def test_opt():
from scipy.optimize import fmin_tnc
x, y = a3200.xyposition()
fmin_tnc(value_and_gradient, (x,y), None,
[(lambda: abs(keyence.measure(1, blocking=True) -
keyence.measure(2, blocking=False))), 0.001],
False, [(x-.008,x+.008), (y-.008,y+.008)],)
def follow_edge(step=1):
a3200.x.freerun(1)
while keyence.measure() is not None: time.sleep(.01)
a3200.x.halt()
a3200.x.relmove(-step/2)
origin, xy = a3200.xyposition(), numpy.array([0, 0])
heading, path, i, dark = 0, [], 0, False
while tuple(xy) not in path[:-10]:
ldark, dark = dark, (keyence.measure(blocking=True) is None)
if dark != ldark: i = 0
if i%2 or not i: heading = (heading + (-1 if dark else 1)) % 4
xy += [(1,0),(0,1),(-1,0),(0,-1)][heading]
a3200.xymove(origin + xy*step)
path.append(tuple(xy))
i += 1
return origin + numpy.array(path, dtype='float')*step
def continuous_spiral_search(continue_func, step=.01, vel=.1):
origin = a3200.xyposition()
heading = 0
for r in itertools.count():
a3200.xyrelmove([(1,0),(0,1),(-1,0),(0,-1)][r%4]*(r//4+1)*step,
blocking = False)
def spiral_search(continue_func, step=.01):
p0 = a3200.xyposition()
s = spiral()
while continue_func():
a3200.xymove(p0+numpy.array(next(s))*step, blocking=True)