class Gait(FeatureExtractorAbstract):
def __init__(self):
self.dc = DistanceCalc()
def getCSVheader(self):
return ["gaitPeriodX", "gaitErrorX", "gaitPeriodY", "gaitErrorY", "gaitPeriodZ", "gaitErrorZ"]
def extract(self, experiment, type, indiv):
filepath = experiment[2] + os.path.sep + PathConfig.traceFolderNormal + os.path.sep + indiv[0] + ".trace"
if not os.path.isfile(filepath):
filepath = experiment[2] + os.path.sep + PathConfig.traceFoldersAlt[type] + os.path.sep + indiv[
0] + ".trace"
if not os.path.isfile(filepath):
return ['NA'] * 3
with open(filepath) as fh:
xs = []
ys = []
zs = []
for line in fh:
linesplit = line.split("\t")
if not self.dc.isValidLine(linesplit):
linesplit = line.split(" ")
if not self.dc.isValidLine(linesplit):
continue
xs.append(linesplit[-3])
ys.append(linesplit[-2])
zs.append(linesplit[-1])
xs = map(float,xs)
ys = map(float,ys)
zs = map(float,zs)
xPeriod, xError = self._getPeriod(xs)
yPeriod, yError = self._getPeriod(ys)
zPeriod, zError = self._getPeriod(zs)
return xPeriod, xError, yPeriod, yError, zPeriod, zError
@staticmethod
def _getPeriod(signal):
if len(signal) == 0:
return 'NA'
signal = np.array(signal)
fft = np.fft.rfft(signal).real
fft = fft[:len(fft) / 2 + 1]
fft[1:] = fft[1:] / (len(signal)/2)
fft[0] = fft[0]/len(signal)
period = np.argmax(fft[1:]) + 1
period_value = fft[1:].max()
linspace = np.linspace(0,len(signal), len(signal))
mse = np.average(signal - (period_value * np.sin(period*linspace+np.average(signal)))**2)
return period, mse
def extract(self, experiment, type, indiv):
errorReturnVal = ['NA'] * 2
filepath = experiment[2] + os.path.sep + PathConfig.traceFolderNormal + os.path.sep + indiv[0] + ".trace"
dc = DistanceCalc()
if not os.path.isfile(filepath):
filepath = experiment[2] + os.path.sep + PathConfig.traceFoldersAlt[type] + os.path.sep + indiv[
0] + ".trace"
if not os.path.isfile(filepath):
return errorReturnVal
with open(filepath, 'r') as inputFile:
traceLines = []
lines = []
for line in inputFile:
lines.append(line)
if len(lines) == 110:
break
if len(lines) < 20:
return errorReturnVal
for line in lines[-10:]:
lineSplit = line.split("\t")
if not dc.isValidLine(lineSplit):
lineSplit = line.split(" ")
if not dc.isValidLine(lineSplit):
continue
traceLines.append(
(float(lineSplit[2]), float(lineSplit[2]))
)
if len(traceLines) < 6:
return errorReturnVal
self._setPoints(traceLines[0], traceLines[3], traceLines[5])
self._getCenter()
self._getRadius()
self._getCircumference()
dist = dc.distanceStep(filepath, "euclidean")
percentCircle = dist / self.circumference
return [self.radius, percentCircle]
def extract(self, experiment, type, indiv):
filepath = experiment[2] + os.path.sep + PathConfig.traceFolderNormal + os.path.sep + indiv[0] + ".trace"
dc = DistanceCalc()
if not os.path.isfile(filepath):
filepath = (
experiment[2] + os.path.sep + PathConfig.traceFoldersAlt[type] + os.path.sep + indiv[0] + ".trace"
)
if not os.path.isfile(filepath):
return ["NA"]
with open(filepath, "r") as inputFile:
monotonyUp = 0
monotonyDown = 0
monotonyLeft = 0
monotonyRight = 0
firstLine = True
xy = (0.0, 0.0)
for line in inputFile:
lineSplit = line.split("\t")
if not dc.isValidLine(lineSplit):
lineSplit = line.split(" ")
if not dc.isValidLine(lineSplit):
continue
if firstLine:
firstLine = False
xy = (lineSplit[2], lineSplit[3])
else:
xyNew = (lineSplit[2], lineSplit[3])
if xyNew[1] > xy[1]:
monotonyUp += 1
if xyNew[1] < xy[1]:
monotonyDown += 1
if xyNew[0] > xy[0]:
monotonyRight += 1
if xyNew[0] < xy[0]:
monotonyLeft += 1
xy = xyNew
return [monotonyRight - monotonyLeft, monotonyUp - monotonyDown]
def dataCollector(self):
listing = glob.glob(self.pattern)
td = TraceDistance()
dc = DistanceCalc()
with open(self.outputFile, "w") as csvfile:
headersWritten = False
print listing # Testing
for nestFile in listing:
print "Experiment folder: " + nestFile # Testing
expFolder = nestFile[-2:]
experimentNumber = str(expFolder)
popListing = glob.glob(nestFile + "/population/*.vxa")
for path in popListing:
indNumber = path[27:-8]
print "Population path: " + path # Testing
print "Ind. ID: " + indNumber # Testing
