This implements a new heuristic to place labels that works as follow:
We find the extent of the data and of the plot
We generate equally spaced positions in between
If the first line to be annotated contains the first position, the second line the second position, ... then we're done
Otherwise, we mix and match the label positions with the lines to maximise the number of lines that contain a position. For the remaining ones, we simply put the label on the line wherever we find a spot.
This implements a new heuristic to place labels that works as follow:
This fixes #58.