music21.common.numberTools

Functions

music21.common.numberTools.fromRoman(num, *, strictModern=False)

Convert a Roman numeral (upper or lower) to an int

http://code.activestate.com/recipes/81611-roman-numerals/

>>> common.fromRoman('ii')
2
>>> common.fromRoman('vii')
7

Works with both IIII and IV forms:

>>> common.fromRoman('MCCCCLXXXIX')
1489
>>> common.fromRoman('MCDLXXXIX')
1489

Some people consider this an error, but you see it in medieval and ancient roman documents:

>>> common.fromRoman('ic')
99

unless strictModern is True

>>> common.fromRoman('ic', strictModern=True)
Traceback (most recent call last):
ValueError: input contains an invalid subtraction element (modern interpretation): ic

But things like this are never seen, and thus cause an error:

>>> common.fromRoman('vx')
Traceback (most recent call last):
ValueError: input contains an invalid subtraction element: vx
Return type

int

music21.common.numberTools.toRoman(num)

Convert a number from 1 to 3999 to a roman numeral

>>> common.toRoman(2)
'II'
>>> common.toRoman(7)
'VII'
>>> common.toRoman(1999)
'MCMXCIX'
>>> common.toRoman('hi')
Traceback (most recent call last):
TypeError: expected integer, got <... 'str'>
Return type

str

music21.common.numberTools.addFloatPrecision(x, grain=0.01)

Given a value that suggests a floating point fraction, like .33, return a Fraction or float that provides greater specification, such as .333333333

>>> import fractions
>>> common.addFloatPrecision(.333)
Fraction(1, 3)
>>> common.addFloatPrecision(.33)
Fraction(1, 3)
>>> common.addFloatPrecision(.35) == fractions.Fraction(1, 3)
False
>>> common.addFloatPrecision(.2) == 0.2
True
>>> common.addFloatPrecision(.125)
0.125
>>> common.addFloatPrecision(1./7) == 1./7
True
Return type

float

music21.common.numberTools.almostEquals(x, y=0.0, grain=1e-07)

almostEquals(x, y) – returns True if x and y are within grain (default 0.0000001) of each other

Allows comparisons between floats that are normally inconsistent.

>>> common.almostEquals(1.000000001, 1)
True
>>> common.almostEquals(1.001, 1)
False
>>> common.almostEquals(1.001, 1, grain=0.1)
True

For very small grains, just compare Fractions without converting…

Return type

bool

music21.common.numberTools.approximateGCD(values, grain=0.0001)

Given a list of values, find the lowest common divisor of floating point values.

>>> common.approximateGCD([2.5, 10, 0.25])
0.25
>>> common.approximateGCD([2.5, 10])
2.5
>>> common.approximateGCD([2, 10])
2.0
>>> common.approximateGCD([1.5, 5, 2, 7])
0.5
>>> common.approximateGCD([2, 5, 10])
1.0
>>> common.approximateGCD([2, 5, 10, 0.25])
0.25
>>> common.strTrimFloat(common.approximateGCD([1/3, 2/3]))
'0.3333'
>>> common.strTrimFloat(common.approximateGCD([5/3, 2/3, 4]))
'0.3333'
>>> common.strTrimFloat(common.approximateGCD([5/3, 2/3, 5]))
'0.3333'
>>> common.strTrimFloat(common.approximateGCD([5/3, 2/3, 5/6, 3/6]))
'0.1667'
Return type

float

music21.common.numberTools.cleanupFloat(floatNum, maxDenominator=65535)

Cleans up a floating point number by converting it to a fractions.Fraction object limited to a denominator of maxDenominator

>>> common.cleanupFloat(0.33333327824)
0.333333333333...
>>> common.cleanupFloat(0.142857)
0.1428571428571...
>>> common.cleanupFloat(1.5)
1.5

