Mercurial Hosting > traffic-intelligence

annotate trafficintelligence/utils.py @ 1028:cc5cb04b04b0

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
major update using the trafficintelligence package name and install through pip
author Nicolas Saunier <nicolas.saunier@polymtl.ca>
date Fri, 15 Jun 2018 11:19:10 -0400
parents python/utils.py@a13f47c8931d
children c6cf75a2ed08
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
1 #! /usr/bin/env python
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
2 # -*- coding: utf-8 -*-
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
3 ''' Generic utilities.'''
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
4
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
5 import matplotlib.pyplot as plt
397
b36b00dd27c3 added function to read scene metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 395
diff changeset
6 from datetime import time, datetime
971
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
7 from argparse import ArgumentTypeError
1021
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
8 from pathlib import Path
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
9 from math import sqrt, ceil, floor
1028
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
10 from scipy.stats import rv_continuous, kruskal, shapiro, lognorm
689
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
11 from scipy.spatial import distance
840
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
12 from scipy.sparse import dok_matrix
1023
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
13 from numpy import zeros, array, exp, sum as npsum, int as npint, arange, cumsum, mean, median, percentile, isnan, ones, convolve, dtype, isnan, NaN, ma, isinf, savez, load as npload, log
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
14
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
15
421
4fce27946c60 first example of video metadata using sqlalchemy
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 405
diff changeset
16 datetimeFormat = "%Y-%m-%d %H:%M:%S"
4fce27946c60 first example of video metadata using sqlalchemy
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 405
diff changeset
17
969
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
18 sjcamDatetimeFormat = "%Y_%m%d_%H%M%S"#2017_0626_143720
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
19
185
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
20 #########################
742
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
21 # Strings
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
22 #########################
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
23
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
24 def upperCaseFirstLetter(s):
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
25 words = s.split(' ')
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
26 lowerWords = [w[0].upper()+w[1:].lower() for w in words]
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
27 return ' '.join(lowerWords)
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
28
971
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
29 class TimeConverter:
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
30 def __init__(self, datetimeFormat = datetimeFormat):
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
31 self.datetimeFormat = datetimeFormat
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
32
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
33 def convert(self, s):
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
34 try:
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
35 return datetime.strptime(s, self.datetimeFormat)
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
36 except ValueError:
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
37 msg = "Not a valid date: '{0}'.".format(s)
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
38 raise ArgumentTypeError(msg)
9897a13772fb added utils to load video sequence in metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 969
diff changeset
39
742
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
40 #########################
185
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
41 # Enumerations
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
42 #########################
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
43
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
44 def inverseEnumeration(l):
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
45 'Returns the dictionary that provides for each element in the input list its index in the input list'
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
46 result = {}
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
47 for i,x in enumerate(l):
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
48 result[x] = i
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
49 return result
155
f03fe3d6d0c8 added functions to parse options
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 152
diff changeset
50
f03fe3d6d0c8 added functions to parse options
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 152
diff changeset
51 #########################
637
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
52 # Simple statistics
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
53 #########################
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
54
680
da1352b89d02 classification is working
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 677
diff changeset
55 def logNormalMeanVar(loc, scale):
687
de278c5e65f6 minor comments for lognormal parameters (numpy and usual names differ)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 686
diff changeset
56 '''location and scale are respectively the mean and standard deviation of the normal in the log-normal distribution
854
33d296984dd8 rework and more info on speed probabilities for classification
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 847
diff changeset
57 https://en.wikipedia.org/wiki/Log-normal_distribution
33d296984dd8 rework and more info on speed probabilities for classification
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 847
diff changeset
58
33d296984dd8 rework and more info on speed probabilities for classification
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 847
diff changeset
59 same as lognorm.stats(scale, 0, exp(loc))'''
680
da1352b89d02 classification is working
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 677
diff changeset
60 mean = exp(loc+(scale**2)/2)
854
33d296984dd8 rework and more info on speed probabilities for classification
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 847
diff changeset
61 var = (exp(scale**2)-1)*exp(2*loc+scale**2)
680
da1352b89d02 classification is working
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 677
diff changeset
62 return mean, var
da1352b89d02 classification is working
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 677
diff changeset
63
855
2277ab1a8141 added utility for lognorm estimation
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 854
diff changeset
64 def fitLogNormal(x):
2277ab1a8141 added utility for lognorm estimation
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 854
diff changeset
65 'returns the fitted location and scale of the lognormal (general definition)'
2277ab1a8141 added utility for lognorm estimation
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 854
diff changeset
66 shape, loc, scale = lognorm.fit(x, floc=0.)
2277ab1a8141 added utility for lognorm estimation
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 854
diff changeset
67 return log(scale), shape
2277ab1a8141 added utility for lognorm estimation
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 854
diff changeset
68
859
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
69 def sampleSize(stdev, tolerance, percentConfidence, nRoundingDigits = None, printLatex = False):
301
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
70 from scipy.stats.distributions import norm
859
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
71 if nRoundingDigits is None:
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
72 k = round(norm.ppf(0.5+percentConfidence/200., 0, 1), 2) # 1.-(100-percentConfidence)/200.
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
73 else:
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
74 k = round(norm.ppf(0.5+percentConfidence/200., 0, 1), nRoundingDigits)
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
75 stdev = round(stdev, nRoundingDigits)
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
76 tolerance = round(tolerance, nRoundingDigits)
423
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
77 if printLatex:
859
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
78 print('$z_{{{}}}^2\\frac{{s^2}}{{e^2}}={}^2\\frac{{{}^2}}{{{}^2}}$'.format(0.5+percentConfidence/200.,k, stdev, tolerance))
423
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
79 return (k*stdev/tolerance)**2
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
80
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
81 def confidenceInterval(mean, stdev, nSamples, percentConfidence, trueStd = True, printLatex = False):
499
0a93afea8243 alternative confidence interval
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 491
diff changeset
82 '''if trueStd, use normal distribution, otherwise, Student
0a93afea8243 alternative confidence interval
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 491
diff changeset
83
973
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 971
diff changeset
84 Use otherwise t.interval or norm.interval for the boundaries
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 971
diff changeset
85 ex: norm.interval(0.95)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 971
diff changeset
86 t.interval(0.95, nSamples-1)'''
423
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
87 from scipy.stats.distributions import norm, t
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
88 if trueStd:
859
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
89 k = round(norm.ppf(0.5+percentConfidence/200., 0, 1), 2)
423
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
90 else: # use Student
859
a8de3c93f6b7 minor modifications to helper stat functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 856
diff changeset
91 k = round(t.ppf(0.5+percentConfidence/200., nSamples-1), 2)
301
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
92 e = k*stdev/sqrt(nSamples)
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
93 if printLatex:
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
94 print('${0} \pm {1}\\frac{{{2}}}{{\sqrt{{{3}}}}}$'.format(mean, k, stdev, nSamples))
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
95 return mean-e, mean+e
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
96
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
97 def computeChi2(expected, observed):
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
98 '''Returns the Chi2 statistics'''
876
c7e72d758049 minor update to avoid integer issue
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 859
diff changeset
99 return sum([((e-o)*(e-o))/float(e) for e, o in zip(expected, observed)])
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
100
1028
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
101 class generateDistribution(rv_continuous):
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
102 def __init__(self, values, probabilities):
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
103 '''The values (and corresponding probabilities) are supposed to be sorted by value
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
104 for v, p in zip(values, probabilities): P(X<=v)=p'''
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
105 assert probabilities[0]==0
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
106 self.values = values
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
107 self.probabilities = probabilities
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
108
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
109 def _cdf(self, x):
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
110 if x < self.values[0]:
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
111 return self.probabilities[0]
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
112 else:
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
113 i=0
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
114 while i+1<len(self.values) and self.values[i+1] < x:
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
115 i += 1
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
116 if i == len(self.values)-1:
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
117 return self.probabilities[-1]
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
118 else:
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
119 return (self.probabilities[i+1]-self.probabilities[i])/(self.values[i+1]-self.values[i])
cc5cb04b04b0 major update using the trafficintelligence package name and install through pip
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1023
diff changeset
120
749
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
121 class DistributionSample(object):
85
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
122 def nSamples(self):
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
123 return sum(self.counts)
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
124
588
c5406edbcf12 added loading ground truth annotations (ground truth) from polytrack format
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 574
diff changeset
125 def cumulativeDensityFunction(sample, normalized = False):
276
78922b4de3bf minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 266
diff changeset
126 '''Returns the cumulative density function of the sample of a random variable'''
588
c5406edbcf12 added loading ground truth annotations (ground truth) from polytrack format
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 574
diff changeset
127 xaxis = sorted(sample)
c5406edbcf12 added loading ground truth annotations (ground truth) from polytrack format
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 574
diff changeset
128 counts = arange(1,len(sample)+1) # dtype = float
c5406edbcf12 added loading ground truth annotations (ground truth) from polytrack format
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 574
diff changeset
129 if normalized:
c5406edbcf12 added loading ground truth annotations (ground truth) from polytrack format
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 574
diff changeset
130 counts /= float(len(sample))
197
2788b2827670 simple cumulatie function distribution computation
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 185
diff changeset
131 return xaxis, counts
85
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
132
749
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
133 class DiscreteDistributionSample(DistributionSample):
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
134 '''Class to represent a sample of a distribution for a discrete random variable'''
85
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
135 def __init__(self, categories, counts):
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
136 self.categories = categories
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
137 self.counts = counts
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
138
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
139 def mean(self):
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
140 result = [float(x*y) for x,y in zip(self.categories, self.counts)]
672
5473b7460375 moved and rationalized imports in modules
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 671
diff changeset
141 return npsum(result)/self.nSamples()
85
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
142
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
143 def var(self, mean = None):
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
144 if not mean:
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
145 m = self.mean()
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
146 else:
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
147 m = mean
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
148 result = 0.
