roiencoder-decoder issue (different roi hashcode ) or just me?

[feelc]

Hi!

I'm not quite sure if I'm saving and recovering rois properly.
When I compare the roi hashcode from before and after saving it seems to change.
Is the hashcode based only on the polygon and position or maybe also in other variables lost or unrecovered in the encode-decode process?
Am I missing something?

To save I use:
    RoiEncoder re = new RoiEncoder(roiPath);
    re.write(roi);      

To load I use:
    roi = RoiDecoder.open(roiPath);


Thanks a lot!

Felipe

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[ctrueden]

Hi Felipe,

> Is the hashcode based only on the polygon and position or maybe also
> in other variables lost or unrecovered in the encode-decode process?

If you look at the source code [1], the Roi classes do not actually
implement the hashCode() method (in fact, nothing whatsoever in ImageJ1
does). So the Object implementation is inherited, which hashes on the
object reference, meaning it will be different every time you create a new
one (even if the coordinates etc. of the ROI are identical).

It's true that Roi implements a different method called getHashCode() [2],
but that does an XOR with hashCode(), so it will also differ every time in
a similar way.

Regards,
Curtis

[1]
https://github.com/imagej/ImageJA/blob/v1.49v/src/main/java/ij/gui/Roi.java
[2]
https://github.com/imagej/ImageJA/blob/v1.49v/src/main/java/ij/gui/Roi.java#L1947-L1952

On Wed, Jul 1, 2015 at 4:48 PM, Felipe Contreras <[hidden email]>
wrote:

> Hi!
>
> I'm not quite sure if I'm saving and recovering rois properly.
> When I compare the roi hashcode from before and after saving it seems to
> change.
> Is the hashcode based only on the polygon and position or maybe also in
> other variables lost or unrecovered in the encode-decode process?
> Am I missing something?
>
> To save I use:
>     RoiEncoder re = new RoiEncoder(roiPath);
>     re.write(roi);
>
> To load I use:
>     roi = RoiDecoder.open(roiPath);
>
>
> Thanks a lot!
>
> Felipe
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[feelc]

Nice! Thanks a lot for the fast answer!

By the way, it would nice if Roi had a hashCode method  that is based on
the polygon and position.
I may have to implement it myself(as of need). I haven't found it in the
polygon class. Is it worth writing it or am I reinventing the wheel?

Thanks a  lot again!

Felipe

2015-07-02 0:04 GMT+02:00 Curtis Rueden <[hidden email]>:

> Hi Felipe,
>
> > Is the hashcode based only on the polygon and position or maybe also
> > in other variables lost or unrecovered in the encode-decode process?
>
> If you look at the source code [1], the Roi classes do not actually
> implement the hashCode() method (in fact, nothing whatsoever in ImageJ1
> does). So the Object implementation is inherited, which hashes on the
> object reference, meaning it will be different every time you create a new
> one (even if the coordinates etc. of the ROI are identical).
>
> It's true that Roi implements a different method called getHashCode() [2],
> but that does an XOR with hashCode(), so it will also differ every time in
> a similar way.
>
> Regards,
> Curtis
>
> [1]
> https://github.com/imagej/ImageJA/blob/v1.49v/src/main/java/ij/gui/Roi.java
> [2]
>
> https://github.com/imagej/ImageJA/blob/v1.49v/src/main/java/ij/gui/Roi.java#L1947-L1952
>
> On Wed, Jul 1, 2015 at 4:48 PM, Felipe Contreras <[hidden email]>
> wrote:
>
> > Hi!
> >
> > I'm not quite sure if I'm saving and recovering rois properly.
> > When I compare the roi hashcode from before and after saving it seems to
> > change.
> > Is the hashcode based only on the polygon and position or maybe also in
> > other variables lost or unrecovered in the encode-decode process?
> > Am I missing something?
> >
> > To save I use:
> >     RoiEncoder re = new RoiEncoder(roiPath);
> >     re.write(roi);
> >
> > To load I use:
> >     roi = RoiDecoder.open(roiPath);
> >
> >
> > Thanks a lot!
> >
> > Felipe
> >
> > --
> > ImageJ mailing list: http://imagej.nih.gov/ij/list.html
> >
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html