Mercurial Hosting > traffic-intelligence

annotate python/utils.py @ 742:fe71639f1ee7 dev

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
merge and added function to up-/lower-case strings
author Nicolas Saunier <nicolas.saunier@polymtl.ca>
date Fri, 28 Aug 2015 10:39:15 -0400
parents 5b91b8d97cf3
children 10dbab1e871d
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
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
7 from math import sqrt, ceil, floor
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
8 from scipy.stats import kruskal, shapiro
689
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
9 from scipy.spatial import distance
741
5b91b8d97cf3 corrected bug
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 739
diff changeset
10 from numpy import zeros, array, exp, sum as npsum, int as npint, arange, cumsum, median, isnan, ones, convolve, dtype, isnan, NaN, mean, ma, isinf
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
11
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
12
421
4fce27946c60 first example of video metadata using sqlalchemy
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 405
diff changeset
13 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
14
185
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
15 #########################
742
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
16 # Strings
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
17 #########################
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
18
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
19 def upperCaseFirstLetter(s):
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
20 words = s.split(' ')
fe71639f1ee7 merge and added function to up-/lower-case strings
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 741
diff changeset
21 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
22 return ' '.join(lowerWords)
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 #########################
185
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
25 # Enumerations
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
26 #########################
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
27
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
28 def inverseEnumeration(l):
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
29 '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
30 result = {}
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
31 for i,x in enumerate(l):
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
32 result[x] = i
c06379f25ab8 utilities for user types
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 181
diff changeset
33 return result
155
f03fe3d6d0c8 added functions to parse options
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 152
diff changeset
34
f03fe3d6d0c8 added functions to parse options
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 152
diff changeset
35 #########################
637
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
36 # Simple statistics
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
37 #########################
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
38
680
da1352b89d02 classification is working
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 677
diff changeset
39 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
40 '''location and scale are respectively the mean and standard deviation of the normal in the log-normal distribution
de278c5e65f6 minor comments for lognormal parameters (numpy and usual names differ)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 686
diff changeset
41 https://en.wikipedia.org/wiki/Log-normal_distribution'''
680
da1352b89d02 classification is working
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 677
diff changeset
42 mean = exp(loc+(scale**2)/2)
da1352b89d02 classification is working
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 677
diff changeset
43 var = (exp(loc**2)-1)*exp(2*loc+scale**2)
da1352b89d02 classification is working
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 677
diff changeset
44 return mean, var
da1352b89d02 classification is working
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 677
diff changeset
45
423
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
46 def sampleSize(stdev, tolerance, percentConfidence, printLatex = False):
301
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
47 from scipy.stats.distributions import norm
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
48 k = round(norm.ppf(0.5+percentConfidence/200., 0, 1)*100)/100. # 1.-(100-percentConfidence)/200.
423
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
49 if printLatex:
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
50 print('${0}^2\\frac{{{1}^2}}{{{2}^2}}$'.format(k, stdev, tolerance))
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
51 return (k*stdev/tolerance)**2
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
52
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
53 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
54 '''if trueStd, use normal distribution, otherwise, Student
0a93afea8243 alternative confidence interval
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 491
diff changeset
55
0a93afea8243 alternative confidence interval
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 491
diff changeset
56 Use otherwise t.interval or norm.interval
0a93afea8243 alternative confidence interval
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 491
diff changeset
57 ex: norm.interval(0.95, loc = 0., scale = 2.3/sqrt(11))
0a93afea8243 alternative confidence interval
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 491
diff changeset
58 t.interval(0.95, 10, loc=1.2, scale = 2.3/sqrt(nSamples))
0a93afea8243 alternative confidence interval
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 491
diff changeset
59 loc is mean, scale is sigma/sqrt(n) (for Student, 10 is df)'''
423
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
60 from scipy.stats.distributions import norm, t
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
61 if trueStd:
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
62 k = round(norm.ppf(0.5+percentConfidence/200., 0, 1)*100)/100. # 1.-(100-percentConfidence)/200.
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
63 else: # use Student
f738fa1b69f0 added sample size and Student distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 421
diff changeset
64 k = round(t.ppf(0.5+percentConfidence/200., nSamples-1)*100)/100.
301
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
65 e = k*stdev/sqrt(nSamples)
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
66 if printLatex:
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
67 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
68 return mean-e, mean+e
27f06d28036d added simple helper for confidence intervals
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 297
diff changeset
69
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
70 def computeChi2(expected, observed):
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
71 '''Returns the Chi2 statistics'''
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
72 result = 0.
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
73 for e, o in zip(expected, observed):
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
74 result += ((e-o)*(e-o))/e
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
75 return result
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
76
665
15e244d2a1b5 corrected bug with circular import for VideoFilenameAddable, moved to base module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 659
diff changeset
77 class EmpiricalDistribution(object):
85
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
78 def nSamples(self):
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
79 return sum(self.counts)
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
80
588
c5406edbcf12 added loading ground truth annotations (ground truth) from polytrack format
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 574
diff changeset
81 def cumulativeDensityFunction(sample, normalized = False):
276
78922b4de3bf minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 266
diff changeset
82 '''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
83 xaxis = sorted(sample)
c5406edbcf12 added loading ground truth annotations (ground truth) from polytrack format
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 574
diff changeset
84 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
85 if normalized:
c5406edbcf12 added loading ground truth annotations (ground truth) from polytrack format
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 574
diff changeset
86 counts /= float(len(sample))
197
2788b2827670 simple cumulatie function distribution computation
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 185
diff changeset
87 return xaxis, counts
85
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
88
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
89 class EmpiricalDiscreteDistribution(EmpiricalDistribution):
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
90 '''Class to represent a sample of a distribution for a discrete random variable
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
91 '''
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
92 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
93 self.categories = categories
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
94 self.counts = counts
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
95
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
96 def mean(self):
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
97 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
98 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
99
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
100 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
101 if not mean:
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
102 m = self.mean()
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
103 else:
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
104 m = mean
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
105 result = 0.
