# -*- tcl -*-
# statistics.test --
#    Test cases for the ::math::statistics package
#
# Note:
#    The tests assume tcltest 2.1, in order to compare
#    floating-point results

if {[lsearch [namespace children] ::tcltest] == -1} {
    package require tcltest 2.1
    namespace import ::tcltest::*
} else {
    # Ensure that 2.1 or higher present.

    if {![package vsatisfies [package present tcltest] 2.1]} {
	puts "Aborting tests for math::statistics."
	puts "Requiring tcltest 2.1, have [package present tcltest]"
	return
    }
}

source [file join [file dirname [info script]] statistics.tcl]

puts "math::statistics [package present math::statistics]"


set ::data_uniform  [list 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0]
set ::data_missing  [list 1.0 1.0 1.0 {} 1.0 {} {} 1.0 1.0 1.0 1.0 1.0 1.0]
set ::data_linear   [list 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0]
set ::data_empty    [list {} {} {}]
set ::data_missing2 [list 1.0 2.0 3.0 {} 4.0 5.0 6.0 7.0 8.0 9.0 10.0]

#
# Create and register (in that order!) custom matching procedures
#
proc matchTolerant { expected actual } {
   set match 1
   foreach a $actual e $expected {
      if { abs($e-$a)>0.0001*abs($e) &&
           abs($e-$a)>0.0001*abs($a)     } {
         set match 0
         break
      }
   }
   return $match
}
proc matchTolerant2 { expected actual } {
   set match 1
   foreach a $actual e $expected {
      if { abs($e-$a)>0.025*abs($e) &&
           abs($e-$a)>0.025*abs($a)     } {
         set match 0
         break
      }
   }
   return $match
}
proc matchAlmostZero { expected actual } {
   set match 1
   foreach a $actual {
      if { abs($a)>1.0e-6 } {
         set match 0
         break
      }
   }
   return $match
}
customMatch tolerant   matchTolerant
customMatch tolerant2  matchTolerant2
customMatch almostzero matchAlmostZero

#
# Test cases
#
test "BasicStats-1.0" "Basic statistics - uniform data" -match tolerant -body {
  set all_data [::math::statistics::BasicStats all $::data_uniform]
} -result [list 1.0 1.0 1.0 [llength $::data_uniform] 0.0 0.0]

test "BasicStats-1.1" "Basic statistics - empty data" -match glob -body {
  catch {
     set all_data [::math::statistics::BasicStats all $::data_empty]
  } msg
  set msg
} -result "Too*"

#
# Result must be the same as for 1.0! Hence ::data_empty and ::data_uniform
#
test "BasicStats-1.2" "Basic statistics - missing data" -match tolerant -body {
  set all_data [::math::statistics::BasicStats all $::data_missing]
} -result [list 1.0 1.0 1.0 [llength $::data_uniform] 0.0 0.0]

test "BasicStats-1.3" "Basic statistics - linear data - mean" -match tolerant -body {
  set value [::math::statistics::mean $::data_linear]
} -result 5.5

test "BasicStats-1.3" "Basic statistics - linear data - min" -match tolerant  -body {
  set value [::math::statistics::min $::data_linear]
} -result 1.0

test "BasicStats-1.4" "Basic statistics - linear data - max" -match tolerant  -body {
  set value [::math::statistics::max $::data_linear]
} -result 10.0

test "BasicStats-1.5" "Basic statistics - linear data - number" -match tolerant  -body {
  set value [::math::statistics::number $::data_linear]
} -result 10

test "BasicStats-1.6" "Basic statistics - missing data - number" -match tolerant  -body {
  set value [::math::statistics::number $::data_missing2]
} -result 10

test "BasicStats-1.7" "Basic statistics - missing data - stdev" -match almostzero -body {
  set value1 [::math::statistics::stdev  $::data_linear]
  set value2 [::math::statistics::stdev  $::data_missing2]
  expr {abs($value1-$value2)}
} -result 0.001 ;# Zero is impossible

test "BasicStats-1.8" "Basic statistics - missing data - var" -match almostzero -body {
  set value1 [::math::statistics::stdev  $::data_linear]
  set value2 [::math::statistics::var    $::data_missing2]
  expr {$value1*$value1-$value2}
} -result 0.001 ;# Zero is impossible

