Marco,
Thank you for the nice words and contribution.
Since I was opening up the code today anyway, I made your small change
in the source -- you are right, it's nice to be able to subclass the
BBox. Subclassing numpy arrays is a bit tricky, thanks for figuring this
out.
I've cc'd this to the FloatCanvas list for the archives.
-Chris
On 10/7/11 11:09 AM, Marco Oster wrote:
> Chris,
>
> I think I have an improvement of your fine FC widget, see patch:
>
> --- /tmp/BBox.py 2011-10-07 19:48:08.000000000 +0200
> +++ BBox.py 2011-10-07 19:18:11.000000000 +0200
> @@ -29,7 +29,7 @@
> fromPoints
>
> """
> - def __new__(subclass, data):
> + def __new__(cls, data):
> """
> Takes Data as an array. Data is any python sequence that can be
> turned into a
> 2x2 numpy array of floats:
> @@ -51,7 +51,7 @@
> if arr[0,0] > arr[1,0] or arr[0,1] > arr[1,1]:
> # note: zero sized BB OK.
> raise ValueError("BBox values not aligned: \n minimum values
> must be less that maximum values")
> - return N.ndarray.__new__(BBox, shape=arr.shape, dtype=arr.dtype,
> buffer=arr)
> + return N.ndarray.__new__(cls, shape=arr.shape, dtype=arr.dtype,
> buffer=arr)
>
> def Overlaps(self, BB):
> """
>
> The change actually enables subclassing of BBox objects. I it for checking of
> hit events on rotated rectangular objects, in conjunction with some BBox
> based hit test code found on the mailing list. It boils down to something
> like that:
>
> class RectBBox(BBox.BBox):
>
> def __new__(self, data, edges=None):
> return BBox.BBox.__new__(self, data)
>
> def __init__(self, data, edges=None):
> ''' assume edgepoints are ordered such you can walk along all edges
> with left rotation sense
> This may be:
> left-top
> left-bottom
> right-bottom
> right-top
>
> or any rotation.
> '''
> BBox.BBox(data)
> self.edges = np.asarray(edges)
>
> print "new rectbbox created"
>
>
> def ac_leftOf_ab(self, a, b, c):
> ab = np.array(b) - np.array(a)
> ac = np.array(c) - np.array(a)
>
> return (ac[0]*ab[1] - ac[1]*ab[0]) <= 0
>
> def PointInside(self, point):
> print "point inside called"
>
> for edge in xrange(4):
> if self.ac_leftOf_ab(self.edges[edge],
> self.edges[(edge+1)%4],
> point):
> continue
> else:
> return False
> return True
>
> I evaluate FloatCanvas for position-aware drawing of electron micrographs
> taken with a TEM. Attached is a screenshot of the results of my code mess :)
> Imaged area is about 3mm in diameter, at an approximate magnification of 160.
> The differently rotated insets are taken at 500x and 6000x. Specimen were
> meant to be vitrified liposomes, but there was next to nothing in.
>
> Thank you for sharing the nice package and please keep up the good work!
> Marco
--
Christopher Barker, Ph.D.
Oceanographer
Emergency Response Division
NOAA/NOS/OR&R (206) 526-6959 voice
7600 Sand Point Way NE (206) 526-6329 fax
Seattle, WA 98115 (206) 526-6317 main reception
[email protected]
_______________________________________________
FloatCanvas mailing list
[email protected]
http://paulmcnett.com/cgi-bin/mailman/listinfo/floatcanvas