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
106 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
107 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
108
86
f03ec4697a09 corrected discrete distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 85
diff changeset
109 def referenceCounts(self, probability):
f03ec4697a09 corrected discrete distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 85
diff changeset
110 '''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
111 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
112 refProba[-1] = 1-npsum(refProba[:-1])
86
f03ec4697a09 corrected discrete distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 85
diff changeset
113 refCounts = [r*self.nSamples() for r in refProba]
f03ec4697a09 corrected discrete distribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 85
diff changeset
114 return refCounts, refProba
85
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
115
7f1e54234f96 added empirical discrete distribution, modified class organization and names
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 77
diff changeset
116 class EmpiricalContinuousDistribution(EmpiricalDistribution):
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
117 '''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
118 with the number of observations for each interval
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
119 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
120 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
121 def __init__(self, categories, counts):
276
78922b4de3bf minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 266
diff changeset
122 # 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
123 self.categories = categories
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
124 self.counts = counts
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
125
35
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
126 def mean(self):
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
127 result = 0.
76
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
128 for i in range(len(self.counts)-1):
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
129 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
130 return result/self.nSamples()
35
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
131
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
132 def var(self, mean = None):
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
133 if not mean:
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
134 m = self.mean()
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
135 else:
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
136 m = mean
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
137 result = 0.
76
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
138 for i in range(len(self.counts)-1):
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
139 mid = (self.categories[i]+self.categories[i+1])/2
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
140 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
141 return result/(self.nSamples()-1)
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
142
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
143 def referenceCounts(self, cdf):
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
144 '''cdf is a cumulative distribution function
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
145 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
146 # refCumulativeCounts = [0]#[cdf(self.categories[0][0])]
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
147 # for inter in self.categories:
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
148 # refCumulativeCounts.append(cdf(inter[1]))
76
64fde2b1f96d simplified intervales in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 75
diff changeset
149 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
150
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
151 refProba = [refCumulativeCounts[0]]
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
152 for i in xrange(1,len(refCumulativeCounts)):
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
153 refProba.append(refCumulativeCounts[i]-refCumulativeCounts[i-1])
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
154 refProba.append(1-refCumulativeCounts[-1])
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
155 refCounts = [p*self.nSamples() for p in refProba]
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
156
8cafee54466f forgotten update of histogram class
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 32
diff changeset
157 return refCounts, refProba
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
158
77
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
159 def printReferenceCounts(self, refCounts=None):
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
160 if refCounts:
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
161 ref = refCounts
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
162 else:
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
163 ref = self.referenceCounts
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
164 for i in xrange(len(ref[0])):
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
165 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
166
5e6cd36a991c added pretty print in empiricalDistribution
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 76
diff changeset
167
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
168 #########################
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
169 # maths section
27
44689029a86f updated segmentIntersection and other
Nicolas Saunier <nico@confins.net>
parents: 24
diff changeset
170 #########################
24
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
171
433
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
172 # def kernelSmoothing(sampleX, X, Y, weightFunc, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
173 # '''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
174 # Sum_x weight(sample_x,x) * y(x)'''
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
175 # from numpy import zeros, array
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
176 # smoothed = zeros(len(sampleX))
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
177 # for i,x in enumerate(sampleX):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
178 # weights = array([weightFunc(x,xx, halfwidth) for xx in X])
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
179 # if sum(weights)>0:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
180 # smoothed[i] = sum(weights*Y)/sum(weights)
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
181 # else:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
182 # smoothed[i] = 0
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
183 # return smoothed
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
184
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
185 def kernelSmoothing(x, X, Y, weightFunc, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
186 '''Returns the smoothed estimate of (X,Y) at x
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
187 Sum_x weight(sample_x,x) * y(x)'''
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
188 weights = array([weightFunc(x,observedx, halfwidth) for observedx in X])
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
189 if sum(weights)>0:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
190 return sum(weights*Y)/sum(weights)
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
191 else:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
192 return 0
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
193
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
194 def uniform(center, x, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
195 if abs(center-x)<halfwidth:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
196 return 1.
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
197 else:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
198 return 0.
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
199
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
200 def gaussian(center, x, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
201 return exp(-((center-x)/halfwidth)**2/2)
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
202
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
203 def epanechnikov(center, x, halfwidth):
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
204 diff = abs(center-x)
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
205 if diff<halfwidth:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
206 return 1.-(diff/halfwidth)**2
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
207 else:
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
208 return 0.
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
209
434
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
210 def triangular(center, x, halfwidth):
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
211 diff = abs(center-x)
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
212 if diff<halfwidth:
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
213 return 1.-abs(diff/halfwidth)
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
214 else:
9a714f32fc9f small updates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 433
diff changeset
215 return 0.