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
149 squares = [float((x-m)*(x-m)*y) for x,y in zip(self.categories, self.counts)]
672
5473b7460375 moved and rationalized imports in modules
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 671
diff changeset
150 return npsum(squares)/(self.nSamples()-1)
85
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
151
86
f03ec4697a09 corrected discrete distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 85
diff changeset
152 def referenceCounts(self, probability):
f03ec4697a09 corrected discrete distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 85
diff changeset
153 '''probability is a function that returns the probability of the random variable for the category values'''
f03ec4697a09 corrected discrete distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 85
diff changeset
154 refProba = [probability(c) for c in self.categories]
672
5473b7460375 moved and rationalized imports in modules
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 671
diff changeset
155 refProba[-1] = 1-npsum(refProba[:-1])
86
f03ec4697a09 corrected discrete distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 85
diff changeset
156 refCounts = [r*self.nSamples() for r in refProba]
f03ec4697a09 corrected discrete distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 85
diff changeset
157 return refCounts, refProba
85
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
158
749
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
159 class ContinuousDistributionSample(DistributionSample):
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
160 '''Class to represent a sample of a distribution for a continuous random variable
76
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
161 with the number of observations for each interval
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
162 intervals (categories variable) are defined by their left limits, the last one being the right limit
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
163 categories contain therefore one more element than the counts'''
35
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
164 def __init__(self, categories, counts):
276
78922b4de3bf minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 266
diff changeset
165 # todo add samples for initialization and everything to None? (or setSamples?)
35
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
166 self.categories = categories
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
167 self.counts = counts
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
168
749
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
169 @staticmethod
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
170 def generate(sample, categories):
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
171 if min(sample) < min(categories):
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
172 print('Sample has lower min than proposed categories ({}, {})'.format(min(sample), min(categories)))
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
173 if max(sample) > max(categories):
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
174 print('Sample has higher max than proposed categories ({}, {})'.format(max(sample), max(categories)))
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
175 dist = ContinuousDistributionSample(sorted(categories), [0]*(len(categories)-1))
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
176 for s in sample:
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
177 i = 0
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
178 while i<len(dist.categories) and dist.categories[i] <= s:
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
179 i += 1
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
180 if i <= len(dist.counts):
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
181 dist.counts[i-1] += 1
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
182 #print('{} in {} {}'.format(s, dist.categories[i-1], dist.categories[i]))
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
183 else:
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
184 print('Element {} is not in the categories'.format(s))
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
185 return dist
10dbab1e871d modifications in samples and distributions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 742
diff changeset
186
35
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
187 def mean(self):
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
188 result = 0.
76
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
189 for i in range(len(self.counts)-1):
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
190 result += self.counts[i]*(self.categories[i]+self.categories[i+1])/2
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
191 return result/self.nSamples()
35
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
192
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
193 def var(self, mean = None):
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
194 if not mean:
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
195 m = self.mean()
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
196 else:
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
197 m = mean
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
198 result = 0.
76
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
199 for i in range(len(self.counts)-1):
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
200 mid = (self.categories[i]+self.categories[i+1])/2
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
201 result += self.counts[i]*(mid - m)*(mid - m)
35
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
202 return result/(self.nSamples()-1)
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
203
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
204 def referenceCounts(self, cdf):
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
205 '''cdf is a cumulative distribution function
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
206 returning the probability of the variable being less that x'''
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
207 # refCumulativeCounts = [0]#[cdf(self.categories[0][0])]
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
208 # for inter in self.categories:
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
209 # refCumulativeCounts.append(cdf(inter[1]))
76
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
210 refCumulativeCounts = [cdf(x) for x in self.categories[1:-1]]
35
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
211
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
212 refProba = [refCumulativeCounts[0]]
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
213 for i in xrange(1,len(refCumulativeCounts)):
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
214 refProba.append(refCumulativeCounts[i]-refCumulativeCounts[i-1])
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
215 refProba.append(1-refCumulativeCounts[-1])
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
216 refCounts = [p*self.nSamples() for p in refProba]
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
217
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
218 return refCounts, refProba
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
219
77
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
220 def printReferenceCounts(self, refCounts=None):
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
221 if refCounts:
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
222 ref = refCounts
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
223 else:
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
224 ref = self.referenceCounts
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
225 for i in xrange(len(ref[0])):
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
226 print('{0}-{1} & {2:0.3} & {3:0.3} \\\\'.format(self.categories[i],self.categories[i+1],ref[1][i], ref[0][i]))
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
227
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
228
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
229 #########################
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
230 # maths section
27
44689029a86f updated segmentIntersection and other
Nicolas Saunier <nico@confins.net>
parents: 24
diff changeset
231 #########################
24
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
232
433
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
233 # def kernelSmoothing(sampleX, X, Y, weightFunc, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
234 # '''Returns a smoothed weighted version of Y at the predefined values of sampleX
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
235 # Sum_x weight(sample_x,x) * y(x)'''
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
236 # from numpy import zeros, array
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
237 # smoothed = zeros(len(sampleX))
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
238 # for i,x in enumerate(sampleX):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
239 # weights = array([weightFunc(x,xx, halfwidth) for xx in X])
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
240 # if sum(weights)>0:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
241 # smoothed[i] = sum(weights*Y)/sum(weights)
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
242 # else:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
243 # smoothed[i] = 0
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
244 # return smoothed
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
245
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
246 def kernelSmoothing(x, X, Y, weightFunc, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
247 '''Returns the smoothed estimate of (X,Y) at x
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
248 Sum_x weight(sample_x,x) * y(x)'''
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
249 weights = array([weightFunc(x,observedx, halfwidth) for observedx in X])
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
250 if sum(weights)>0:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
251 return sum(weights*Y)/sum(weights)
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
252 else:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
253 return 0
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
254
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
255 def uniform(center, x, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
256 if abs(center-x)<halfwidth:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
257 return 1.
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
258 else:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
259 return 0.
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
260
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
261 def gaussian(center, x, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
262 return exp(-((center-x)/halfwidth)**2/2)
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
263
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
264 def epanechnikov(center, x, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
265 diff = abs(center-x)
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
266 if diff<halfwidth:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
267 return 1.-(diff/halfwidth)**2
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
268 else:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
269 return 0.
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
270
434
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
271 def triangular(center, x, halfwidth):
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
272 diff = abs(center-x)
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
273 if diff<halfwidth:
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
274 return 1.-abs(diff/halfwidth)
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
275 else:
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
276 return 0.
