segpy.ibm_float module

class segpy.ibm_float.IBMFloat

Bases: numbers.Real

as_integer_ratio()
exp16

The base 16 exponent.

frexp()

Obtain the fraction and exponent.

Returns:A pair where the first item is the fraction in the range -1.0 and +1.0 and the exponent is an integer such that f = fraction * 2**exponent
classmethod from_bytes(b)
classmethod from_float(f)

Construct an IBMFloat from an IEEE float.

Parameters:

f (float) – The value to be converted.
Returns:

An IBMFloat.
Raises:
  • OverflowError - If f is outside the representable range.
  • ValueError - If f is NaN or infinite.
  • FloatingPointError - If f cannot be represented without total loss of precision.
classmethod from_float_without_underflow(f)

Construct an IBMFloat from an IEEE float.

Parameters:

f (float) – The value to be converted. If the provided IEEE value underflows the smallest representable IBM value, this function returns zero.
Returns:

An IBMFloat.
Raises:
  • OverflowError - If f is outside the representable range.
  • ValueError - If f is NaN or infinite.
classmethod from_real(f)
int_mantissa
is_subnormal()
is_zero()
classmethod ldexp(fraction, exponent)

Make an IBMFloat from fraction and exponent.

The is the inverse function of IBMFloat.frexp()

Parameters:
  • fraction – A Real in the range -1.0 to 1.0.
  • exponent – An integer in the range -256 to 255 inclusive.
normalize()

Normalize the floating point representation.

Returns:A normalized IBMFloat equal in value to this object.
Raises:FloatingPointError - If the number could not be normalized.
signbit

True if the value is negative, otherwise False.

try_normalize()

Normalize if possible.

If it is not possible to normalize the representation, it remains unmodified.

zero_subnormal()
segpy.ibm_float.ibm2ieee(big_endian_bytes)

Interpret a byte string as a big-endian IBM float.

Parameters:big_endian_bytes (bytes) – A byte-string containing at least four bytes.
Returns:The floating point value.
segpy.ibm_float.ieee2ibm(f)

Convert a float to four big-endian bytes representing an IBM float.

Parameters:

f (float) – The value to be converted.
Returns:

A bytes object (Python 3) or a string (Python 2) containing four bytes representing a big-endian IBM float.
Raises:
  • OverflowError - If f is outside the representable range.
  • ValueError - If f is NaN or infinite.
  • FloatingPointError - If f cannot be represented without total loss of precision.