Fractions are passed through silently…

>>> import fractions
>>> common.cleanupFloat(fractions.Fraction(4, 3))
Fraction(4, 3)
music21.common.numberTools.contiguousList(inputListOrTuple)

returns bool True or False if a list containing ints contains only contiguous (increasing) values

requires the list to be sorted first

>>> l = [3, 4, 5, 6]
>>> common.contiguousList(l)
True
>>> l.append(8)
>>> common.contiguousList(l)
False

Sorting matters

>>> l.append(7)
>>> common.contiguousList(l)
False
>>> common.contiguousList(sorted(l))
True
Return type

bool

music21.common.numberTools.decimalToTuplet(decNum)

For simple decimals (usually > 1), a quick way to figure out the fraction in lowest terms that gives a valid tuplet.

No it does not work really fast. No it does not return tuplets with denominators over 100. Too bad, math geeks. This is real life. :-)

returns (numerator, denominator)

>>> common.decimalToTuplet(1.5)
(3, 2)
>>> common.decimalToTuplet(1.25)
(5, 4)

If decNum is < 1, the denominator will be greater than the numerator:

>>> common.decimalToTuplet(.8)
(4, 5)

If decNum is <= 0, returns a ZeroDivisionError:

>>> common.decimalToTuplet(-.02)
Traceback (most recent call last):
ZeroDivisionError: number must be greater than zero

TODO: replace with fractions…

Return type

tuple(int)

music21.common.numberTools.dotMultiplier(dots)

dotMultiplier(dots) returns how long to multiply the note length of a note in order to get the note length with n dots

Since dotMultiplier always returns a power of two in the denominator, the float will be exact.

>>> from fractions import Fraction
>>> Fraction(common.dotMultiplier(1))
Fraction(3, 2)
>>> Fraction(common.dotMultiplier(2))
Fraction(7, 4)
>>> Fraction(common.dotMultiplier(3))
Fraction(15, 8)
>>> common.dotMultiplier(0)
1.0
Return type

float

music21.common.numberTools.euclidGCD(a, b)

use Euclid’s algorithm to find the GCD of a and b

>>> common.euclidGCD(2, 4)
2
>>> common.euclidGCD(20, 8)
4
>>> common.euclidGCD(20, 16)
4
Return type

int

music21.common.numberTools.groupContiguousIntegers(src)

Given a list of integers, group contiguous values into sub lists

>>> common.groupContiguousIntegers([3, 5, 6])
[[3], [5, 6]]
>>> common.groupContiguousIntegers([3, 4, 6])
[[3, 4], [6]]
>>> common.groupContiguousIntegers([3, 4, 6, 7])
[[3, 4], [6, 7]]
>>> common.groupContiguousIntegers([3, 4, 6, 7, 20])
[[3, 4], [6, 7], [20]]
>>> common.groupContiguousIntegers([3, 4, 5, 6, 7])
[[3, 4, 5, 6, 7]]
>>> common.groupContiguousIntegers([3])
[[3]]
>>> common.groupContiguousIntegers([3, 200])
[[3], [200]]
music21.common.numberTools.lcm(filterList)

Find the least common multiple of a list of values

>>> common.lcm([3, 4, 5])
60
>>> common.lcm([3, 4])
12
>>> common.lcm([1, 2])
2
>>> common.lcm([3, 6])
6
Return type

int

music21.common.numberTools.mixedNumeral(expr, limitDenominator=65535)

Returns a string representing a mixedNumeral form of a number

>>> common.mixedNumeral(1.333333)
'1 1/3'
>>> common.mixedNumeral(0.333333)
'1/3'
>>> common.mixedNumeral(-1.333333)
'-1 1/3'
>>> common.mixedNumeral(-0.333333)
'-1/3'
>>> common.mixedNumeral(0)
'0'
>>> common.mixedNumeral(-0)
'0'

Works with Fraction objects too

>>> from fractions import Fraction
>>> common.mixedNumeral(Fraction(31, 7))
'4 3/7'
>>> common.mixedNumeral(Fraction(1, 5))
'1/5'
>>> common.mixedNumeral(Fraction(-1, 5))
'-1/5'
>>> common.mixedNumeral(Fraction(-31, 7))
'-4 3/7'

Denominator is limited by default but can be changed.

