Utilities for Sage-mpmath interaction

Also patches some mpmath functions for speed

sage.libs.mpmath.utils.bitcount(n)

Bitcount of a Sage Integer or Python int/long.

EXAMPLES:

sage: from mpmath.libmp import bitcount
sage: bitcount(0)
0
sage: bitcount(1)
1
sage: bitcount(100)
7
sage: bitcount(-100)
7
sage: bitcount(2r)
2
sage: bitcount(2L)
2

sage.libs.mpmath.utils.call(func, *args, **kwargs)

Call an mpmath function with Sage objects as inputs and convert the result back to a Sage real or complex number.

By default, a RealNumber or ComplexNumber with the current working precision of mpmath (mpmath.mp.prec) will be returned.

If prec=n is passed among the keyword arguments, the temporary working precision will be set to n and the result will also have this precision.

If parent=P is passed, P.prec() will be used as working precision and the result will be coerced to P (or the corresponding complex field if necessary).

Arguments should be Sage objects that can be coerced into RealField or ComplexField elements. Arguments may also be tuples, lists or dicts (which are converted recursively), or any type that mpmath understands natively (e.g. Python floats, strings for options).

EXAMPLES:

sage: import sage.libs.mpmath.all as a
sage: a.mp.prec = 53
sage: a.call(a.erf, 3+4*I)
-120.186991395079 - 27.7503372936239*I
sage: a.call(a.polylog, 2, 1/3+4/5*I)
0.153548951541433 + 0.875114412499637*I
sage: a.call(a.barnesg, 3+4*I)
-0.000676375932234244 - 0.0000442236140124728*I
sage: a.call(a.barnesg, -4)
0.000000000000000
sage: a.call(a.hyper, [2,3], [4,5], 1/3)
1.10703578162508
sage: a.call(a.hyper, [2,3], [4,(2,3)], 1/3)
1.95762943509305
sage: a.call(a.quad, a.erf, [0,1])
0.486064958112256
sage: a.call(a.gammainc, 3+4*I, 2/3, 1-pi*I, prec=100)
-274.18871130777160922270612331 + 101.59521032382593402947725236*I
sage: x = (3+4*I).n(100)
sage: y = (2/3).n(100)
sage: z = (1-pi*I).n(100)
sage: a.call(a.gammainc, x, y, z, prec=100)
-274.18871130777160922270612331 + 101.59521032382593402947725236*I
sage: a.call(a.erf, infinity)
1.00000000000000
sage: a.call(a.erf, -infinity)
-1.00000000000000
sage: a.call(a.gamma, infinity)
+infinity
sage: a.call(a.polylog, 2, 1/2, parent=RR)
0.582240526465012
sage: a.call(a.polylog, 2, 2, parent=RR)
2.46740110027234 - 2.17758609030360*I
sage: a.call(a.polylog, 2, 1/2, parent=RealField(100))
0.58224052646501250590265632016
sage: a.call(a.polylog, 2, 2, parent=RealField(100))
2.4674011002723396547086227500 - 2.1775860903036021305006888982*I
sage: a.call(a.polylog, 2, 1/2, parent=CC)
0.582240526465012
sage: type(_)
<class 'sage.rings.complex_mpfr.ComplexNumber'>
sage: a.call(a.polylog, 2, 1/2, parent=RDF)
0.5822405264650125
sage: type(_)
<class 'sage.rings.real_double...RealDoubleElement...'>

Check that github issue #11885 is fixed:

sage: a.call(a.ei, 1.0r, parent=float)
1.8951178163559366

Check that github issue #14984 is fixed:

sage: a.call(a.log, -1.0r, parent=float)
3.141592653589793j

sage.libs.mpmath.utils.from_man_exp(man, exp, prec=0, rnd='d')

Create normalized mpf value tuple from mantissa and exponent.

With prec > 0, rounds the result in the desired direction if necessary.

EXAMPLES:

sage: from mpmath.libmp import from_man_exp
sage: from_man_exp(-6, -1)
(1, 3, 0, 2)
sage: from_man_exp(-6, -1, 1, 'd')
(1, 1, 1, 1)
sage: from_man_exp(-6, -1, 1, 'u')
(1, 1, 2, 1)

sage.libs.mpmath.utils.isqrt(n)

Square root (rounded to floor) of a Sage Integer or Python int/long. The result is a Sage Integer.