#
# Histograms
#
test "Histogram-1.0" "Histogram - uniform data" -match glob -body {
  set values [::math::statistics::histogram {0 2} $::data_uniform]
} -result [list 0 [llength $::data_uniform] 0]

test "Histogram-1.1" "Histogram - missing data" -match glob -body {
  set values [::math::statistics::histogram {0 2} $::data_missing]
} -result [list 0 [::math::statistics::number $::data_missing] 0]

test "Histogram-1.2" "Histogram - linear data" -match glob -body {
  set values [::math::statistics::histogram {1.5 4.5 9.5} $::data_linear]
} -result {1 3 5 1}

test "Histogram-1.3" "Histogram - linear data 2" -match glob -body {
  set values [::math::statistics::histogram {1.5 2.5 10.5} $::data_linear]
} -result {1 1 8 0}

#
# Quantiles
#
test "Quantiles-1.0" "Quantiles - raw data" -match tolerant -body {
  set values [::math::statistics::quantiles $::data_linear {0.25 0.55 0.95}]
} -result {3.0 6.0 10.0}

  set limits    {1.0 2.0 3.0 4.0}
  set data_hist {0 10 20 10 0}
  set values [::math::statistics::quantiles $limits $data_hist {0.25 0.5 0.9}]

test "Quantiles-1.1" "Quantiles - histogram" -match tolerant -body {
  set limits    {1.0 2.0 3.0 4.0}
  set data_hist {0 10 20 10 0}
  set values [::math::statistics::quantiles $limits $data_hist {0.25 0.5 0.9}]
} -result {2.0 2.5 3.6}

#
# Generate histogram limits
#

test "Limits-1.0" "Limits - based on mean/stdev" -match tolerant -body {
  set values [::math::statistics::mean-histogram-limits 1.0 1.0 4]
} -result {0.0 0.75 1.25 2.0}

test "Limits-1.1" "Limits - based on mean/stdev" -match tolerant -body {
  set values [::math::statistics::mean-histogram-limits 1.0 1.0 9]
} -result {-2.0 -1.0 0.0 0.75 1.0 1.25 2.0 3.0 4.0}

test "Limits-1.2" "Limits - based on mean/stdev" -match tolerant -body {
  set values [::math::statistics::mean-histogram-limits 0.0 1.0 11]
} -result {-3.0 -2.4 -1.8 -1.2 -0.6 0.0 0.6 1.2 1.8 2.4 3.0}

test "Limits-2.0" "Limits - based on min/max" -match tolerant -body {
  set values [::math::statistics::minmax-histogram-limits -2.0 2.0 9]
} -result {-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0}

test "Limits-2.1" "Limits - based on min/max" -match tolerant -body {
  set values [::math::statistics::minmax-histogram-limits -2.0 2.0 2]
} -result {-2.0 2.0}

#
# To do: design test cases for the following functions:
# - t-test-mean
# - estimate-mean-stdev
# - autocorr
# - crosscorr
# - linear-model
# - linear-residuals
# - pdf-*
# - cdf-*
# - random-*
# - histogram-*
#
# Crude test cases for Student's t test
#
test "Students-t-test-1.0" "Student's t - same sample" -match glob -body {
  set sample [::math::statistics::random-normal 0.0 1.0 40]
  set mean   0.0
  set stdev  1.0
  set confidence 0.95

  set result [::math::statistics::t-test-mean $sample $mean $stdev $confidence]
} -result 1

test "Students-t-test-1.1" "Student's t - different sample" -match glob -body {
  set sample [::math::statistics::random-normal 0.0 1.0 40]
  set mean   10.0
  set stdev   1.0
  set confidence 0.95

  set result [::math::statistics::t-test-mean $sample $mean $stdev $confidence]
} -result 0

test "Students-t-test-1.2" "Student's t - small sample" -match glob -body {
  set sample [::math::statistics::random-normal 0.0 1.0 2]
  set mean    2.0
  set stdev   1.0
  set confidence 0.90

  set result [::math::statistics::t-test-mean $sample $mean $stdev $confidence]
} -result 1