433
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
277
518
0c86c73f3c09 median smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 511
diff changeset
278 def medianSmoothing(x, X, Y, halfwidth):
0c86c73f3c09 median smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 511
diff changeset
279 '''Returns the media of Y's corresponding to X's in the interval [x-halfwidth, x+halfwidth]'''
0c86c73f3c09 median smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 511
diff changeset
280 return median([y for observedx, y in zip(X,Y) if abs(x-observedx)<halfwidth])
0c86c73f3c09 median smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 511
diff changeset
281
521
3707eeb20f25 changed argMaxDict name to argmaxDict
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 518
diff changeset
282 def argmaxDict(d):
561
ee45c6eb6d49 added Mohamed Gomaa Mohamed function to smooth object trajectories
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 553
diff changeset
283 return max(d, key=d.get)
279
3af4c267a7bf generic simple LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 276
diff changeset
284
837
e01cabca4c55 minor modifications to merge-features
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 749
diff changeset
285 def deltaFrames(t1, t2, frameRate):
e01cabca4c55 minor modifications to merge-features
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 749
diff changeset
286 '''Returns the number of frames between t1 and t2
e01cabca4c55 minor modifications to merge-features
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 749
diff changeset
287 positive if t1<=t2, negative otherwise'''
e01cabca4c55 minor modifications to merge-features
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 749
diff changeset
288 if t1 > t2:
e01cabca4c55 minor modifications to merge-features
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 749
diff changeset
289 return -(t1-t2).seconds*frameRate
e01cabca4c55 minor modifications to merge-features
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 749
diff changeset
290 else:
e01cabca4c55 minor modifications to merge-features
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 749
diff changeset
291 return (t2-t1).seconds*frameRate
e01cabca4c55 minor modifications to merge-features
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 749
diff changeset
292
395
6fba1ab040f1 minor modification to framestotime
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 391
diff changeset
293 def framesToTime(nFrames, frameRate, initialTime = time()):
6fba1ab040f1 minor modification to framestotime
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 391
diff changeset
294 '''returns a datetime.time for the time in hour, minutes and seconds
6fba1ab040f1 minor modification to framestotime
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 391
diff changeset
295 initialTime is a datetime.time'''
6fba1ab040f1 minor modification to framestotime
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 391
diff changeset
296 seconds = int(floor(float(nFrames)/float(frameRate))+initialTime.hour*3600+initialTime.minute*60+initialTime.second)
261
4aa792cb0fa9 changing framesToTime to return a datetime.time
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 248
diff changeset
297 h = int(floor(seconds/3600.))
248
571ba5ed22e2 added utils for bus processing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 241
diff changeset
298 seconds = seconds - h*3600
261
4aa792cb0fa9 changing framesToTime to return a datetime.time
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 248
diff changeset
299 m = int(floor(seconds/60))
248
571ba5ed22e2 added utils for bus processing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 241
diff changeset
300 seconds = seconds - m*60
262
a048066bd20f correcting bug in framesToTime
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 261
diff changeset
301 return time(h, m, seconds)
248
571ba5ed22e2 added utils for bus processing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 241
diff changeset
302
381
387cc0142211 script to replay event annotations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 376
diff changeset
303 def timeToFrames(t, frameRate):
387cc0142211 script to replay event annotations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 376
diff changeset
304 return frameRate*(t.hour*3600+t.minute*60+t.second)
387cc0142211 script to replay event annotations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 376
diff changeset
305
241
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
306 def sortXY(X,Y):
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
307 'returns the sorted (x, Y(x)) sorted on X'
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
308 D = {}
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
309 for x, y in zip(X,Y):
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
310 D[x]=y
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
311 xsorted = sorted(D.keys())
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
312 return xsorted, [D[x] for x in xsorted]
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
313
733
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
314 def compareLengthForSort(i, j):
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
315 if len(i) < len(j):
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
316 return -1
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
317 elif len(i) == len(j):
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
318 return 0
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
319 else:
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
320 return 1
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
321
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
322 def sortByLength(instances, reverse = False):
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
323 '''Returns a new list with the instances sorted by length (method __len__)
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
324 reverse is passed to sorted'''
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
325 return sorted(instances, key = len, reverse = reverse)
733
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
326
32
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
327 def ceilDecimals(v, nDecimals):
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
328 '''Rounds the number at the nth decimal
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
329 eg 1.23 at 0 decimal is 2, at 1 decimal is 1.3'''
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
330 tens = 10**nDecimals
32
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
331 return ceil(v*tens)/tens
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
332
152
74b1fc68d4df re-organized code to avoid cyclic python module dependencies
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 116
diff changeset
333 def inBetween(bound1, bound2, x):
698
8d99a9e16644 added clarification comments
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 689
diff changeset
334 'useful if one does not know the order of bound1/bound2'
569
0057c04f94d5 work in progress on intersections (for PET)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 561
diff changeset
335 return bound1 <= x <= bound2 or bound2 <= x <= bound1
0057c04f94d5 work in progress on intersections (for PET)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 561
diff changeset
336
0057c04f94d5 work in progress on intersections (for PET)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 561
diff changeset
337 def pointDistanceL2(x1,y1,x2,y2):
0057c04f94d5 work in progress on intersections (for PET)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 561
diff changeset
338 ''' Compute point-to-point distance (L2 norm, ie Euclidean distance)'''
0057c04f94d5 work in progress on intersections (for PET)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 561
diff changeset
339 return sqrt((x2-x1)**2+(y2-y1)**2)
24
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
340
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
341 def crossProduct(l1, l2):
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
342 return l1[0]*l2[1]-l1[1]*l2[0]
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
343
574
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
344 def cat_mvgavg(cat_list, halfWidth):
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
345 ''' Return a list of categories/values smoothed according to a window.
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
346 halfWidth is the search radius on either side'''
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
347 from copy import deepcopy
659
784298512b60 minor modifications
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 637
diff changeset
348 smoothed = deepcopy(cat_list)
784298512b60 minor modifications
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 637
diff changeset
349 for point in range(len(cat_list)):
574
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
350 lower_bound_check = max(0,point-halfWidth)
659
784298512b60 minor modifications
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 637
diff changeset
351 upper_bound_check = min(len(cat_list)-1,point+halfWidth+1)
784298512b60 minor modifications
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 637
diff changeset
352 window_values = cat_list[lower_bound_check:upper_bound_check]
574
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
353 smoothed[point] = max(set(window_values), key=window_values.count)
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
354 return smoothed
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
355
547
97c5fef5b2d6 corrected bugs
MohamedGomaa
parents: 521
diff changeset
356 def filterMovingWindow(inputSignal, halfWidth):
29
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
357 '''Returns an array obtained after the smoothing of the input by a moving average
