agilent: Implement new uncertainty and other fixes

[PeterKraus]

• Implements Better uncertainties for yadg-7.0 #240 for agilent.csv, agilent.ch, agilent.dx parsers:
  • For agilent.ch and agilent.dx this is trivial, as a scaling factor has been used previously.
  • For agilent.csv, we now pick the maximum value from the str -> float conversion for each trace; the test file that is available shows that the number of decimal places is constant in the whole file and scientific notation is not used, i.e. the uncertainty is absolute.
• In agilent.csv, we now assign elution_time as a proper coordinate of signal. We also remove any manual padding. The elution_time is now expanded as necessary on concatenation (using xr.concat), and np.nans are inserted into signal at the necessary positions automatically.