433
d40ad901b272 added kernel smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 423
diff changeset
216
518
0c86c73f3c09 median smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 511
diff changeset
217 def medianSmoothing(x, X, Y, halfwidth):
0c86c73f3c09 median smoothing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 511
diff changeset
218 '''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
219 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
220
521
3707eeb20f25 changed argMaxDict name to argmaxDict
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 518
diff changeset
221 def argmaxDict(d):
561
ee45c6eb6d49 added Mohamed Gomaa Mohamed function to smooth object trajectories
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 553
diff changeset
222 return max(d, key=d.get)
279
3af4c267a7bf generic simple LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 276
diff changeset
223
395
6fba1ab040f1 minor modification to framestotime
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 391
diff changeset
224 def framesToTime(nFrames, frameRate, initialTime = time()):
6fba1ab040f1 minor modification to framestotime
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 391
diff changeset
225 '''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
226 initialTime is a datetime.time'''
6fba1ab040f1 minor modification to framestotime
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 391
diff changeset
227 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
228 h = int(floor(seconds/3600.))
248
571ba5ed22e2 added utils for bus processing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 241
diff changeset
229 seconds = seconds - h*3600
261
4aa792cb0fa9 changing framesToTime to return a datetime.time
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 248
diff changeset
230 m = int(floor(seconds/60))
248
571ba5ed22e2 added utils for bus processing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 241
diff changeset
231 seconds = seconds - m*60
262
a048066bd20f correcting bug in framesToTime
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 261
diff changeset
232 return time(h, m, seconds)
248
571ba5ed22e2 added utils for bus processing
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 241
diff changeset
233
381
387cc0142211 script to replay event annotations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 376
diff changeset
234 def timeToFrames(t, frameRate):
387cc0142211 script to replay event annotations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 376
diff changeset
235 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
236
241
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
237 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
238 '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
239 D = {}
ee1caff48b03 added function to sort to list of paired data X,Y
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 235
diff changeset
240 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
241 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
242 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
243 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
244
733
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
245 def compareLengthForSort(i, j):
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
246 if len(i) < len(j):
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
247 return -1
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
248 elif len(i) == len(j):
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
249 return 0
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
250 else:
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
251 return 1
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
252
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
253 def sortByLength(instances, reverse = False):
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
254 '''Returns a new list with the instances sorted by length (method __len__)
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
255 reverse is passed to sorted'''
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
256 return sorted(instances, cmp = compareLengthForSort, reverse = reverse)
c35e4a4b199d sortbylength
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 698
diff changeset
257
32
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
258 def ceilDecimals(v, nDecimals):
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
259 '''Rounds the number at the nth decimal
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
260 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
261 tens = 10**nDecimals
32
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
262 return ceil(v*tens)/tens
48e56179c39e added ceil function
Nicolas Saunier <nico@confins.net>
parents: 31
diff changeset
263
152
74b1fc68d4df re-organized code to avoid cyclic python module dependencies
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 116
diff changeset
264 def inBetween(bound1, bound2, x):
698
8d99a9e16644 added clarification comments
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 689
diff changeset
265 '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
266 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
267
0057c04f94d5 work in progress on intersections (for PET)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 561
diff changeset
268 def pointDistanceL2(x1,y1,x2,y2):
0057c04f94d5 work in progress on intersections (for PET)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 561
diff changeset
269 ''' 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
270 return sqrt((x2-x1)**2+(y2-y1)**2)
24
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
271
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
272 def crossProduct(l1, l2):
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
273 return l1[0]*l2[1]-l1[1]*l2[0]
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
274
574
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
275 def cat_mvgavg(cat_list, halfWidth):
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
276 ''' 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
277 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
278 from copy import deepcopy
659
784298512b60 minor modifications
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 637
diff changeset
279 smoothed = deepcopy(cat_list)
784298512b60 minor modifications
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 637
diff changeset
280 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
281 lower_bound_check = max(0,point-halfWidth)
659
784298512b60 minor modifications
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 637
diff changeset
282 upper_bound_check = min(len(cat_list)-1,point+halfWidth+1)
784298512b60 minor modifications
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 637
diff changeset
283 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
284 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
285 return smoothed
e24eeb244698 first implementation of projection to curvilinear coordinates
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 569
diff changeset
286
547
97c5fef5b2d6 corrected bugs
MohamedGomaa
parents: 521
diff changeset
287 def filterMovingWindow(inputSignal, halfWidth):
