Source: 📖 Effective Python item 69

Using decimal

Python's float type is not always exact, which can lead to surprising results. When absolute precision is required, instead of using a float type you can pass the required number (before calculation) to the Decimal constructor.

from decimcal import Decimal

print(Decimal('1.45'))		# passed as str
print(Decimal(1.45))		#passed as float

>>>
1.45
1.4499999999999999555910790149937383830547332763671875

The resulting number can then be used for calculations with absolute accuracy (be sure that all other numbers involved in the calculation are Decimal types in order to guarantee accuracy).