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
358 The first and last points are copied from the original.'''
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
359 width = float(halfWidth*2+1)
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
360 win = ones(width,'d')
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
361 result = convolve(win/width,array(inputSignal),'same')
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
362 result[:halfWidth] = inputSignal[:halfWidth]
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
363 result[-halfWidth:] = inputSignal[-halfWidth:]
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
364 return result
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
365
199
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
366 def linearRegression(x, y, deg = 1, plotData = False):
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
367 '''returns the least square estimation of the linear regression of y = ax+b
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
368 as well as the plot'''
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
369 from numpy.lib.polynomial import polyfit
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
370 from numpy.core.multiarray import arange
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
371 coef = polyfit(x, y, deg)
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
372 if plotData:
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
373 def poly(x):
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
374 result = 0
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
375 for i in range(len(coef)):
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
376 result += coef[i]*x**(len(coef)-i-1)
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
377 return result
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
378 plt.plot(x, y, 'x')
199
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
379 xx = arange(min(x), max(x),(max(x)-min(x))/1000)
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
380 plt.plot(xx, [poly(z) for z in xx])
199
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
381 return coef
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
382
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
383 def correlation(data, correlationMethod = 'pearson', plotFigure = False, displayNames = None, figureFilename = None):
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
384 '''Computes (and displays) the correlation matrix for a pandas DataFrame'''
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
385 columns = data.columns.tolist()
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
386 for var in data.columns:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
387 uniqueValues = data[var].unique()
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
388 if len(uniqueValues) == 1 or data.dtypes[var] == dtype('O') or (len(uniqueValues) == 2 and len(data.loc[~isnan(data[var]), var].unique()) == 1): # last condition: only one other value than nan
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
389 columns.remove(var)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
390 c=data[columns].corr(correlationMethod)
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
391 if plotFigure:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
392 fig = plt.figure(figsize=(4+0.4*c.shape[0], 0.4*c.shape[0]))
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
393 fig.add_subplot(1,1,1)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
394 #plt.imshow(np.fabs(c), interpolation='none')
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
395 plt.imshow(c, vmin=-1., vmax = 1., interpolation='none', cmap = 'RdYlBu_r') # coolwarm
847
36c5bee9a887 bug correction
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 841
diff changeset
396 if displayNames is not None:
36c5bee9a887 bug correction
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 841
diff changeset
397 colnames = [displayNames.get(s.strip(), s.strip()) for s in columns]
36c5bee9a887 bug correction
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 841
diff changeset
398 else:
36c5bee9a887 bug correction
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 841
diff changeset
399 colnames = columns
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
400 #correlation.plot_corr(c, xnames = colnames, normcolor=True, title = filename)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
401 plt.xticks(range(len(colnames)), colnames, rotation=90)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
402 plt.yticks(range(len(colnames)), colnames)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
403 plt.tick_params('both', length=0)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
404 plt.subplots_adjust(bottom = 0.29)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
405 plt.colorbar()
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
406 plt.title('Correlation ({})'.format(correlationMethod))
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
407 plt.tight_layout()
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
408 if len(colnames) > 50:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
409 plt.subplots_adjust(left=.06)
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
410 if figureFilename is not None:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
411 plt.savefig(figureFilename, dpi = 150, transparent = True)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
412 return c
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
413
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
414 def addDummies(data, variables, allVariables = True):
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
415 '''Add binary dummy variables for each value of a nominal variable
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
416 in a pandas DataFrame'''
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
417 newVariables = []
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
418 for var in variables:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
419 if var in data.columns and data.dtypes[var] == dtype('O') and len(data[var].unique()) > 2:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
420 values = data[var].unique()
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
421 if not allVariables:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
422 values = values[:-1]
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
423 for val in values:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
424 if val is not NaN:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
425 newVariable = (var+'_{}'.format(val)).replace('.','').replace(' ','').replace('-','')
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
426 data[newVariable] = (data[var] == val)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
427 newVariables.append(newVariable)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
428 return newVariables
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
429
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
430 def kruskalWallis(data, dependentVariable, independentVariable, plotFigure = False, filenamePrefix = None, figureFileType = 'pdf', saveLatex = False, renameVariables = lambda s: s, kwCaption = ''):
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
431 '''Studies the influence of (nominal) independent variable over the dependent variable
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
432
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
433 Makes tests if the conditional distributions are normal
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
434 using the Shapiro-Wilk test (in which case ANOVA could be used)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
435 Implements uses the non-parametric Kruskal Wallis test'''
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
436 tmp = data[data[independentVariable].notnull()]
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
437 independentVariableValues = sorted(tmp[independentVariable].unique().tolist())
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
438 if len(independentVariableValues) >= 2:
674
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
439 if saveLatex:
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
440 from storage import openCheck
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
441 out = openCheck(filenamePrefix+'-{}-{}.tex'.format(dependentVariable, independentVariable), 'w')
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
442 for x in independentVariableValues:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
443 print('Shapiro-Wilk normality test for {} when {}={}: {} obs'.format(dependentVariable,independentVariable, x, len(tmp.loc[tmp[independentVariable] == x, dependentVariable])))
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
444 if len(tmp.loc[tmp[independentVariable] == x, dependentVariable]) >= 3:
978
184f1dd307f9 corrected print and exception statements for Python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 973
diff changeset
445 print(shapiro(tmp.loc[tmp[independentVariable] == x, dependentVariable]))
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
446 if plotFigure:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
447 plt.figure()
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
448 plt.boxplot([tmp.loc[tmp[independentVariable] == x, dependentVariable] for x in independentVariableValues])
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
449 plt.xticks(range(1,len(independentVariableValues)+1), independentVariableValues)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
450 plt.title('{} vs {}'.format(dependentVariable, independentVariable))
674
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
451 if filenamePrefix is not None:
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
452 plt.savefig(filenamePrefix+'-{}-{}.{}'.format(dependentVariable, independentVariable, figureFileType))
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
453 table = tmp.groupby([independentVariable])[dependentVariable].describe().unstack().sort(['50%'], ascending = False)
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
454 table['count'] = table['count'].astype(int)
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
455 testResult = kruskal(*[tmp.loc[tmp[independentVariable] == x, dependentVariable] for x in independentVariableValues])
674
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
456 if saveLatex:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
457 out.write('\\begin{minipage}{\\linewidth}\n'
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
458 +'\\centering\n'
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
459 +'\\captionof{table}{'+(kwCaption.format(dependentVariable, independentVariable, *testResult))+'}\n'
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
460 +table.to_latex(float_format = lambda x: '{:.3f}'.format(x)).encode('ascii')+'\n'
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
461 +'\\end{minipage}\n'
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
462 +'\\ \\vspace{0.5cm}\n')
674
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
463 else:
978
184f1dd307f9 corrected print and exception statements for Python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 973
diff changeset
464 print(table)
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
465 return testResult
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
466 else:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
467 return None
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
468
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
469 def prepareRegression(data, dependentVariable, independentVariables, maxCorrelationThreshold, correlations, maxCorrelationP, correlationFunc, stdoutText = ['Removing {} (constant: {})', 'Removing {} (correlation {} with {})', 'Removing {} (no correlation: {}, p={})'], saveFiles = False, filenamePrefix = None, latexHeader = '', latexTable = None, latexFooter=''):
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
470 '''Removes variables from candidate independent variables if
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
471 - if two independent variables are correlated (> maxCorrelationThreshold), one is removed
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
472 - if an independent variable is not correlated with the dependent variable (p>maxCorrelationP)
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
473 Returns the remaining non-correlated variables, correlated with the dependent variable
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
474
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
475 correlationFunc is spearmanr or pearsonr from scipy.stats
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
476 text is the template to display for the two types of printout (see default): 3 elements if no saving to latex file, 8 otherwise
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
477
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
478 TODO: pass the dummies for nominal variables and remove if all dummies are correlated, or none is correlated with the dependentvariable'''
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
479 from copy import copy
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
480 from pandas import DataFrame
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
481 result = copy(independentVariables)
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
482 table1 = ''
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
483 table2 = {}
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
484 # constant variables
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
485 for var in independentVariables:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
486 uniqueValues = data[var].unique()
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
487 if (len(uniqueValues) == 1) or (len(uniqueValues) == 2 and uniqueValues.dtype != dtype('O') and len(data.loc[~isnan(data[var]), var].unique()) == 1):
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
488 print(stdoutText[0].format(var, uniqueValues))
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
489 if saveFiles:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
490 table1 += latexTable[0].format(var, *uniqueValues)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
491 result.remove(var)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
492 # correlated variables
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
493 for v1 in copy(result):
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
494 if v1 in correlations.index:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
495 for v2 in copy(result):
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
496 if v2 != v1 and v2 in correlations.index:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
497 if abs(correlations.loc[v1, v2]) > maxCorrelationThreshold:
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
498 if v1 in result and v2 in result:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
499 if saveFiles:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
500 table1 += latexTable[1].format(v2, v1, correlations.loc[v1, v2])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
501 print(stdoutText[1].format(v2, v1, correlations.loc[v1, v2]))
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
502 result.remove(v2)
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
503 # not correlated with dependent variable
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
504 table2['Correlations'] = []
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
505 table2['Valeurs p'] = []
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
506 for var in copy(result):
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
507 if data.dtypes[var] != dtype('O'):
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
508 cor, p = correlationFunc(data[dependentVariable], data[var])
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
509 if p > maxCorrelationP:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
510 if saveFiles:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
511 table1 += latexTable[2].format(var, cor, p)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
512 print(stdoutText[2].format(var, cor, p))
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
513 result.remove(var)
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
514 else:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
515 table2['Correlations'].append(cor)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
516 table2['Valeurs p'].append(p)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
517
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
518 if saveFiles:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
519 from storage import openCheck
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
520 out = openCheck(filenamePrefix+'-removed-variables.tex', 'w')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
521 out.write(latexHeader)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
522 out.write(table1)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
523 out.write(latexFooter)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
524 out.close()
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
525 out = openCheck(filenamePrefix+'-correlations.html', 'w')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
526 table2['Variables'] = [var for var in result if data.dtypes[var] != dtype('O')]
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
527 out.write(DataFrame(table2)[['Variables', 'Correlations', 'Valeurs p']].to_html(formatters = {'Correlations': lambda x: '{:.2f}'.format(x), 'Valeurs p': lambda x: '{:.3f}'.format(x)}, index = False))
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
528 out.close()
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
529 return result
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
530
841
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 840
diff changeset
531 def saveDokMatrix(filename, m, lowerTriangle = False):
840
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
532 'Saves a dok_matrix using savez'
841
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 840
diff changeset
533 if lowerTriangle:
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
534 keys = [k for k in m if k[0] > k[1]]
841
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 840
diff changeset
535 savez(filename, shape = m.shape, keys = keys, values = [m[k[0],k[1]] for k in keys])
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 840
diff changeset
536 else:
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
537 savez(filename, shape = m.shape, keys = list(m.keys()), values = list(m.values()))
840
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
538
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
539 def loadDokMatrix(filename):
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
540 'Loads a dok_matrix saved using the above saveDokMatrix'
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
541 data = npload(filename)
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
542 m = dok_matrix(tuple(data['shape']))
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
543 for k, v in zip(data['keys'], data['values']):
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
544 m[tuple(k)] = v
15a82ebc62c4 utils for sparse matrix
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 837
diff changeset
545 return m
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
546
1023
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
547 def aggregationFunction(funcStr, centile = 50):
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
548 '''return the numpy function corresponding to funcStr
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
549 centile can be a list of centiles to compute at once, eg [25, 50, 75] for the 3 quartiles'''
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
550 if funcStr == 'median':
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
551 return median
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
552 elif funcStr == 'mean':
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
553 return mean
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
554 elif funcStr == 'centile':
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
555 return lambda x: percentile(x, centile)
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
556 elif funcStr == '85centile':
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
557 return lambda x: percentile(x, 85)
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
558 else:
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
559 print('Unknown aggregation method: {}'.format(funcStr))
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
560 return None
a13f47c8931d work on processing large datasets (generate speed data)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 1022
diff changeset
561
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
562 #########################
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
563 # regression analysis using statsmodels (and pandas)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
564 #########################
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
565
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
566 # TODO make class for experiments?
