.join(os.path.join(os.getcwd(), os.pardir), 'results'))
machinePolicyPath = os.path.abspath(
os.path.join(os.path.join(os.getcwd(), os.pardir), 'machinePolicy'))
dataPath = os.path.abspath(
os.path.join(os.path.join(os.getcwd(), os.pardir), 'conditionData'))
df = pd.read_csv(
os.path.join(dataPath, 'DesignConditionForAvoidCommitmentZone.csv'))
df['intentionedDisToTargetMin'] = df.apply(
lambda x: x['minDis'] - x['avoidCommitmentZone'], axis=1)
gridSize = 15
screenWidth = 600
screenHeight = 600
fullScreen = False
renderOn = False
initializeScreen = InitializeScreen(screenWidth, screenHeight, fullScreen)
screen = initializeScreen()
# pg.mouse.set_visible(False)
leaveEdgeSpace = 2
lineWidth = 1
backgroundColor = [205, 255, 204]
lineColor = [0, 0, 0]
targetColor = [255, 50, 50]
playerColor = [50, 50, 255]
targetRadius = 10
playerRadius = 10
textColorTuple = (255, 50, 50)
drawBackground = DrawBackground(screen, gridSize, leaveEdgeSpace,
backgroundColor, lineColor, lineWidth,
textColorTuple)
drawText = DrawText(screen, drawBackground)
drawNewState = DrawNewState(screen, drawBackground, targetColor,
playerColor, targetRadius, playerRadius)
drawImage = DrawImage(screen)
# condition
width = [5]
height = [5]
intentionDis = [2, 3, 4]
decisionSteps = [0, 1, 2, 4, 6, 8]
# decisionSteps = [1]
targetDiffs = [0, 0, 1, 2]
rotateAngles = [0, 90, 180, 270]
obstaclesMapStep0 = [[(1, 4), (1, 5), (1, 6), (3, 2), (4, 2), (5, 2),
(6, 2), (7, 2)]]
obstaclesMapStep1 = [[(1, 2), (1, 4), (1, 5), (1, 6), (3, 2), (4, 2),
(5, 2), (6, 2)]]
obstaclesMapStep2 = [[(2, 2), (2, 4), (3, 5), (3, 6), (4, 2), (5, 3),
(6, 3)],
[(2, 2), (2, 4), (2, 5), (3, 6), (4, 2), (5, 2),
(6, 3)],
[(2, 2), (2, 4), (3, 5), (2, 6), (4, 2), (5, 3),
(6, 2)]]
obstaclesMapStep4 = [[(3, 3), (4, 1), (1, 4), (5, 3), (3, 5), (6, 3),
(3, 6)],
[(3, 3), (5, 1), (1, 5), (5, 3), (3, 5), (6, 3),
(3, 6)],
[(3, 3), (3, 1), (1, 3), (5, 3), (3, 5), (6, 3),
(3, 6)]]
obstaclesMapStep6 = [[(4, 4), (4, 1), (4, 2), (6, 4), (4, 6), (1, 4),
(2, 4)],
[(4, 4), (5, 1), (4, 2), (6, 4), (4, 6), (1, 5),
(2, 4)],
[(4, 4), (3, 1), (4, 2), (6, 4), (4, 6), (1, 3),
(2, 4)]]
obstaclesMapStep8 = [[(5, 5), (4, 2), (2, 4), (6, 1), (6, 2), (6, 3),
(1, 6), (2, 6), (3, 6), (7, 5), (5, 7)]]
obstaclesCondition = [
obstaclesMapStep0, obstaclesMapStep1, obstaclesMapStep2,
obstaclesMapStep4, obstaclesMapStep6, obstaclesMapStep8
]
obstaclesMaps = dict(zip(decisionSteps, obstaclesCondition))
rotatePoint = RotatePoint(gridSize)
checkBoundary = CheckBoundary([0, gridSize - 1], [0, gridSize - 1])
noiseActionSpace = [(0, -1), (0, 1), (-1, 0), (1, 0), (1, 1), (1, -1),
(-1, -1), (-1, 1)]
normalNoise = AimActionWithNoise(noiseActionSpace, gridSize)
specialNoise = backToCrossPointNoise
initPrior = [0.5, 0.5]
# inferGoalPosterior = InferGoalPosterior(goalPolicy)
softmaxBetaList = [5]
noise = 0
gamma = 0.9
goalReward = [50, 50]
actionSpace = [(0, -1), (0, 1), (-1, 0), (1, 0)]
runVI = RunVI(gridSize, actionSpace, noiseActionSpace, noise, gamma,
goalReward)
for softmaxBeta in softmaxBetaList:
# for noise in noiseList:
for i in range(20):
print(i)
numBlocks = 3
expDesignValues = [[b, h, d, m, diff] for b in width
for h in height for d in intentionDis
for m in decisionSteps
for diff in targetDiffs] * numBlocks
random.shuffle(expDesignValues)
numExpTrial = len(expDesignValues)
condition = namedtuple('condition', 'name decisionSteps')
expCondition = condition(name='expCondition',
decisionSteps=decisionSteps)
lineCondition = condition(name='lineCondition',
