Better handling of generic csv tables

Output of ZOLA, thunderstorm or csv convert from Picasso are handled adequately. Adding drift columns to the end where missing.

Corrected progress bar

v0.4.2

version patch

Improved fit using floating polynomial fit

Now predetection by peak of xcorr is only used ones, firther initialization for the fit is taken from the last frame.

Improved fit using 3D gaussian for xcorr peak

v0.3.0

logging back to warning

old release

0.2.0

matplotlib debug