Sounds like a lot of work.  Let me ask you this: do you know if the calibration parameter is a simple multiplier that is applied consistently across the entire range of values or is it more complicated?    For example, if the actual TDS were say 90 and the sensor was reading 100, would the calibration factor be a simple zero point shift (ie -10), a multiplier (ie x 0.9), or some combination of the two?  

I'm thinking this function could me incorporated directly into the arduino code as it might be simpler then altering the TDS library. 

Sounds like I might have to do some experiments incorporating a range of calibration fluids to see what's up.