29
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
288 '''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
289 The first and last points are copied from the original.'''
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
290 width = float(halfWidth*2+1)
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
291 win = ones(width,'d')
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
292 result = convolve(win/width,array(inputSignal),'same')
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
293 result[:halfWidth] = inputSignal[:halfWidth]
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
294 result[-halfWidth:] = inputSignal[-halfWidth:]
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
295 return result
ca8e716cc231 added moving average filter
Nicolas Saunier <nico@confins.net>
parents: 27
diff changeset
296
199
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
297 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
298 '''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
299 as well as the plot'''
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
300 from numpy.lib.polynomial import polyfit
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
301 from numpy.core.multiarray import arange
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
302 coef = polyfit(x, y, deg)
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
303 if plotData:
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
304 def poly(x):
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
305 result = 0
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
306 for i in range(len(coef)):
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
307 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
308 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
309 plt.plot(x, y, 'x')
199
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
310 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
311 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
312 return coef
ca9d9104afba added utility to calibrate polynoms and plot
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 197
diff changeset
313
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
314 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
315 '''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
316 columns = data.columns.tolist()
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
317 for var in data.columns:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
318 uniqueValues = data[var].unique()
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
319 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
320 columns.remove(var)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
321 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
322 if plotFigure:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
323 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
324 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
325 #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
326 plt.imshow(c, vmin=-1., vmax = 1., interpolation='none', cmap = 'RdYlBu_r') # coolwarm
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
327 colnames = [displayNames.get(s.strip(), s.strip()) for s in 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
328 #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
329 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
330 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
331 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
332 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
333 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
334 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
335 plt.tight_layout()
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
336 if len(colnames) > 50:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
337 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
338 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
339 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
340 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
341
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
342 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
343 '''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
344 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
345 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
346 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
347 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
348 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
349 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
350 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
351 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
352 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
353 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
354 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
355 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
356 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
357
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
358 def kruskalWallis(data, dependentVariable, independentVariable, plotFigure = False, filenamePrefix = None, figureFileType = 'pdf', saveLatex = False, renameVariables = lambda s: s, kwCaption = u''):
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
359 '''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
360
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
361 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
362 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
363 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
364 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
365 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
366 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
367 if saveLatex:
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
368 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
369 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
370 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
371 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
372 if len(tmp.loc[tmp[independentVariable] == x, dependentVariable]) >= 3:
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
373 print shapiro(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
374 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
375 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
376 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
377 #q25, q75 = tmp[dependentVariable].quantile([.25, .75])
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.ylim(ymax = q75+1.5*(q75-q25))
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
379 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
380 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
381 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
382 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
383 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
384 table['count'] = table['count'].astype(int)
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
385 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
386 if saveLatex:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
387 out.write('\\begin{minipage}{\\linewidth}\n'
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
388 +'\\centering\n'
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
389 +'\\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
390 +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
391 +'\\end{minipage}\n'
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
392 +'\\ \\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
393 else:
01b89182891a corrected bug for intersection of lines (thanks to Paul for finding)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 672
diff changeset
394 print table
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
395 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
396 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
397 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
398
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
399 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
400 '''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
401 - 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
402 - 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
403 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
404
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 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
406 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
407
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
408 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
409 from copy import copy
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
410 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
411 result = copy(independentVariables)
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
412 table1 = ''
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
413 table2 = {}
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
414 # constant variables
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
415 for var in independentVariables:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
416 uniqueValues = data[var].unique()
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
417 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
418 print(stdoutText[0].format(var, uniqueValues))
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
419 if saveFiles:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
420 table1 += latexTable[0].format(var, *uniqueValues)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
421 result.remove(var)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
422 # correlated variables
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
423 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
424 if v1 in correlations.index:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
425 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
426 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
427 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
428 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
429 if saveFiles:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
430 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
431 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
432 result.remove(v2)
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
433 # not correlated with dependent variable
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
434 table2['Correlations'] = []
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
435 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
436 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
437 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
438 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
439 if p > maxCorrelationP:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
440 if saveFiles:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
441 table1 += latexTable[2].format(var, cor, p)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
442 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
443 result.remove(var)
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
444 else:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
445 table2['Correlations'].append(cor)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
446 table2['Valeurs p'].append(p)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
447
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
448 if saveFiles:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
449 from storage import openCheck
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
450 out = openCheck(filenamePrefix+'-removed-variables.tex', 'w')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
451 out.write(latexHeader)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
452 out.write(table1)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
453 out.write(latexFooter)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
454 out.close()
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
455 out = openCheck(filenamePrefix+'-correlations.html', 'w')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
456 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
457 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
458 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
459 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
460
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
461
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
462 #########################
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
463 # regression analysis using statsmodels (and pandas)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
464 #########################
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
465
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
466 # 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
467 # 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
468 def modelString(experiment, dependentVariable, independentVariables):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
469 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
470
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
471 def runModel(experiment, data, dependentVariable, independentVariables, regressionType = 'ols'):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
472 import statsmodels.formula.api as smf
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
473 modelStr = modelString(experiment, dependentVariable, independentVariables)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
474 if regressionType == 'ols':
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
475 model = smf.ols(modelStr, data = data)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
476 elif regressionType == 'gls':
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
477 model = smf.gls(modelStr, data = data)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
478 elif regressionType == 'rlm':
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
479 model = smf.rlm(modelStr, data = data)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
480 else:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
481 print('Unknown regression type {}. Exiting'.format(regressionType))
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
482 import sys
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
483 sys.exit()
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
484 return model.fit()
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
485
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
486 def runModels(experiments, data, dependentVariable, independentVariables, regressionType = 'ols'):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
487 '''Runs several models and stores 3 statistics
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
488 adjusted R2, condition number (should be small, eg < 1000)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
489 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
490 for i,experiment in experiments.iterrows():
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
491 if experiment[independentVariables].any():
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
492 results = runModel(experiment, data, dependentVariable, independentVariables, regressionType = 'ols')
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
493 experiments.loc[i,'r2adj'] = results.rsquared_adj
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
494 experiments.loc[i,'condNum'] = results.condition_number
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
495 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
496 experiments.loc[i,'nobs'] = int(results.nobs)
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
497 return experiments
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
498
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
499 def generateExperiments(independentVariables):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
500 '''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
501 from pandas import DataFrame
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
502 experiments = {}
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
503 nIndependentVariables = len(independentVariables)
669
df6be882f325 bug corrected
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 668
diff changeset
504 if nIndependentVariables != len(set(independentVariables)):
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
505 print("Duplicate variables. Exiting")
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
506 import sys
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
507 sys.exit()
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
508 nModels = 2**nIndependentVariables
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
509 for i,var in enumerate(independentVariables):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
510 pattern = [False]*(2**i)+[True]*(2**i)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
511 experiments[var] = pattern*(2**(nIndependentVariables-i-1))
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
512 experiments = DataFrame(experiments)
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
513 experiments['r2adj'] = 0.