backupBasePath = nestFile + "/population_{condition}/" + indNumber + "_vox.vxa"
backup_path = os.path.abspath(backupBasePath.format(condition="beforePL"))
if not os.path.isfile(backup_path):
backup_path = os.path.abspath(backupBasePath.format(condition="MUT"))
if not os.path.isfile(backup_path):
backup_path = False
print "Backups path: " + backup_path # Testing
voxProbability, voxLifetime, voxCounts = self.voxCounter(experimentNumber, path, indNumber)
voxProb_orig, voxProbLife_orig, voxCounts_orig = self.voxCounter(
experimentNumber, backup_path, indNumber
)
traceFilename = os.path.abspath(nestFile + "/traces_afterPP/" + indNumber + ".trace")
mutatedBasePath = nestFile + "/traces_{condition}/" + indNumber + ".trace"
NOMUT_traceFilename = os.path.abspath(mutatedBasePath.format(condition="MUT"))
variant = "mut"
if not os.path.isfile(NOMUT_traceFilename):
NOMUT_traceFilename = os.path.abspath(mutatedBasePath.format(condition="NOMUT"))
variant = "nomut"
if not os.path.isfile(NOMUT_traceFilename):
NOMUT_traceFilename = False
if variant == "nomut":
fieldnames = [
"Ind_ID",
"birthtime",
"Exp_Num",
"probability",
"euclideanStep",
"manhattanStep",
"euclideanTotal",
"manhattanTotal",
"lifetime",
"size",
"totalMuscles",
"totalFat",
"totalBone",
"diseasedMuscle",
"diseasedFat",
"diseasedBone",
"diseasedEuclideanStep",
"diseasedManhattanStep",
"diseasedEuclideanTotal",
"diseasedManhattanTotal",
]
else:
fieldnames = [
"Ind_ID",
"birthtime",
"Exp_Num",
"probability",
"euclideanStep",
"manhattanStep",
"euclideanTotal",
"manhattanTotal",
"lifetime",
"size",
"totalMuscles",
"totalFat",
"totalBone",
"originalMuscle",
"originalFat",
"originalBone",
"originalEuclideanStep",
"originalManhattanStep",
"originalEuclideanTotal",
"originalManhattanTotal",
]
if not headersWritten:
writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
writer.writeheader()
headersWritten = True
if os.path.isfile(traceFilename):
distances = td.calcDistance(traceFilename)
birthtime = dc.getBirthTime(traceFilename)
if not birthtime:
birthtime = "NA"
print distances # Testing
print "Born at: " + str(birthtime)
else:
distances = ["NA", "NA", "NA", "NA", "NA"] # Batman
birthtime = ["NA"]
print indNumber + " trace file missing in /traces_afterPP/ of experiment " + experimentNumber
if NOMUT_traceFilename:
NOMUT_distances = td.calcDistance(NOMUT_traceFilename)
print NOMUT_distances
else:
NOMUT_distances = ["NA", "NA", "NA", "NA", "NA"] # Batman
print indNumber + " trace file missing in /traces_NOMUT/ of experiment " + experimentNumber
if variant == "mut":
columnNames = {
"Ind_ID": indNumber,
"birthtime": birthtime,
"Exp_Num": experimentNumber,
"probability": voxProbability,
"euclideanStep": distances[0],
"manhattanStep": distances[1],
"euclideanTotal": distances[2],
"manhattanTotal": distances[3],
"lifetime": voxLifetime,
"size": voxCounts[4],
"totalMuscles": voxCounts[3],
"totalFat": voxCounts[1],
"totalBone": voxCounts[2],
"originalMuscle": voxCounts_orig[3],
"originalFat": voxCounts_orig[1],
"originalBone": voxCounts_orig[2],
"originalEuclideanStep": NOMUT_distances[0],
"originalManhattanStep": NOMUT_distances[1],
"originalEuclideanTotal": NOMUT_distances[2],
"originalManhattanTotal": NOMUT_distances[3],
}
elif variant == "nomut":
columnNames = {
"Ind_ID": indNumber,
"birthtime": birthtime,
"Exp_Num": experimentNumber,
"probability": voxProbability,
"euclideanStep": distances[0],
"manhattanStep": distances[1],
"euclideanTotal": distances[2],
"manhattanTotal": distances[3],
"lifetime": voxLifetime,
"size": voxCounts[4],
"totalMuscles": voxCounts[3],
"totalFat": voxCounts[1],
"totalBone": voxCounts[2],
"diseasedMuscle": voxCounts_orig[3],
"diseasedFat": voxCounts_orig[1],
"diseasedBone": voxCounts_orig[2],
"diseasedEuclideanStep": NOMUT_distances[0],
"diseasedManhattanStep": NOMUT_distances[1],
"diseasedEuclideanTotal": NOMUT_distances[2],
"diseasedManhattanTotal": NOMUT_distances[3],
}
else:
print "File reading (mut/nomut) error!"
break
writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
writer.writerow(columnNames)
def __init__(self):
self.dc = DistanceCalc()