>>> common.mixedNumeral(2.0000001)
'2'
>>> common.mixedNumeral(2.0000001, limitDenominator=10000000)
'2 1/10000000'
music21.common.numberTools.nearestMultiple(n, unit)

Given a positive value n, return the nearest multiple of the supplied unit as well as the absolute difference (error) to seven significant digits and the signed difference.

>>> print(common.nearestMultiple(.25, .25))
(0.25, 0.0, 0.0)
>>> print(common.nearestMultiple(.35, .25))
(0.25, 0.1..., 0.1...)
>>> print(common.nearestMultiple(.20, .25))
(0.25, 0.05..., -0.05...)

Note that this one also has an error of .1 but it’s a positive error off of 0.5 >>> print(common.nearestMultiple(.4, .25)) (0.5, 0.1…, -0.1…)

>>> common.nearestMultiple(.4, .25)[0]
0.5
>>> common.nearestMultiple(23404.001, .125)[0]
23404.0
>>> common.nearestMultiple(23404.134, .125)[0]
23404.125

Error is always positive, but signed difference can be negative.

>>> common.nearestMultiple(23404 - 0.0625, .125)
(23403.875, 0.0625, 0.0625)
>>> common.nearestMultiple(.001, .125)[0]
0.0
>>> common.almostEquals(common.nearestMultiple(0.25, 1 / 3)[0], .33333333)
True
>>> common.almostEquals(common.nearestMultiple(0.55, 1 / 3)[0], .66666666)
True
>>> common.almostEquals(common.nearestMultiple(234.69, 1 / 3)[0], 234.6666666)
True
>>> common.almostEquals(common.nearestMultiple(18.123, 1 / 6)[0], 18.16666666)
True
>>> common.nearestMultiple(-0.5, 0.125)
Traceback (most recent call last):
ValueError: n (-0.5) is less than zero. Thus cannot find nearest
    multiple for a value less than the unit, 0.125
Return type

tuple(float)

music21.common.numberTools.numToIntOrFloat(value)

Given a number, return an integer if it is very close to an integer, otherwise, return a float.

This routine is very important for conversion of Accidental objects’ .alter attribute in musicXML must be 1 (not 1.0) for sharp and -1 (not -1.0) for flat, but allows for 0.5 for half-sharp.

>>> common.numToIntOrFloat(1.0)
1
>>> common.numToIntOrFloat(1.00003)
1.00003
>>> common.numToIntOrFloat(1.5)
1.5
>>> common.numToIntOrFloat(1.0000000005)
1
>>> common.numToIntOrFloat(0.999999999)
1
>>> sharp = pitch.Accidental('sharp')
>>> common.numToIntOrFloat(sharp.alter)
1
>>> halfFlat = pitch.Accidental('half-flat')
>>> common.numToIntOrFloat(halfFlat.alter)
-0.5

Also can take in a string representing an int or float

>>> common.numToIntOrFloat('1.0')
1
>>> common.numToIntOrFloat('1')
1
>>> common.numToIntOrFloat('1.25')
1.25
Return type

float

music21.common.numberTools.opFrac(num)

opFrac -> optionally convert a number to a fraction or back.

Important music21 function for working with offsets and quarterLengths

Takes in a number (or None) and converts it to a Fraction with denominator less than limitDenominator if it is not binary expressible; otherwise return a float. Or if the Fraction can be converted back to a binary expressable float then do so.

This function should be called often to ensure that values being passed around are floats and ints wherever possible and fractions where needed.

The naming of this method violates music21’s general rule of no abbreviations, but it is important to make it short enough so that no one will be afraid of calling it often. It also doesn’t have a setting for maxDenominator so that it will expand in Code Completion easily. That is to say, this function has been set up to be used, so please use it.