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
514 experiments['condNum'] = NaN
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
515 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
516 experiments['nobs'] = -1
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
517 return experiments
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
518
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
519 def findBestModel(data, dependentVariable, independentVariables, regressionType = 'ols', nProcesses = 1):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
520 '''Generates all possible model with the independentVariables
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
521 and runs them, saving the results in experiments
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
522 with multiprocess option'''
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
523 from pandas import concat
671
849f5f8bf4b9 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 670
diff changeset
524 from multiprocessing import Pool
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
525 experiments = generateExperiments(independentVariables)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
526 nModels = len(experiments)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
527 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
528 print("IndependentVariables: {}".format(independentVariables))
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
529 if nProcesses == 1:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
530 return runModels(experiments, data, dependentVariable, independentVariables, regressionType)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
531 else:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
532 pool = Pool(processes = nProcesses)
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
533 chunkSize = int(ceil(nModels/nProcesses))
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
534 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
535 return concat([job.get() for job in jobs])
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
536
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
537 def findBestModelFwd(data, dependentVariable, independentVariables, modelFunc, experiments = None):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
538 '''Forward search for best model (based on adjusted R2)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
539 Randomly starting with one variable and adding randomly variables
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
540 if they improve the model
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
541
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
542 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
543 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
544 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
545 from numpy.random import permutation as nppermutation
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
546 if experiments is None:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
547 experiments = generateExperiments(independentVariables)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
548 nIndependentVariables = len(independentVariables)
670
f72ed51c6b65 corrected other missing imports
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 669
diff changeset
549 permutation = nppermutation(range(nIndependentVariables)).tolist()
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
550 variableMapping = {j: independentVariables[i] for i,j in enumerate(permutation)}
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
551 print('Tested variables '+', '.join([variableMapping[i] for i in xrange(nIndependentVariables)]))
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
552 bestModel = [False]*nIndependentVariables
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
553 currentVarNum = 0
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
554 currentR2Adj = 0.
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
555 for currentVarNum in xrange(nIndependentVariables):
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
556 currentModel = [i for i in bestModel]
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
557 currentModel[currentVarNum] = True
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
558 rowIdx = sum([0]+[2**i for i in xrange(nIndependentVariables) if currentModel[permutation[i]]])
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
559 #print currentVarNum, sum(currentModel), ', '.join([independentVariables[i] for i in xrange(nIndependentVariables) if currentModel[permutation[i]]])
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
560 if experiments.loc[rowIdx, 'shapiroP'] < 0:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
561 modelStr = modelString(experiments.loc[rowIdx], dependentVariable, independentVariables)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
562 model = modelFunc(modelStr, data = data)
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
563 results = model.fit()
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
564 experiments.loc[rowIdx, 'r2adj'] = results.rsquared_adj
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
565 experiments.loc[rowIdx, 'condNum'] = results.condition_number
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
566 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
567 experiments.loc[rowIdx, 'nobs'] = int(results.nobs)
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
568 if currentR2Adj < experiments.loc[rowIdx, 'r2adj']:
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
569 currentR2Adj = experiments.loc[rowIdx, 'r2adj']
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
570 bestModel[currentVarNum] = True
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
571 return experiments
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
572
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
573 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
574 import statsmodels.api as sm
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
575 '''Displays some model results
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
576
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
577 3 graphics, true-predicted, residuals-predicted, '''
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
578 print(results.summary())
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
579 shapiroResult = shapiro(results.resid)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
580 print(shapiroResult)
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
581 if plotFigures:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
582 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
583 if model is not None:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
584 ax = fig.add_subplot(3,1,1)
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
585 plt.plot(results.predict(), model.endog, 'x')
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
586 x=plt.xlim()
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
587 y=plt.ylim()
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
588 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
589 #plt.axis('equal')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
590 if text is not None:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
591 plt.title(text['title-shapiro'].format(*shapiroResult))
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
592 #plt.title(text['true-predicted.title'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
593 plt.xlabel(text['true-predicted.xlabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
594 plt.ylabel(text['true-predicted.ylabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
595 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
596 plt.plot(results.predict(), results.resid, 'x')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
597 nextSubplotNum = 3
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
598 else:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
599 fig.add_subplot(2,1,1)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
600 plt.plot(results.predict(), results.resid, 'x')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
601 nextSubplotNum = 2
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
602 if text is not None:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
603 if model is None:
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
604 plt.title(text['title-shapiro'].format(*shapiroResult))
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
605 plt.xlabel(text['residuals-predicted.xlabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
606 plt.ylabel(text['residuals-predicted.ylabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
607 qqAx = fig.add_subplot(nextSubplotNum,1,nextSubplotNum)
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
608 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
609 plt.axis('equal')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
610 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
611 plt.xlabel(text['qqplot.xlabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
612 plt.ylabel(text['qqplot.ylabel'])
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
613 plt.tight_layout()
676
58b9ac2f262f fine tuning
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 675
diff changeset
614 if filenamePrefix is not None:
677
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
615 from storage import openCheck
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
616 out = openCheck(filenamePrefix+'-coefficients.html', 'w')
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
617 out.write(results.summary().as_html())
ae07c7b4cf87 update to utils for pavement results
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 676
diff changeset
618 plt.savefig(filenamePrefix+'-model-results.'+figureFileType)
667
179b81faa1f8 added regression analysis functions
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 665
diff changeset
619
27
44689029a86f updated segmentIntersection and other
Nicolas Saunier <nico@confins.net>
parents: 24
diff changeset
620 #########################
455
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
621 # iterable section
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
622 #########################
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
623
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
624 def mostCommon(L):
456
825e5d49325d slight update
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 455
diff changeset
625 '''Returns the most frequent element in a iterable
825e5d49325d slight update
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 455
diff changeset