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
567 # TODO add tests with public dataset downloaded from Internet (IRIS et al)
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
568 def modelString(experiment, dependentVariable, independentVariables):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
569 return dependentVariable+' ~ '+' + '.join([independentVariable for independentVariable in independentVariables if experiment[independentVariable]])
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
570
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
571 def runModel(experiment, data, dependentVariable, independentVariables, regressionType = 'ols'):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
572 import statsmodels.formula.api as smf
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
573 modelStr = modelString(experiment, dependentVariable, independentVariables)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
574 if regressionType == 'ols':
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
575 model = smf.ols(modelStr, data = data)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
576 elif regressionType == 'gls':
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
577 model = smf.gls(modelStr, data = data)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
578 elif regressionType == 'rlm':
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
579 model = smf.rlm(modelStr, data = data)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
580 else:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
581 print('Unknown regression type {}. Exiting'.format(regressionType))
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
582 import sys
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
583 sys.exit()
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
584 return model.fit()
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
585
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
586 def runModels(experiments, data, dependentVariable, independentVariables, regressionType = 'ols'):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
587 '''Runs several models and stores 3 statistics
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
588 adjusted R2, condition number (should be small, eg < 1000)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
589 and p-value for Shapiro-Wilk test of residual normality'''
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
590 for i,experiment in experiments.iterrows():
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
591 if experiment[independentVariables].any():
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
592 results = runModel(experiment, data, dependentVariable, independentVariables, regressionType = 'ols')
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
593 experiments.loc[i,'r2adj'] = results.rsquared_adj
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
594 experiments.loc[i,'condNum'] = results.condition_number
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
595 experiments.loc[i, 'shapiroP'] = shapiro(results.resid)[1]
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
596 experiments.loc[i,'nobs'] = int(results.nobs)
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
597 return experiments
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
598
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
599 def generateExperiments(independentVariables):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
600 '''Generates all possible models for including or not each independent variable'''
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
601 from pandas import DataFrame
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
602 experiments = {}
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
603 nIndependentVariables = len(independentVariables)
669
df6be882f325 bug corrected
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 668
diff changeset
604 if nIndependentVariables != len(set(independentVariables)):
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
605 print("Duplicate variables. Exiting")
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
606 import sys
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
607 sys.exit()
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
608 nModels = 2**nIndependentVariables
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
609 for i,var in enumerate(independentVariables):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
610 pattern = [False]*(2**i)+[True]*(2**i)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
611 experiments[var] = pattern*(2**(nIndependentVariables-i-1))
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
612 experiments = DataFrame(experiments)
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
613 experiments['r2adj'] = 0.
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
614 experiments['condNum'] = NaN
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
615 experiments['shapiroP'] = -1
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
616 experiments['nobs'] = -1
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
617 return experiments
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
618
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
619 def findBestModel(data, dependentVariable, independentVariables, regressionType = 'ols', nProcesses = 1):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
620 '''Generates all possible model with the independentVariables
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
621 and runs them, saving the results in experiments
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
622 with multiprocess option'''
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
623 from pandas import concat
671
849f5f8bf4b9 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 670
diff changeset
624 from multiprocessing import Pool
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
625 experiments = generateExperiments(independentVariables)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
626 nModels = len(experiments)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
627 print("Running {} models with {} processes".format(nModels, nProcesses))
674
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
628 print("IndependentVariables: {}".format(independentVariables))
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
629 if nProcesses == 1:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
630 return runModels(experiments, data, dependentVariable, independentVariables, regressionType)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
631 else:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
632 pool = Pool(processes = nProcesses)
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
633 chunkSize = int(ceil(nModels/nProcesses))
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
634 jobs = [pool.apply_async(runModels, args = (experiments[i*chunkSize:(i+1)*chunkSize], data, dependentVariable, independentVariables, regressionType)) for i in range(nProcesses)]
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
635 return concat([job.get() for job in jobs])
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
636
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
637 def findBestModelFwd(data, dependentVariable, independentVariables, modelFunc, experiments = None):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
638 '''Forward search for best model (based on adjusted R2)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
639 Randomly starting with one variable and adding randomly variables
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
640 if they improve the model
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
641
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
642 The results are added to experiments if provided as argument
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
643 Storing in experiment relies on the index being the number equal
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
644 to the binary code derived from the independent variables'''
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
645 from numpy.random import permutation as nppermutation
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
646 if experiments is None:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
647 experiments = generateExperiments(independentVariables)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
648 nIndependentVariables = len(independentVariables)
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
649 permutation = nppermutation(list(range(nIndependentVariables)))
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
650 variableMapping = {j: independentVariables[i] for i,j in enumerate(permutation)}
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
651 print('Tested variables '+', '.join([variableMapping[i] for i in range(nIndependentVariables)]))
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
652 bestModel = [False]*nIndependentVariables
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
653 currentVarNum = 0
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
654 currentR2Adj = 0.