decisionSteps=decisionSteps)
conditionList = [expCondition
] * numExpTrial # + [lineCondition] * numExpTrial
random.shuffle(conditionList)
numNormalTrials = len(conditionList)
noiseDesignValues = [0] * numNormalTrials
if len(conditionList) != len(noiseDesignValues):
raise Exception("unmatch condition design")
# deubg
# expDesignValues = [specialDesign] * 10
# noiseDesignValues = ['special'] * 10
# conditionList = [expCondition] * 10
# debug
isInBoundary = IsInBoundary([0, gridSize - 1], [0, gridSize - 1])
creatRectMap = CreatRectMap(rotateAngles, gridSize, obstaclesMaps,
rotatePoint)
creatLineMap = CreatLineMap(rotateAngles, gridSize, rotatePoint,
isInBoundary)
samplePositionFromCondition = SamplePositionFromCondition(
creatRectMap, creatLineMap, expDesignValues)
modelController = ModelControllerOnline(softmaxBeta)
# modelController = AvoidCommitModel(softmaxBeta, actionSpace, checkBoundary)
controller = modelController
renderOn = 1
normalTrial = NormalTrial(renderOn, controller, drawNewState,
drawText, normalNoise, checkBoundary)
experimentValues = co.OrderedDict()
experimentValues["name"] = "noise" + str(
noise) + '_' + "softmaxBeta" + str(softmaxBeta) + '_' + str(i)
resultsDirPath = os.path.join(
resultsPath,
"noise" + str(noise) + '_' + "softmaxBeta" + str(softmaxBeta))
if not os.path.exists(resultsDirPath):
os.mkdir(resultsDirPath)
writerPath = os.path.join(resultsDirPath,
experimentValues["name"] + '.csv')
writer = WriteDataFrameToCSV(writerPath)
experiment = ObstacleModelSimulation(normalTrial, writer,
experimentValues,
samplePositionFromCondition,
drawImage, resultsPath, runVI)
experiment(noiseDesignValues, conditionList)
def main():
picturePath = os.path.abspath(
os.path.join(os.path.join(os.getcwd(), os.pardir), 'pictures'))
resultsPath = os.path.abspath(
os.path.join(os.path.join(os.getcwd(), os.pardir), 'results'))
machinePolicyPath = os.path.abspath(
os.path.join(os.path.join(os.getcwd(), os.pardir), 'machinePolicy'))
dataPath = os.path.abspath(
os.path.join(os.path.join(os.getcwd(), os.pardir), 'conditionData'))
df = pd.read_csv(
os.path.join(dataPath, 'DesignConditionForAvoidCommitmentZone.csv'))
df['intentionedDisToTargetMin'] = df.apply(
lambda x: x['minDis'] - x['avoidCommitmentZone'], axis=1)
gridSize = 15
screenWidth = 600
screenHeight = 600
fullScreen = False
renderOn = False
initializeScreen = InitializeScreen(screenWidth, screenHeight, fullScreen)
screen = initializeScreen()
pg.mouse.set_visible(False)
leaveEdgeSpace = 2
lineWidth = 1
backgroundColor = [205, 255, 204]
lineColor = [0, 0, 0]
targetColor = [255, 50, 50]
playerColor = [50, 50, 255]
targetRadius = 10
playerRadius = 10
textColorTuple = (255, 50, 50)
introductionImage = pg.image.load(
os.path.join(picturePath, 'introduction.png'))
finishImage = pg.image.load(os.path.join(picturePath, 'finish.png'))
introductionImage = pg.transform.scale(introductionImage,
(screenWidth, screenHeight))
finishImage = pg.transform.scale(
finishImage, (int(screenWidth * 2 / 3), int(screenHeight / 4)))
drawBackground = DrawBackground(screen, gridSize, leaveEdgeSpace,
backgroundColor, lineColor, lineWidth,
textColorTuple)
drawText = DrawText(screen, drawBackground)
drawNewState = DrawNewState(screen, drawBackground, targetColor,
playerColor, targetRadius, playerRadius)
drawImage = DrawImage(screen)
# condition
width = [5]
height = [5]
intentionDis = [3, 4, 5]
decisionSteps = [2, 4, 6, 10]
targetDiffs = [0, 2, 4]
rotateAngles = [0, 90, 180, 270]