626
825e5d49325d slight update
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 455
diff changeset
627 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
628 from itertools import groupby
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
629 from operator import itemgetter
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
630 # get an iterable of (item, iterable) pairs
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
631 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
632 # print 'SL:', SL
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
633 groups = groupby(SL, key=itemgetter(0))
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
634 # auxiliary function to get "quality" for an item
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
635 def _auxfun(g):
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
636 item, iterable = g
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
637 count = 0
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
638 min_index = len(L)
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
639 for _, where in iterable:
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
640 count += 1
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
641 min_index = min(min_index, where)
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
642 # 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
643 return count, -min_index
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
644 # pick the highest-count/earliest item
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
645 return max(groups, key=_auxfun)[0]
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
646
abe0b2347d4c added most common utility function
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 434
diff changeset
647 #########################
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
648 # sequence section
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
649 #########################
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
650
665
15e244d2a1b5 corrected bug with circular import for VideoFilenameAddable, moved to base module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 659
diff changeset
651 class LCSS(object):
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
652 '''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
653 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
654
cdee6a3a47b4 allowing alternate database and filename for classify-objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 680
diff changeset
655 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
656 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
657 ie compute and computeXX methods'''
689
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
658 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
659 '''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
660 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
661
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
662 eg distance(p1, p2) < epsilon
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
663
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
664 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
665 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
666 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
667 import sys
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
668 sys.exit()
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
669 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
670 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
671 import sys
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
672 sys.exit()
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
673 else:
741
5b91b8d97cf3 corrected bug
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 739
diff changeset
674 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
675 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
676 self.similarityFunc = similarityFunc
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
677 self.metric = metric
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
678 self.epsilon = epsilon
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
679 self.aligned = aligned
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
680 self.delta = delta
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
681 self.lengthFunc = lengthFunc
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
682 self.subSequenceIndices = [(0,0)]
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
683
373
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
684 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
685 n1 = len(l1)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
686 n2 = len(l2)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
687 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
688 if self.similarityFunc is not None:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
689 for i in xrange(1,n1+1):
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
690 for j in xrange(max(1,i-jshift-self.delta),min(n2,i-jshift+self.delta)+1):
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
691 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
692 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
693 else:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
694 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
695 elif self.metric is not None:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
696 similarElements = distance.cdist(l1, l2, self.metric) <= self.epsilon
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
697 for i in xrange(1,n1+1):
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
698 for j in xrange(max(1,i-jshift-self.delta),min(n2,i-jshift+self.delta)+1):
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
699 if similarElements[i-1, j-1]:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
700 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
701 else:
9990ef119bce added version of LCSS with cdist computations
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 687
diff changeset
702 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
703
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
704
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
705 def subSequence(self, i, j):
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
706 '''Returns the subsequence of two sequences
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
707 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
708 if i == 0 or j == 0:
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
709 return []
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
710 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
711 return self.subSequence(i, j-1)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
712 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
713 return self.subSequence(i-1, j)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
714 else:
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
715 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
716
373
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
717 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
718 '''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
719 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
720 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
721 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
722
607
84690dfe5560 add some functions for behaviour analysis
MohamedGomaa
parents: 574
diff changeset
723 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
724 '''
372
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
725 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
726 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
727 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
728 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
729 else:
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
730 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
731 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
732 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
733 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
734 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
735
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
736 if self.aligned:
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
737 lcssValues = {}
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
738 similarityTables = {}
374
a7af3519687e finished implementation of aligned LCSS with matching sequence decoded
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 373
diff changeset
739 for i in xrange(-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
740 self.similarities(l1, l2, i)
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
741 lcssValues[i] = self.similarityTable.max()
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
742 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
743 #print self.similarityTable
521
3707eeb20f25 changed argMaxDict name to argmaxDict
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 518
diff changeset
744 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
745 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
746 else:
389
6d26dcc7bba0 modifications to compute alignment for None indicators
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 381
diff changeset
747 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
748 self.similarities(l1, l2)
373
d0b86ed50f32 work in progress on LCSS
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 372
diff changeset
749
374
a7af3519687e finished implementation of aligned LCSS with matching sequence decoded
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 373
diff changeset
750 # 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
751 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
752
372
349eb1e09f45 Cleaned the methods/functions indicating if a point is in a polygon
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 371
diff changeset
753 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
754 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
755 if revertIndices:
374
a7af3519687e finished implementation of aligned LCSS with matching sequence decoded
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 373
diff changeset
756 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
757 return self.similarityTable[-1,-1]
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
758
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
759 def compute(self, l1, l2, computeSubSequence = False):
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
760 '''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
761 return self._compute(l1, l2, computeSubSequence)