This is a performance critical operation. Do not alter it in any way without running many timing tests.

>>> from fractions import Fraction
>>> defaults.limitOffsetDenominator
65535
>>> common.opFrac(3)
3.0
>>> common.opFrac(1/3)
Fraction(1, 3)
>>> common.opFrac(1/4)
0.25
>>> f = Fraction(1, 3)
>>> common.opFrac(f + f + f)
1.0
>>> common.opFrac(0.123456789)
Fraction(10, 81)
>>> common.opFrac(None) is None
True
music21.common.numberTools.ordinalAbbreviation(value, plural=False)

Return the ordinal abbreviations for integers

>>> common.ordinalAbbreviation(3)
'rd'
>>> common.ordinalAbbreviation(255)
'th'
>>> common.ordinalAbbreviation(255, plural=True)
'ths'
Return type

str

music21.common.numberTools.roundToHalfInteger(num)

Given a floating-point number, round to the nearest half-integer. Returns int or float

>>> common.roundToHalfInteger(1.2)
1
>>> common.roundToHalfInteger(1.35)
1.5
>>> common.roundToHalfInteger(1.8)
2
>>> common.roundToHalfInteger(1.6234)
1.5

.25 rounds up:

>>> common.roundToHalfInteger(0.25)
0.5

as does .75

>>> common.roundToHalfInteger(0.75)
1

unlike python round function, does the same for 1.25 and 1.75

>>> common.roundToHalfInteger(1.25)
1.5
>>> common.roundToHalfInteger(1.75)
2

negative numbers however, round up on the boundaries

>>> common.roundToHalfInteger(-0.26)
-0.5
>>> common.roundToHalfInteger(-0.25)
0
Return type

float

music21.common.numberTools.standardDeviation(coll, bassel=False)

DEPRECATED: use statistics.stdev instead.

Given a collection of values, return the standard deviation.

Return type

float

music21.common.numberTools.strTrimFloat(floatNum, maxNum=4)

returns a string from a float that is at most maxNum of decimal digits long, but never less than 1.

>>> common.strTrimFloat(42.3333333333)
'42.3333'
>>> common.strTrimFloat(42.3333333333, 2)
'42.33'
>>> common.strTrimFloat(6.66666666666666, 2)
'6.67'
>>> common.strTrimFloat(2.0)
'2.0'
>>> common.strTrimFloat(-5)
'-5.0'
Return type

str

music21.common.numberTools.unitBoundaryProportion(series)

Take a series of parts with an implied sum, and create unit-interval boundaries proportional to the series components.

>>> common.unitBoundaryProportion([1, 1, 2])
[(0, 0.25), (0.25, 0.5), (0.5, 1.0)]
>>> common.unitBoundaryProportion([8, 1, 1])
[(0, 0.8...), (0.8..., 0.9...), (0.9..., 1.0)]
Return type

list(tuple(float))

music21.common.numberTools.unitNormalizeProportion(values)

Normalize values within the unit interval, where max is determined by the sum of the series.

>>> common.unitNormalizeProportion([0, 3, 4])
[0.0, 0.42857142857142855, 0.5714285714285714]
>>> common.unitNormalizeProportion([1, 1, 1])
[0.3333333..., 0.333333..., 0.333333...]

On 32-bit computers this number is inexact. On 64-bit it works fine.

#>>> common.unitNormalizeProportion([.2, .6, .2]) #[0.20000000000000001, 0.59999999999999998, 0.20000000000000001]

Return type

list(float)

music21.common.numberTools.weightedSelection(values, weights, randomGenerator=None)

Given a list of values and an equal-sized list of weights, return a randomly selected value using the weight.

Example: sum -1 and 1 for 100 values; should be around 0 or at least between -50 and 50 (99.99999% of the time)

>>> -50 < sum([common.weightedSelection([-1, 1], [1, 1]) for x in range(100)]) < 50
True
Return type

int