#! /usr/bin/env python3
import argparse
import numpy as np
from sklearn import mixture
import matplotlib.pyplot as plt
from trafficintelligence import storage, ml
parser = argparse.ArgumentParser(description='The program learns and displays Gaussians fit to beginnings and ends of object trajectories (based on Mohamed Gomaa Mohamed 2015 PhD).')
parser.add_argument('-d', dest = 'databaseFilename', help = 'name of the Sqlite database file', required = True)
parser.add_argument('-t', dest = 'trajectoryType', help = 'type of trajectories to display', choices = ['feature', 'object'], default = 'object')
parser.add_argument('-n', dest = 'nObjects', help = 'number of objects to display', type = int)
parser.add_argument('-norigins', dest = 'nOriginClusters', help = 'number of clusters for trajectory origins', required = True, type = int)
parser.add_argument('-ndestinations', dest = 'nDestinationClusters', help = 'number of clusters for trajectory destinations (=norigins if not provided)', type = int)
parser.add_argument('--covariance-type', dest = 'covarianceType', help = 'type of covariance of Gaussian model', default = "full")
parser.add_argument('-w', dest = 'worldImageFilename', help = 'filename of the world image')
parser.add_argument('-u', dest = 'unitsPerPixel', help = 'number of units of distance per pixel', type = float, default = 1.)
parser.add_argument('--display', dest = 'display', help = 'displays points of interests', action = 'store_true') # default is manhattan distance
parser.add_argument('--assign', dest = 'assign', help = 'assigns the trajectories to the POIs and saves the assignments', action = 'store_true')
parser.add_argument('--display-paths', dest = 'displayPaths', help = 'displays all possible origin destination if assignment is done', action = 'store_true')
# TODO test Variational Bayesian Gaussian Mixture BayesianGaussianMixture
args = parser.parse_args()
objects = storage.loadTrajectoriesFromSqlite(args.databaseFilename, args.trajectoryType, args.nObjects)
beginnings = []
ends = []
for o in objects:
beginnings.append(o.getPositionAt(0).aslist())
ends.append(o.getPositionAt(int(o.length())-1).aslist())
if args.assign:
o.od = [-1, -1]
beginnings = np.array(beginnings)
ends = np.array(ends)
nDestinationClusters = args.nDestinationClusters
if args.nDestinationClusters is None:
nDestinationClusters = args.nOriginClusters
gmmId=0
models = {}
for nClusters, points, gmmType in zip([args.nOriginClusters, nDestinationClusters],
[beginnings, ends],
['beginning', 'end']):
# estimation
gmm = mixture.GaussianMixture(n_components=nClusters, covariance_type = args.covarianceType)
models[gmmType]=gmm.fit(points)
if not models[gmmType].converged_:
print('Warning: model for '+gmmType+' points did not converge')
if args.display or args.assign:
labels = models[gmmType].predict(points)
# plot
if args.display:
fig = plt.figure()
if args.worldImageFilename is not None and args.unitsPerPixel is not None:
img = plt.imread(args.worldImageFilename)
plt.imshow(img)
ml.plotGMMClusters(models[gmmType], labels, points, fig, nUnitsPerPixel = args.unitsPerPixel)
plt.axis('image')
plt.title(gmmType)
print(gmmType+' Clusters:\n{}'.format(ml.computeClusterSizes(labels, range(models[gmmType].n_components))))
# save
storage.savePOIsToSqlite(args.databaseFilename, models[gmmType], gmmType, gmmId)
# save assignments
if args.assign:
for o, l in zip(objects, labels):
if gmmType == 'beginning':
o.od[0] = l
elif gmmType == 'end':
o.od[1] = l
gmmId += 1
if args.assign:
storage.savePOIAssignments(args.databaseFilename, objects)
if args.displayPaths:
for i in range(args.nOriginClusters):
for j in range(args.nDestinationClusters):
odObjects = [o for o in objects if o.od[0] == i and o.od[1] == j]
if len(odObjects) > 0:
fig = plt.figure()
ax = fig.add_subplot(111)
ml.plotGMM(models['beginning'].means_[i], models['beginning'].covariances_[i], i, fig, 'b')
ml.plotGMM(models['end'].means_[j], models['end'].covariances_[j], j, fig, 'r')
for o in odObjects:
o.plot(withOrigin = True)
plt.title('OD {} to {}'.format(i,j))
plt.axis('equal')
plt.show()
if args.display:
plt.axis('equal')
plt.show()