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
762
375
2ea8584aa80a making indicator LCSS work
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 374
diff changeset
763 def computeAlignment(self):
374
a7af3519687e finished implementation of aligned LCSS with matching sequence decoded
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 373
diff changeset
764 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
765
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
766 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
767 ''' compute the normalized LCSS
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
768 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
769 lengthFunc = lambda x,y:float(x,y)/2'''
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
770 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
771
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
772 def computeNormalized(self, l1, l2, computeSubSequence = False):
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
773 return self._computeNormalized(l1, l2, computeSubSequence)
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
774
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
775 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
776 ''' compute the LCSS distance'''
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
777 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
778
376
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
779 def computeDistance(self, l1, l2, computeSubSequence = False):
2e6b8610bcaa work on indicator similarity
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 375
diff changeset
780 return self._computeDistance(l1, l2, computeSubSequence)
370
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
781
97e8fa0ee9a1 work in progress for complete alignment
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 369
diff changeset
782 #########################
45
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
783 # plotting section
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
784 #########################
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
785
332
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
786 def plotPolygon(poly, options = ''):
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
787 'Plots shapely polygon poly'
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
788 from numpy.core.multiarray import array
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
789 from matplotlib.pyplot import plot
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
790 from shapely.geometry import Polygon
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
791
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
792 tmp = array(poly.exterior)
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
793 plot(tmp[:,0], tmp[:,1], options)
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
794
324
99ca91a46007 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 323
diff changeset
795 def stepPlot(X, firstX, lastX, initialCount = 0, increment = 1):
99ca91a46007 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 323
diff changeset
796 '''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
797 returns the lists that can be plotted
324
99ca91a46007 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 323
diff changeset
798 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
799 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
800
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
801 sortedX = []
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
802 counts = [initialCount]
f6f423e25c7f adding function to generate step plots (for cumulative number of vehicles)
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 286
diff changeset
803 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
804 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
805 counts.append(counts[-1])
324
99ca91a46007 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 323
diff changeset
806 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
807 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
808 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
809
665
15e244d2a1b5 corrected bug with circular import for VideoFilenameAddable, moved to base module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 659
diff changeset
810 class PlottingPropertyValues(object):
45
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
811 def __init__(self, values):
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
812 self.values = values
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
813
116
2bf5b76320c0 moved intersection plotting and added markers for scatter plots
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 115
diff changeset
814 def __getitem__(self, i):
45
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
815 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
816
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
817 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
818 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
819
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
820 linestyles = PlottingPropertyValues(['-', '--', '-.', ':'])
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
821
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
822 colors = PlottingPropertyValues('brgmyck') # 'w'
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
823
115
550556378466 added functionalities to indicator maps
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 86
diff changeset
824 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
825 from matplotlib.pyplot import pcolor
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
826 coords = array(indicatorMap.keys())
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
827 minX = min(coords[:,0])
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
828 minY = min(coords[:,1])
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
829 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
830 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
831 C = defaultValue*ones((len(Y), len(X)))
65
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
832 for k,v in indicatorMap.iteritems():
75cf537b8d88 moved and generalized map making functions to the library
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 48
diff changeset
833 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
834 if masked:
550556378466 added functionalities to indicator maps
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 86
diff changeset
835 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
836 else:
550556378466 added functionalities to indicator maps
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 86
diff changeset
837 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
838
45
74d2de078baf added colors, linestyles and markers to pick from
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 42
diff changeset
839 #########################
637
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
840 # Data download
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
841 #########################
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
842
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
843 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
844 '''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
845 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
846 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
847
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
848 Example: MONTREAL MCTAVISH 10761
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
849 MONTREALPIERRE ELLIOTT TRUDEAU INTL A 5415
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
850
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
851 To get daily data for 2010 and 2011, downloadECWeather(10761, [2010,2011], [], '/tmp')
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
852 To get hourly data for 2009 and 2012, January, March and October, downloadECWeather(10761, [2009,2012], [1,3,10], '/tmp')'''
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
853 import urllib2
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
854 if english:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
855 language = 'e'
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
856 else:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
857 language = 'f'
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
858 if len(months) == 0:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
859 timeFrame = 2
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
860 months = [1]
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
861 else:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
862 timeFrame = 1
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
863
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
864 for year in years:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
865 for month in months:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
866 url = urllib2.urlopen('http://climat.meteo.gc.ca/climateData/bulkdata_{}.html?format=csv&stationID={}&Year={}&Month={}&Day=1&timeframe={}&submit=++T%C3%A9l%C3%A9charger+%0D%0Ades+donn%C3%A9es'.format(language, stationID, year, month, timeFrame))
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
867 data = url.read()
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
868 outFilename = '{}/{}-{}'.format(outputDirectoryname, stationID, year)
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
869 if timeFrame == 1:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
870 outFilename += '-{}-hourly'.format(month)
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
871 else:
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
872 outFilename += '-daily'
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
873 outFilename += '.csv'
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
874 out = open(outFilename, 'w')
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
875 out.write(data)
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
876 out.close()
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
877
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
878 #########################
c9a0b72979fd added function to get canadian public weather data
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 615
diff changeset
879 # File I/O
27
44689029a86f updated segmentIntersection and other
Nicolas Saunier <nico@confins.net>
parents: 24
diff changeset
880 #########################
24
6fb59cfb201e first version of segmentIntersection
Nicolas Saunier <nico@confins.net>
parents: 19
diff changeset
881
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