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
655 for currentVarNum in range(nIndependentVariables):
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
656 currentModel = [i for i in bestModel]
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
657 currentModel[currentVarNum] = True
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
658 rowIdx = sum([0]+[2**i for i in range(nIndependentVariables) if currentModel[permutation[i]]])
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
659 #print currentVarNum, sum(currentModel), ', '.join([independentVariables[i] for i in range(nIndependentVariables) if currentModel[permutation[i]]])
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
660 if experiments.loc[rowIdx, 'shapiroP'] < 0:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
661 modelStr = modelString(experiments.loc[rowIdx], dependentVariable, independentVariables)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
662 model = modelFunc(modelStr, data = data)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
663 results = model.fit()
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
664 experiments.loc[rowIdx, 'r2adj'] = results.rsquared_adj
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
665 experiments.loc[rowIdx, 'condNum'] = results.condition_number
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
666 experiments.loc[rowIdx, 'shapiroP'] = shapiro(results.resid)[1]
668
f8dcf483b296 code to prepare regression variables (remove correlated variables) and record dataset size in experimnets
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 667
diff changeset
667 experiments.loc[rowIdx, 'nobs'] = int(results.nobs)
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
668 if currentR2Adj < experiments.loc[rowIdx, 'r2adj']:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
669 currentR2Adj = experiments.loc[rowIdx, 'r2adj']
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
670 bestModel[currentVarNum] = True
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
671 return experiments
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
672
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
673 def displayModelResults(results, model = None, plotFigures = True, filenamePrefix = None, figureFileType = 'pdf', text = {'title-shapiro': 'Shapiro-Wilk normality test for residuals: {:.2f} (p={:.3f})', 'true-predicted.xlabel': 'Predicted values', 'true-predicted.ylabel': 'True values', 'residuals-predicted.xlabel': 'Predicted values', 'residuals-predicted.ylabel': 'Residuals'}):
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
674 import statsmodels.api as sm
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
675 '''Displays some model results
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
676
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
677 3 graphics, true-predicted, residuals-predicted, '''
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
678 print(results.summary())
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
679 shapiroResult = shapiro(results.resid)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
680 print(shapiroResult)
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
681 if plotFigures:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
682 fig = plt.figure(figsize=(7,6.3*(2+int(model is not None))))
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
683 if model is not None:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
684 ax = fig.add_subplot(3,1,1)
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
685 plt.plot(results.predict(), model.endog, 'x')
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
686 x=plt.xlim()
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
687 y=plt.ylim()
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
688 plt.plot([max(x[0], y[0]), min(x[1], y[1])], [max(x[0], y[0]), min(x[1], y[1])], 'r')
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
689 #plt.axis('equal')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
690 if text is not None:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
691 plt.title(text['title-shapiro'].format(*shapiroResult))
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
692 #plt.title(text['true-predicted.title'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
693 plt.xlabel(text['true-predicted.xlabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
694 plt.ylabel(text['true-predicted.ylabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
695 fig.add_subplot(3,1,2, sharex = ax)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
696 plt.plot(results.predict(), results.resid, 'x')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
697 nextSubplotNum = 3
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
698 else:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
699 fig.add_subplot(2,1,1)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
700 plt.plot(results.predict(), results.resid, 'x')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
701 nextSubplotNum = 2
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
702 if text is not None:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
703 if model is None:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
704 plt.title(text['title-shapiro'].format(*shapiroResult))
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
705 plt.xlabel(text['residuals-predicted.xlabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
706 plt.ylabel(text['residuals-predicted.ylabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
707 qqAx = fig.add_subplot(nextSubplotNum,1,nextSubplotNum)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
708 sm.qqplot(results.resid, fit = True, line = '45', ax = qqAx)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
709 plt.axis('equal')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
710 if text is not None and 'qqplot.xlabel' in text:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
711 plt.xlabel(text['qqplot.xlabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
712 plt.ylabel(text['qqplot.ylabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
713 plt.tight_layout()
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
714 if filenamePrefix is not None:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
715 from storage import openCheck
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
716 out = openCheck(filenamePrefix+'-coefficients.html', 'w')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
717 out.write(results.summary().as_html())
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
718 plt.savefig(filenamePrefix+'-model-results.'+figureFileType)
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
719
27
44689029a86f updated segmentIntersection and other
Nicolas Saunier <nico@confins.net>
parents: 24
diff changeset
720 #########################
455
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
721 # iterable section
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
722 #########################
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
723
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
724 def mostCommon(L):
456
825e5d49325d slight update
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 455
diff changeset
725 '''Returns the most frequent element in a iterable
825e5d49325d slight update
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 455
diff changeset
726
825e5d49325d slight update
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 455
diff changeset
727 taken from http://stackoverflow.com/questions/1518522/python-most-common-element-in-a-list'''
455
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
728 from itertools import groupby
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
729 from operator import itemgetter
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
730 # get an iterable of (item, iterable) pairs
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
731 SL = sorted((x, i) for i, x in enumerate(L))
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
732 # print 'SL:', SL
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
733 groups = groupby(SL, key=itemgetter(0))
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
734 # auxiliary function to get "quality" for an item
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
735 def _auxfun(g):
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
736 item, iterable = g
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
737 count = 0
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
738 min_index = len(L)
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
739 for _, where in iterable:
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
740 count += 1
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
741 min_index = min(min_index, where)
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
742 # print 'item %r, count %r, minind %r' % (item, count, min_index)
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
743 return count, -min_index
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
744 # pick the highest-count/earliest item
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
745 return max(groups, key=_auxfun)[0]
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
746
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
747 #########################
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
748 # sequence section
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
749 #########################
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
750
665
15e244d2a1b5 corrected bug with circular import for VideoFilenameAddable, moved to base module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 659
diff changeset
751 class LCSS(object):
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
752 '''Class that keeps the LCSS parameters
686
cdee6a3a47b4 allowing alternate database and filename for classify-objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 680
diff changeset
753 and puts together the various computations
cdee6a3a47b4 allowing alternate database and filename for classify-objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 680
diff changeset
754
cdee6a3a47b4 allowing alternate database and filename for classify-objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 680
diff changeset
755 the methods with names starting with _ are not to be shadowed
cdee6a3a47b4 allowing alternate database and filename for classify-objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 680
diff changeset
756 in child classes, who will shadow the other methods,
cdee6a3a47b4 allowing alternate database and filename for classify-objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 680
diff changeset
757 ie compute and computeXX methods'''
689
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
758 def __init__(self, similarityFunc = None, metric = None, epsilon = None, delta = float('inf'), aligned = False, lengthFunc = min):
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
759 '''One should provide either a similarity function
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
760 that indicates (return bool) whether elements in the compares lists are similar
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
761
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
762 eg distance(p1, p2) < epsilon
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
763
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
764 or a type of metric usable in scipy.spatial.distance.cdist with an epsilon'''
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
765 if similarityFunc is None and metric is None:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
766 print("No way to compute LCSS, similarityFunc and metric are None. Exiting")
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
767 import sys
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
768 sys.exit()
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
769 elif metric is not None and epsilon is None:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
770 print("Please provide a value for epsilon if using a cdist metric. Exiting")
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
771 import sys
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
772 sys.exit()
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
773 else:
741
5b91b8d97cf3 corrected bug
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 739
diff changeset
774 if similarityFunc is None and metric is not None and not isinf(delta):
737
fb60b54e1041 added warning for finite delta
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 733
diff changeset
775 print('Warning: you are using a cdist metric and a finite delta, which will make probably computation slower than using the equivalent similarityFunc (since all pairwise distances will be computed by cdist).')
689
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
776 self.similarityFunc = similarityFunc
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
777 self.metric = metric
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
778 self.epsilon = epsilon
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
779 self.aligned = aligned
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
780 self.delta = delta
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
781 self.lengthFunc = lengthFunc
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
782 self.subSequenceIndices = [(0,0)]
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
783
373
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
784 def similarities(self, l1, l2, jshift=0):
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
785 n1 = len(l1)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
786 n2 = len(l2)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
787 self.similarityTable = zeros((n1+1,n2+1), dtype = npint)
689
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
788 if self.similarityFunc is not None:
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
789 for i in range(1,n1+1):
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
790 for j in range(max(1,i-jshift-self.delta),min(n2,i-jshift+self.delta)+1):
689
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
791 if self.similarityFunc(l1[i-1], l2[j-1]):
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
792 self.similarityTable[i,j] = self.similarityTable[i-1,j-1]+1
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
793 else:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