EXAMPLES:

sage: from mpmath.libmp import isqrt
sage: isqrt(0)
0
sage: isqrt(100)
10
sage: isqrt(10)
3
sage: isqrt(10r)
3
sage: isqrt(10L)
3

sage.libs.mpmath.utils.mpmath_to_sage(x, prec)

Convert any mpmath number (mpf or mpc) to a Sage RealNumber or ComplexNumber of the given precision.

EXAMPLES:

sage: import sage.libs.mpmath.all as a
sage: a.mpmath_to_sage(a.mpf('2.5'), 53)
2.50000000000000
sage: a.mpmath_to_sage(a.mpc('2.5','-3.5'), 53)
2.50000000000000 - 3.50000000000000*I
sage: a.mpmath_to_sage(a.mpf('inf'), 53)
+infinity
sage: a.mpmath_to_sage(a.mpf('-inf'), 53)
-infinity
sage: a.mpmath_to_sage(a.mpf('nan'), 53)
NaN
sage: a.mpmath_to_sage(a.mpf('0'), 53)
0.000000000000000

A real example:

sage: RealField(100)(pi)
3.1415926535897932384626433833
sage: t = RealField(100)(pi)._mpmath_(); t
mpf('3.1415926535897932')
sage: a.mpmath_to_sage(t, 100)
3.1415926535897932384626433833

We can ask for more precision, but the result is undefined:

sage: a.mpmath_to_sage(t, 140) # random
3.1415926535897932384626433832793333156440
sage: ComplexField(140)(pi)
3.1415926535897932384626433832795028841972

A complex example:

sage: ComplexField(100)([0, pi])
3.1415926535897932384626433833*I
sage: t = ComplexField(100)([0, pi])._mpmath_(); t
mpc(real='0.0', imag='3.1415926535897932')
sage: sage.libs.mpmath.all.mpmath_to_sage(t, 100)
3.1415926535897932384626433833*I

Again, we can ask for more precision, but the result is undefined:

sage: sage.libs.mpmath.all.mpmath_to_sage(t, 140) # random
3.1415926535897932384626433832793333156440*I
sage: ComplexField(140)([0, pi])
3.1415926535897932384626433832795028841972*I

sage.libs.mpmath.utils.normalize(sign, man, exp, bc, prec, rnd)

Create normalized mpf value tuple from full list of components.

EXAMPLES:

sage: from mpmath.libmp import normalize
sage: normalize(0, 4, 5, 3, 53, 'n')
(0, 1, 7, 1)

sage.libs.mpmath.utils.sage_to_mpmath(x, prec)

Convert any Sage number that can be coerced into a RealNumber or ComplexNumber of the given precision into an mpmath mpf or mpc. Integers are currently converted to int.

Lists, tuples and dicts passed as input are converted recursively.

EXAMPLES:

sage: import sage.libs.mpmath.all as a
sage: a.mp.dps = 15
sage: print(a.sage_to_mpmath(2/3, 53))
0.666666666666667
sage: print(a.sage_to_mpmath(2./3, 53))
0.666666666666667
sage: print(a.sage_to_mpmath(3+4*I, 53))
(3.0 + 4.0j)
sage: print(a.sage_to_mpmath(1+pi, 53))
4.14159265358979
sage: a.sage_to_mpmath(infinity, 53)
mpf('+inf')
sage: a.sage_to_mpmath(-infinity, 53)
mpf('-inf')
sage: a.sage_to_mpmath(NaN, 53)
mpf('nan')
sage: a.sage_to_mpmath(0, 53)
0
sage: a.sage_to_mpmath([0.5, 1.5], 53)
[mpf('0.5'), mpf('1.5')]
sage: a.sage_to_mpmath((0.5, 1.5), 53)
(mpf('0.5'), mpf('1.5'))
sage: a.sage_to_mpmath({'n':0.5}, 53)
{'n': mpf('0.5')}