#
# Test private procedures
#
test "Cdf-toms322-1.0" "TOMS322 - erf(x)" -match tolerant2 -body {
  set result {}
  foreach z {4.417 3.891 3.291 2.576 2.241 1.960 1.645 1.150 0.674
             0.319 0.126 0.063 0.0125} {
     set prob [::math::statistics::Cdf-toms322 1 5000 [expr {$z*$z}]]
     lappend result [expr {1.0-$prob}]
  }
  set result
} -result {1.e-5 1.e-4 1.e-3 1.e-2 0.025 0.050 0.100 0.250 0.500
           0.750 0.900 0.950 0.990 }

test "Cdf-toms322-2.0" "TOMS322 - inverse erf(x)" -match tolerant2 -body {
  set result {}
  foreach p {0.5120 0.5948 0.7019 0.7996  0.8997  0.9505  0.9901  0.9980 } {
     set z [::math::statistics::Inverse-cdf-normal 0.0 1.0 $p]
     lappend result $z
  }
  set result
} -result    {0.03  0.24   0.53   0.84    1.28    1.65    2.33    2.88 }

#
# Correlation coefficients
#
test "Correlation-1.0" "Correlation - linear data" -match tolerant -body {
  set corr [::math::statistics::corr $::data_linear $::data_linear]
} -result 1.0
test "Correlation-1.1" "Correlation - linear/uniform" -match almostzero -body {
  set corr [::math::statistics::corr $::data_linear $::data_uniform]
} -result 0.0

#
# Test list procedures
#
proc matchListElements { expected actual } {
   if { [llength $expected] != [llength $actual] } {
      return 0
   } else {
      set match 1
      foreach a $actual e $expected {
         if { $a != $e } {
            set match 0
            break
         }
      }
   }
   return $match
}
customMatch matchList  matchListElements

set ::data_list {1 2 3 4 5 6 7 8 9 10}
set ::data_pairs {{1 2} {3 4} {5 6} {7 8} {9 10}}

test "Filter-1.0" "True filter" -match matchList -body {
   set data [::math::statistics::filter x $::data_list 1]
} -result $::data_list

test "Filter-1.1" "False filter" -match matchList -body {
   set data [::math::statistics::filter x $::data_list 0]
} -result {}

test "Filter-1.2" "Even filter" -match matchList -body {
   set data [::math::statistics::filter x $::data_list {$x%2==0}]
} -result {2 4 6 8 10}

test "Filter-2.1" "filter with parameter" -match matchList -body {
   set param 3.0
   set data [::math::statistics::filter x $::data_list {$x > $param}]
} -result {4 5 6 7 8 9 10}

test "Map-1.0" "Identity map" -match matchList -body {
   set data [::math::statistics::map x $::data_list {$x}]
} -result $::data_list

test "Map-1.1" "Is-even map" -match matchList -body {
   set data [::math::statistics::map x $::data_list {$x%2==0}]
} -result {0 1 0 1 0 1 0 1 0 1}

test "Map-1.2" "Double map" -match matchList -body {
   set data [::math::statistics::map x $::data_list {$x*2}]
} -result {2 4 6 8 10 12 14 16 18 20}

test "Map-2.1" "map with parameter" -match matchList -body {
   set param 3.0
   set data [::math::statistics::map x $::data_list {$x + $param}]
} -result {4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0}

test "Samplescount-1.0" "Single sublist" -match matchList -body {
   set data [::math::statistics::samplescount x [list $::data_list]]
} -result {10}

test "Samplescount-1.0" "List of singleton sublist" -match matchList -body {
   set data [::math::statistics::samplescount x $::data_list]
} -result {1 1 1 1 1 1 1 1 1 1}

test "Samplescount-1.1" "Pairs sublist" -match matchList -body {
   set data [::math::statistics::samplescount x $::data_pairs]
} -result {2 2 2 2 2}

test "Samplescount-1.2" "Select uneven sublist" -match matchList -body {
   set data [::math::statistics::samplescount x $::data_pairs {$x%2}]
} -result {1 1 1 1 1}

test "Samplescount-2.1" "Count with parameter" -match matchList -body {
   set param 3.0
   set data [::math::statistics::samplescount x $::data_pairs {$x>$param}]
} -result {0 1 2 2 2}