obstaclesMap1 = [(2, 2), (2, 4), (2, 5), (2, 6), (4, 2), (5, 2), (6, 2)]
obstaclesMap2 = [(3, 3), (4, 1), (1, 4), (5, 3), (3, 5), (6, 3), (3, 6)]
obstaclesMap3 = [(4, 4), (4, 1), (4, 2), (6, 4), (4, 6), (1, 4), (2, 4)]
speicalObstacleMap = [(4, 1), (4, 2), (6, 3), (6, 4), (1, 4), (2, 4),
(3, 6), (4, 6)]
obstaclesCondition = [
obstaclesMap1, obstaclesMap2, obstaclesMap3, speicalObstacleMap
]
obstaclesMaps = dict(zip(decisionSteps, obstaclesCondition))
rotatePoint = RotatePoint(gridSize)
checkBoundary = CheckBoundary([0, gridSize - 1], [0, gridSize - 1])
noiseActionSpace = [(0, -1), (0, 1), (-1, 0), (1, 0), (1, 1), (1, -1),
(-1, -1), (-1, 1)]
normalNoise = AimActionWithNoise(noiseActionSpace, gridSize)
specialNoise = backToCrossPointNoise
initPrior = [0.5, 0.5]
# inferGoalPosterior = InferGoalPosterior(goalPolicy)
softmaxBetaList = [6]
noiseList = [0]
# noise = 0.067
# for softmaxBeta in softmaxBetaList:
softmaxBeta = 6
for noise in noiseList:
for i in range(20):
print(i)
numBlocks = 5
expDesignValues = [[b, h, d, m, diff] for b in width
for h in height for d in intentionDis
for m in decisionSteps
for diff in targetDiffs] * numBlocks
random.shuffle(expDesignValues)
numExpTrial = len(expDesignValues)
specialDesign = [5, 5, 4, 10, 0]
expDesignValues.append(specialDesign)
condition = namedtuple('condition', 'name decisionSteps')
expCondition = condition(name='expCondition',
decisionSteps=decisionSteps[:-1])
lineCondition = condition(name='lineCondition',
decisionSteps=decisionSteps[:-1])
specialCondition = condition(name='specialCondition',
decisionSteps=[10])
numControlTrial = int(numExpTrial / 2)
conditionList = [expCondition] * numControlTrial + [
lineCondition
] * numControlTrial
conditionList = [lineCondition] * numControlTrial
random.shuffle(conditionList)
numNormalTrials = len(conditionList)
numTrialsPerBlock = 3
noiseCondition = list(permutations([1, 2, 0], numTrialsPerBlock))
noiseCondition.append((1, 1, 1))
blockNumber = int(numNormalTrials / numTrialsPerBlock)
noiseDesignValues = createNoiseDesignValue(noiseCondition,
blockNumber)
conditionList.append(specialCondition)
if noise == 0:
noiseDesignValues = [0] * len(conditionList)
if len(conditionList) != len(noiseDesignValues):
raise Exception("unmatch condition design")
# deubg
# expDesignValues = [specialDesign] * 10
# noiseDesignValues = ['special'] * 10
# conditionList = [expCondition] * 10
# debug
isInBoundary = IsInBoundary([0, gridSize - 1], [0, gridSize - 1])
creatRectMap = CreatRectMap(rotateAngles, gridSize, obstaclesMaps,
rotatePoint)
creatLineMap = CreatLineMap(rotateAngles, gridSize, rotatePoint,
isInBoundary)
samplePositionFromCondition = SamplePositionFromCondition(
creatRectMap, creatLineMap, expDesignValues)
runVI = RunVI(gridSize, noise, noiseActionSpace)
modelController = ModelControllerOnline(softmaxBeta, runVI)
controller = modelController
renderOn = 0
normalTrial = NormalTrial(renderOn, controller, drawNewState,
drawText, normalNoise, checkBoundary)
specialTrial = SpecialTrial(renderOn, controller, drawNewState,
drawText, specialNoise, checkBoundary)
experimentValues = co.OrderedDict()
experimentValues["name"] = "noise" + str(
noise) + '_' + "softmaxBeta" + str(softmaxBeta) + '_' + str(i)
writerPath = os.path.join(resultsPath,
experimentValues["name"] + '.csv')
writer = WriteDataFrameToCSV(writerPath)
experiment = ObstacleExperiment(normalTrial, specialTrial, writer,
experimentValues,
samplePositionFromCondition,
drawImage, resultsPath)
experiment(noiseDesignValues, conditionList)