Rational approximation : float2rational exact

float2rational (or exact) takes as argument an expression.
float2rational returns a rational approximation of all the floating point numbers r contained in this expression, such that | r - float2rational(r)| < $\epsilon$, where $\epsilon$ is defined by epsilon in the cas configuration (Cfg menu, or cas_setup command).
Input :

float2rational(1.5)

Output :

3/2

Input :

float2rational(1.414)

Output :

707/500

Input :

float2rational(0.156381102937*2)

Output :

5144/16447

Input :

float2rational(1.41421356237)

Output :

114243/80782

Input :

float2rational(1.41421356237^2)

Output :

2