I think the pendulum arc could be long enough to accommodate 2 magnets (one of each pole)

@Gotoluc
Some observations.

When the approach of the rotating magnet is very precisely
at right angles (in all planes) to the sliding magnet....
there is practically no work done against magnetic forces...
for rotator to approach / exit.
................................................
Even the very slight off angle / changing angle ....of  approach by the rotating
magnet (which is due to the curvature of the bicycle rim) is having a significant
effect.

Using longer magnets will,  increase the work input, more than
one might expect due to this feature.  (longer magnets will have a more dramatic
change in distance from / angle to .... the sliding magnet... at a given curvature of
their approach line  (I think)....
.....................................................
The greater the diameter of the rotating rim.... the more nearly,
the rotating magnet's approach to the sliding magnet....
comes toward being a straight on / 180 degree (2 x 90 degree) approach.

There will be a diminishing return in the benefits of a longer pendulum
(from pivot point to magnets).  At some particular length, any further increase
in length... will give almost no decrease in the force needed to approach
/ escape.

The more perfect the alignments the more nearly one will approach
a zero for the input.  There is a limit... even the magnets / their fields are
not be perfect.

Also, at close proximity "magnetic domain flipping" within the  magnets
might have some undesirable effect. ..... But when 2 magnetic poles approach
2 other magnetic poles... these effect might completely self cancel.