794 self.similarityTable[i,j] = max(self.similarityTable[i-1,j], self.similarityTable[i,j-1])
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
795 elif self.metric is not None:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
796 similarElements = distance.cdist(l1, l2, self.metric) <= self.epsilon
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
797 for i in range(1,n1+1):
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
798 for j in range(max(1,i-jshift-self.delta),min(n2,i-jshift+self.delta)+1):
689
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
799 if similarElements[i-1, j-1]:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
800 self.similarityTable[i,j] = self.similarityTable[i-1,j-1]+1
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
801 else:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
802 self.similarityTable[i,j] = max(self.similarityTable[i-1,j], self.similarityTable[i,j-1])
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
803
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
804
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
805 def subSequence(self, i, j):
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
806 '''Returns the subsequence of two sequences
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
807 http://en.wikipedia.org/wiki/Longest_common_subsequence_problem'''
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
808 if i == 0 or j == 0:
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
809 return []
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
810 elif self.similarityTable[i][j] == self.similarityTable[i][j-1]:
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
811 return self.subSequence(i, j-1)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
812 elif self.similarityTable[i][j] == self.similarityTable[i-1][j]:
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
813 return self.subSequence(i-1, j)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
814 else:
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
815 return self.subSequence(i-1, j-1) + [(i-1,j-1)]
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
816
373
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
817 def _compute(self, _l1, _l2, computeSubSequence = False):
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
818 '''returns the longest common subsequence similarity
689
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
819 l1 and l2 should be the right format
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
820 eg list of tuple points for cdist
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
821 or elements that can be compare using similarityFunc
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
822
607
84690dfe5560 add some functions for behaviour analysis
MohamedGomaa
parents: 574
diff changeset
823 if aligned, returns the best matching if using a finite delta by shifting the series alignments
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
824 '''
372
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
825 if len(_l2) < len(_l1): # l1 is the shortest
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
826 l1 = _l2
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
827 l2 = _l1
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
828 revertIndices = True
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
829 else:
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
830 l1 = _l1
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
831 l2 = _l2
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
832 revertIndices = False
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
833 n1 = len(l1)
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
834 n2 = len(l2)
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
835
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
836 if self.aligned:
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
837 lcssValues = {}
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
838 similarityTables = {}
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
839 for i in range(-n2-self.delta+1, n1+self.delta): # interval such that [i-shift-delta, i-shift+delta] is never empty, which happens when i-shift+delta < 1 or when i-shift-delta > n2
373
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
840 self.similarities(l1, l2, i)
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
841 lcssValues[i] = self.similarityTable.max()
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
842 similarityTables[i] = self.similarityTable
374
a7af3519687e finished implementation of aligned LCSS with matching sequence decoded
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 373
diff changeset
843 #print self.similarityTable
521
3707eeb20f25 changed argMaxDict name to argmaxDict
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 518
diff changeset
844 alignmentShift = argmaxDict(lcssValues) # ideally get the medium alignment shift, the one that minimizes distance
389
6d26dcc7bba0 modifications to compute alignment for None indicators
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 381
diff changeset
845 self.similarityTable = similarityTables[alignmentShift]
372
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
846 else:
389
6d26dcc7bba0 modifications to compute alignment for None indicators
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 381
diff changeset
847 alignmentShift = 0
372
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
848 self.similarities(l1, l2)
373
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
849
374
a7af3519687e finished implementation of aligned LCSS with matching sequence decoded
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 373
diff changeset
850 # threshold values for the useful part of the similarity table are n2-n1-delta and n1-n2-delta
389
6d26dcc7bba0 modifications to compute alignment for None indicators
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 381
diff changeset
851 self.similarityTable = self.similarityTable[:min(n1, n2+alignmentShift+self.delta)+1, :min(n2, n1-alignmentShift+self.delta)+1]
373
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
852
372
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
853 if computeSubSequence:
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
854 self.subSequenceIndices = self.subSequence(self.similarityTable.shape[0]-1, self.similarityTable.shape[1]-1)
371
924e38c9f70e work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 370
diff changeset
855 if revertIndices:
374
a7af3519687e finished implementation of aligned LCSS with matching sequence decoded
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 373
diff changeset
856 self.subSequenceIndices = [(j,i) for i,j in self.subSequenceIndices]
372
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
857 return self.similarityTable[-1,-1]
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
858
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
859 def compute(self, l1, l2, computeSubSequence = False):
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
860 '''get methods are to be shadowed in child classes '''
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
861 return self._compute(l1, l2, computeSubSequence)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
862
375
2ea8584aa80a making indicator LCSS work
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 374
diff changeset
863 def computeAlignment(self):
374
a7af3519687e finished implementation of aligned LCSS with matching sequence decoded
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 373
diff changeset
864 return mean([j-i for i,j in self.subSequenceIndices])
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
865
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
866 def _computeNormalized(self, l1, l2, computeSubSequence = False):
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
867 ''' compute the normalized LCSS
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
868 ie, the LCSS divided by the min or mean of the indicator lengths (using lengthFunc)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
869 lengthFunc = lambda x,y:float(x,y)/2'''
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
870 return float(self._compute(l1, l2, computeSubSequence))/self.lengthFunc(len(l1), len(l2))
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
871
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
872 def computeNormalized(self, l1, l2, computeSubSequence = False):
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
873 return self._computeNormalized(l1, l2, computeSubSequence)
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
874
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
875 def _computeDistance(self, l1, l2, computeSubSequence = False):
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
876 ''' compute the LCSS distance'''
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
877 return 1-self._computeNormalized(l1, l2, computeSubSequence)
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
878
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
879 def computeDistance(self, l1, l2, computeSubSequence = False):
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
880 return self._computeDistance(l1, l2, computeSubSequence)
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
881
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
882 #########################
45
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
883 # plotting section
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
884 #########################
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
885
940
d8ab183a7351 verified motion prediction with prototypes at constant speed (test needed)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 876
diff changeset
886 def plotPolygon(poly, options = '', **kwargs):
332
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
887 'Plots shapely polygon poly'
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
888 from matplotlib.pyplot import plot
940
d8ab183a7351 verified motion prediction with prototypes at constant speed (test needed)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 876
diff changeset
889 x,y = poly.exterior.xy
d8ab183a7351 verified motion prediction with prototypes at constant speed (test needed)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 876
diff changeset
890 plot(x, y, options, **kwargs)
332
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
891
324
99ca91a46007 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 323
diff changeset
892 def stepPlot(X, firstX, lastX, initialCount = 0, increment = 1):
99ca91a46007 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 323
diff changeset
893 '''for each value in X, increment by increment the initial count
297
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
894 returns the lists that can be plotted
324
99ca91a46007 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 323
diff changeset
895 to obtain a step plot increasing by one for each value in x, from first to last value
99ca91a46007 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 323
diff changeset
896 firstX and lastX should be respectively smaller and larger than all elements in X'''
297
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
897
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
898 sortedX = []
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
899 counts = [initialCount]
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
900 for x in sorted(X):
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
901 sortedX += [x,x]
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
902 counts.append(counts[-1])
324
99ca91a46007 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 323
diff changeset
903 counts.append(counts[-1]+increment)
297
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
904 counts.append(counts[-1])
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
905 return [firstX]+sortedX+[lastX], counts
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
906
665
15e244d2a1b5 corrected bug with circular import for VideoFilenameAddable, moved to base module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 659
diff changeset
907 class PlottingPropertyValues(object):
45
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
908 def __init__(self, values):
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
909 self.values = values
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
910
116
2bf5b76320c0 moved intersection plotting and added markers for scatter plots
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 115
diff changeset
911 def __getitem__(self, i):
45
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
912 return self.values[i%len(self.values)]
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
913
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
914 markers = PlottingPropertyValues(['+', '*', ',', '.', 'x', 'D', 's', 'o'])
116
2bf5b76320c0 moved intersection plotting and added markers for scatter plots
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 115
diff changeset
915 scatterMarkers = PlottingPropertyValues(['s','o','^','>','v','<','d','p','h','8','+','x'])
45
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
916
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
917 linestyles = PlottingPropertyValues(['-', '--', '-.', ':'])
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
918
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
919 colors = PlottingPropertyValues('brgmyck') # 'w'
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
920
990
94bee7b604eb addition
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 978
diff changeset
921 def monochromeCycler(withMarker = False):
94bee7b604eb addition
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 978
diff changeset
922 from cycler import cycler
94bee7b604eb addition
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 978
diff changeset
923 if withMarker:
94bee7b604eb addition
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 978
diff changeset
924 monochrome = (cycler('color', ['k']) * cycler('linestyle', ['-', '--', ':', '-.']) * cycler('marker', ['^',',', '.']))
94bee7b604eb addition
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 978
diff changeset
925 else:
94bee7b604eb addition
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 978
diff changeset
926 monochrome = (cycler('color', ['k']) * cycler('linestyle', ['-', '--', ':', '-.']))
