Is it possible to deduce a function from its fourier series ?
You can recover (1/2)(f(x-)+f(x+)) from the Fourier series of f. Among other things the Fourier series does not distinguish between functions that differ on a set of measure zero. This is not a major problem as we often either know what f does on sets of measure zero or do not care. If you mean can we find the equation of a certain form for f, generally yes in principle but it can be very difficult. Like if you were given some Fourier series of say (3x^3+2x +4)e^(3x^2+9x)cos((3x^2+5x-7)sin(9x^2+4x+3) you would have a hard time guessing that.
Separate names with a comma.