882 def removeExtension(filename, delimiter = '.'):
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
883 '''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
884 i = filename.rfind(delimiter)
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
885 if i>0:
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
886 return filename[:i]
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
887 else:
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
888 return filename
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
889
46
b5d007612e16 added filename util
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 45
diff changeset
890 def cleanFilename(s):
b5d007612e16 added filename util
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 45
diff changeset
891 'cleans filenames obtained when contatenating figure characteristics'
739
25e78d756823 minor change
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 737
diff changeset
892 return s.replace(' ','-').replace('.','').replace('/','-').replace(',','')
46
b5d007612e16 added filename util
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 45
diff changeset
893
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
894 def listfiles(dirname, extension, remove = False):
14
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
895 '''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
896 If remove is True, the filenames are stripped from the extension'''
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
897 from os import listdir
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
898 tmp = [f for f in listdir(dirname) if f.endswith(extension)]
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
899 tmp.sort()
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
900 if remove:
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
901 return [removeExtension(f, extension) for f in tmp]
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
902 else:
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
903 return tmp
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
904
266
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
905 def mkdir(dirname):
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
906 'Creates a directory if it does not exist'
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
907 import os
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
908 if not os.path.exists(dirname):
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
909 os.mkdir(dirname)
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
910 else:
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
911 print(dirname+' already exists')
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
912
14
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
913 def removeFile(filename):
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
914 '''Deletes the file while avoiding raising an error
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
915 if the file does not exist'''
266
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
916 import os
14
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
917 if (os.path.exists(filename)):
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
918 os.remove(filename)
266
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
919 else:
aba9711b3149 small modificatons and reorganization
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 262
diff changeset
920 print(filename+' does not exist')
14
e7bbe8465591 homography and other utils
Nicolas Saunier <nico@confins.net>
parents: 7
diff changeset
921
42
1a2ac2d4f53a added loading of the rest of the data for objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 35
diff changeset
922 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
923 '''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
924 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
925
1a2ac2d4f53a added loading of the rest of the data for objects
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 35
diff changeset
926 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
927 '''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
928 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
929
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
930 #########################
332
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
931 # CLI utils
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
932 #########################
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
933
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
934 def parseCLIOptions(helpMessage, options, cliArgs, optionalOptions=[]):
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
935 ''' Simple function to handle similar argument parsing
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
936 Returns the dictionary of options and their values
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
937
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
938 * cliArgs are most likely directly sys.argv
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
939 (only the elements after the first one are considered)
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
940
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
941 * options should be a list of strings for getopt options,
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
942 eg ['frame=','correspondences=','video=']
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
943 A value must be provided for each option, or the program quits'''
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
944 import sys, getopt
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
945 from numpy.core.fromnumeric import all
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
946 optionValues, args = getopt.getopt(cliArgs[1:], 'h', ['help']+options+optionalOptions)
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
947 optionValues = dict(optionValues)
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
948
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
949 if '--help' in optionValues.keys() or '-h' in optionValues.keys():
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
950 print(helpMessage+
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
951 '\n - Compulsory options: '+' '.join([opt.replace('=','') for opt in options])+
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
952 '\n - Non-compulsory options: '+' '.join([opt.replace('=','') for opt in optionalOptions]))
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
953 sys.exit()
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
954
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
955 missingArgument = [('--'+opt.replace('=','') in optionValues.keys()) for opt in options]
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
956 if not all(missingArgument):
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
957 print('Missing argument')
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
958 print(optionValues)
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
959 sys.exit()
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
960
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
961 return optionValues
a6ca86107f27 reorganized utils module
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 324
diff changeset
962
397
b36b00dd27c3 added function to read scene metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 395
diff changeset
963
b36b00dd27c3 added function to read scene metadata
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 395
diff changeset
964 #########################
553
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
965 # Profiling
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
966 #########################
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
967
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
968 def analyzeProfile(profileFilename, stripDirs = True):
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
969 '''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
970
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
971 obtained by for example:
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
972 - 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
973 import cProfile, os
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
974 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
975
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
976 - or on the command line:
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
977 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
978 import pstats, os
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
979 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
980 if stripDirs:
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
981 p.strip_dirs()
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
982 p.sort_stats('time')
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
983 p.print_stats(.2)
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
984 #p.sort_stats('time')
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
985 # 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
986 return p
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
987
3622a5653ee9 added basic info and function to profile code
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 547
diff changeset
988 #########################
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
989 # running tests
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
990 #########################
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
991
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
992 if __name__ == "__main__":
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
993 import doctest
2
de5642925615 started implementation of TimeInterval and Spatio-temporal object
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 0
diff changeset
994 import unittest
31
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
995 suite = doctest.DocFileSuite('tests/utils.txt')
c000f37c316d moved tests to independent file, added chi2 computation
Nicolas Saunier <nico@confins.net>
parents: 29
diff changeset
996 #suite = doctest.DocTestSuite()
2
de5642925615 started implementation of TimeInterval and Spatio-temporal object
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 0
diff changeset
997 unittest.TextTestRunner().run(suite)
de5642925615 started implementation of TimeInterval and Spatio-temporal object
Nicolas Saunier <nicolas.saunier@polymtl.ca>
parents: 0
diff changeset
998 #doctest.testmod()
0
aed8eb63cdde initial commit with non-functional python code for NGSIM
Nicolas Saunier <nico@confins.net>
parents:
diff changeset
999 #doctest.testfile("example.txt")