94bee7b604eb addition
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 978
diff changeset
927 plt.rc('axes', prop_cycle=monochrome)
94bee7b604eb addition
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 978
diff changeset
928
115
550556378466 added functionalities to indicator maps
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 86
diff changeset
929 def plotIndicatorMap(indicatorMap, squareSize, masked = True, defaultValue=-1):
65
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
930 from matplotlib.pyplot import pcolor
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
931 coords = array(list(indicatorMap.keys()))
65
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
932 minX = min(coords[:,0])
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
933 minY = min(coords[:,1])
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
934 X = arange(minX, max(coords[:,0])+1.1)*squareSize
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
935 Y = arange(minY, max(coords[:,1])+1.1)*squareSize
115
550556378466 added functionalities to indicator maps
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 86
diff changeset
936 C = defaultValue*ones((len(Y), len(X)))
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
937 for k,v in indicatorMap.items():
65
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
938 C[k[1]-minY,k[0]-minX] = v
115
550556378466 added functionalities to indicator maps
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 86
diff changeset
939 if masked:
550556378466 added functionalities to indicator maps
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 86
diff changeset
940 pcolor(X, Y, ma.masked_where(C==defaultValue,C))
550556378466 added functionalities to indicator maps
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 86
diff changeset
941 else:
550556378466 added functionalities to indicator maps
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 86
diff changeset
942 pcolor(X, Y, C)
65
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
943
45
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
944 #########################
637
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
945 # Data download
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
946 #########################
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
947
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
948 def downloadECWeather(stationID, years, months = [], outputDirectoryname = '.', english = True):
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
949 '''Downloads monthly weather data from Environment Canada
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
950 If month is provided (number 1 to 12), it means hourly data for the whole month
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
951 Otherwise, means the data for each day, for the whole year
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
952
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
953 Example: MONTREAL MCTAVISH 10761
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
954 MONTREALPIERRE ELLIOTT TRUDEAU INTL A 5415
856
e310577cc0b8 updated function (url) for weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 855
diff changeset
955 see ftp://client_climate@ftp.tor.ec.gc.ca/Pub/Get_More_Data_Plus_de_donnees/Station%20Inventory%20EN.csv
637
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
956
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
957 To get daily data for 2010 and 2011, downloadECWeather(10761, [2010,2011], [], '/tmp')
973
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 971
diff changeset
958 To get hourly data for 2009 and 2012, January, March and October, downloadECWeather(10761, [2009,2012], [1,3,10], '/tmp')
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 971
diff changeset
959
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 971
diff changeset
960 for annee in `seq 2016 2017`;do wget --content-disposition "http://climat.meteo.gc.ca/climate_data/bulk_data_f.html?format=csv&stationID=10761&Year=${annee}&timeframe=2&submit=++T%C3%A9l%C3%A9charger+%0D%0Ades+donn%C3%A9es" ;done
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 971
diff changeset
961 for annee in `seq 2016 2017`;do for mois in `seq 1 12`;do wget --content-disposition "http://climat.meteo.gc.ca/climate_data/bulk_data_f.html?format=csv&stationID=10761&Year=${annee}&Month=${mois}&timeframe=1&submit=++T%C3%A9l%C3%A9charger+%0D%0Ades+donn%C3%A9es" ;done;done
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 971
diff changeset
962 '''
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
963 import urllib.request
637
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
964 if english:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
965 language = 'e'
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
966 else:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
967 language = 'f'
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
968 if len(months) == 0:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
969 timeFrame = 2
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
970 months = [1]
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
971 else:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
972 timeFrame = 1
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
973
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
974 for year in years:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
975 for month in months:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
976 outFilename = '{}/{}-{}'.format(outputDirectoryname, stationID, year)
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
977 if timeFrame == 1:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
978 outFilename += '-{}-hourly'.format(month)
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
979 else:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
980 outFilename += '-daily'
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
981 outFilename += '.csv'
997
4f3387a242a1 updated utils to python 3
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 990
diff changeset
982 url = urllib.request.urlretrieve('http://climate.weather.gc.ca/climate_data/bulk_data_{}.html?format=csv&stationID={}&Year={}&Month={}&Day=1&timeframe={}&submit=Download+Data'.format(language, stationID, year, month, timeFrame), outFilename)
637
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
983
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
984 #########################
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
985 # File I/O
27
44689029a86f updated segmentIntersection and other
Nicolas Saunier <nico@confins.net>
parents: 24
diff changeset
986 #########################
24
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
987
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
988 def removeExtension(filename, delimiter = '.'):
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
989 '''Returns the filename minus the extension (all characters after last .)'''
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
990 i = filename.rfind(delimiter)
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
991 if i>0:
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
992 return filename[:i]
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
993 else:
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
994 return filename
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
995
969
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
996 def getExtension(filename, delimiter = '.'):
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
997 '''Returns the filename minus the extension (all characters after last .)'''
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
998 i = filename.rfind(delimiter)
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
999 if i>0:
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
1000 return filename[i+1:]
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
1001 else:
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
1002 return ''
5d788d2e8ffc work in progress
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 940
diff changeset
1003
46
b5d007612e16 added filename util
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 45
diff changeset
1004 def cleanFilename(s):
b5d007612e16 added filename util
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 45
diff changeset
1005 'cleans filenames obtained when contatenating figure characteristics'
739
25e78d756823 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 737
diff changeset
1006 return s.replace(' ','-').replace('.','').replace('/','-').replace(',','')
46
b5d007612e16 added filename util
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 45
diff changeset
1007
1021
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1008 def getRelativeFilename(parentPath, filename):
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1009 'Returns filename if absolute, otherwise parentPath/filename as string'
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1010 filePath = Path(filename)
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1011 if filePath.is_absolute():
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1012 return filename
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1013 else:
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1014 return str(parentPath/filePath)
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1015
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
1016 def listfiles(dirname, extension, remove = False):
14
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
1017 '''Returns the list of files with the extension in the directory dirname
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
1018 If remove is True, the filenames are stripped from the extension'''
1021
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1019 d = Path(dirname)
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1020 if d.is_dir():
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1021 tmp = [str(f) for f in d.glob('*.extension')]
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1022 if remove:
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1023 return [removeExtension(f, extension) for f in tmp]
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1024 else:
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1025 return tmp
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
1026 else:
1021
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1027 print(dirname+' is not a directory')
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1028 return []
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
1029
266
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
1030 def mkdir(dirname):
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
1031 'Creates a directory if it does not exist'
1021
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1032 p = Path(dirname)
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1033 if not p.exists():
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1034 p.mkdir()
266
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
1035 else:
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
1036 print(dirname+' already exists')
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
1037
14
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
1038 def removeFile(filename):
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
1039 '''Deletes the file while avoiding raising an error
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
1040 if the file does not exist'''
1021
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1041 f = Path(filename)
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1042 if (f.exists()):
16932cefabc1 work on paths in line with new configurations from tracker
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 997
diff changeset
1043 f.unlink()
266
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
1044 else:
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
1045 print(filename+' does not exist')
14
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
1046
42
1a2ac2d4f53a added loading of the rest of the data for objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 35
diff changeset
1047 def line2Floats(l, separator=' '):
1a2ac2d4f53a added loading of the rest of the data for objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 35
diff changeset
1048 '''Returns the list of floats corresponding to the string'''
1a2ac2d4f53a added loading of the rest of the data for objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 35
diff changeset
1049 return [float(x) for x in l.split(separator)]
1a2ac2d4f53a added loading of the rest of the data for objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 35
diff changeset
1050
1a2ac2d4f53a added loading of the rest of the data for objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 35
diff changeset
1051 def line2Ints(l, separator=' '):
1a2ac2d4f53a added loading of the rest of the data for objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 35
diff changeset
1052 '''Returns the list of ints corresponding to the string'''
1a2ac2d4f53a added loading of the rest of the data for objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 35
diff changeset
1053 return [int(x) for x in l.split(separator)]
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
1054
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
1055 #########################
553
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1056 # Profiling
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1057 #########################
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1058
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1059 def analyzeProfile(profileFilename, stripDirs = True):
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1060 '''Analyze the file produced by cProfile
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1061
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1062 obtained by for example:
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1063 - call in script (for main() function in script)
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1064 import cProfile, os
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1065 cProfile.run('main()', os.path.join(os.getcwd(),'main.profile'))
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1066
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1067 - or on the command line:
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1068 python -m cProfile [-o profile.bin] [-s sort] scriptfile [arg]'''
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1069 import pstats, os
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1070 p = pstats.Stats(os.path.join(os.pardir, profileFilename))
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1071 if stripDirs:
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1072 p.strip_dirs()
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1073 p.sort_stats('time')
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1074 p.print_stats(.2)
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1075 #p.sort_stats('time')
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1076 # p.print_callees(.1, 'int_prediction.py:')
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1077 return p
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1078
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
1079 #########################
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
1080 # running tests
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
1081 #########################
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
1082
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
1083 if __name__ == "__main__":
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
1084 import doctest
2
de5642925615 started implementation of TimeInterval and Spatio-temporal object
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 0
diff changeset
1085 import unittest
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
1086 suite = doctest.DocFileSuite('tests/utils.txt')
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
1087 #suite = doctest.DocTestSuite()
2
de5642925615 started implementation of TimeInterval and Spatio-temporal object
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 0
diff changeset
1088 unittest.TextTestRunner().run(suite)
de5642925615 started implementation of TimeInterval and Spatio-temporal object
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 0
diff changeset
1089 #doctest.testmod()
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
1090 #doctest.testfile("example.txt")