[ome-devel] Scanr patch

[Melissa Linkert]

Hi Rubén,

> I have tried just it with two different sets that do not form a square. 
> Wow. These last changes are brilliant. The names and the conversion are precise.

I am very glad to hear it!

> Just a minor thing, in line 520, the item "loop count" is the number of cycles but " 1" should be added.
>
> core[0].sizeT = Integer.parseInt(value) + 1;

Unfortunately, that is not always the case; for example, the latest dataset
that you sent (20100827_004) clearly has 10 timepoints and the value of
"timeloop count" is 10.

What I have done with r6967 is to set core[0].sizeT as you suggest, but then adjust the value later if it does not match up with the actual number of TIFF files.  This seems to work with all of the datasets that I have, but as always please let me know if it does not work for you.

Regards,
-Melissa

On Mon, Sep 20, 2010 at 09:49:49PM +0200, Rubén Muñoz wrote:
> Hi Melissa,
> I have tried just it with two different sets that do not form a square. 
> Wow. These last changes are brilliant. The names and the conversion are precise.
> 
> Just a minor thing, in line 520, the item "loop count" is the number of cycles but " 1" should be added.
> 
> core[0].sizeT = Integer.parseInt(value) + 1;
> 
> Actually sometimes: 
> I have noticed that, when the acquisition was aborted and 'core[0].sizeT' is parsed, missing frames of one serie are taken from next serie.
> Of course the code expects the coherency with the descriptor, so I update it manually in those cases.
> 
> Best regards,
> Rubén
> 
> 
> On Sep 20, 2010, at 6:36 PM, Melissa Linkert wrote:
> 
> > Hi Rubén,
> > 
> >> That's a good point, actually we still need it.
> >> I certainly forgot to recalculate this value.
> > 
> > OK, no problem.
> > 
> > With r6959 (which closes ticket #550), non-acquired wells should no
> > longer be included in the series list.  As far as I can tell, acquired
> > wells are also being properly labelled now, but if you could confirm that I
> > would greatly appreciate it.
> > 
> > Regards,
> > -Melissa
> > 
> > On Fri, Sep 17, 2010 at 09:08:44AM +0200, Rubén Muñoz wrote:
> >> Hi Melissa,
> >> 
> >> That's a good point, actually we still need it. 
> >> I certainly forgot to recalculate this value. What about?:
> >> 
> >> int nSpot = well * nPos + pos + 1;
> >> names[realPosCount] = "Well " + (well + 1) + ", Field " + (pos +1) + " (Spot " + nSpot + ")";
> >> 
> >> The code doesn't label non-acquired wells.
> >> 
> >> Best,
> >> Rubén
> >> 
> >> On Sep 16, 2010, at 11:58 PM, Melissa Linkert wrote:
> >> 
> >>> Hi Rubén,
> >>> 
> >>> Applying this patch is not a problem, but I just want to verify that the
> >>> format of the Image names is correct.
> >>> 
> >>> In a previous thread, you had asked to have the spot number included in
> >>> the Image name (as is currently being done).  See:
> >>> 
> >>> http://lists.openmicroscopy.org.uk/pipermail/ome-devel/2009-December/001536.html
> >>> http://lists.openmicroscopy.org.uk/pipermail/ome-devel/2009-December/001538.html
> >>> 
> >>> Are you sure that the spot number should be removed from the Image name?
> >>> 
> >>> Regards,
> >>> -Melissa
> >>> 
> >>> On Thu, Sep 16, 2010 at 09:47:45AM +0200, Rubén Muñoz wrote:
> >>>> Hi Melissa, 
> >>>> 
> >>>> After using the ScanR importer with some non-captured wells, I realized that the names were incorrect.
> >>>> I did the following patch to fix the problem. 
> >>>> 
> >>>> This is not preventing the converter to create more metadata and images than actually needed, but is already a help.
> >>>> 
> >>>> Thanks!
> >>>> 
> >>>> Rubén
> >>>> 
> >>>> 337a338,339
> >>>>>   String [] names = new String [nWells * nPos];
> >>>>> 
> >>>> 361a364
> >>>>>                 names[realPosCount] = "Well " + (well + 1) + ", Field " + (pos +1);
> >>>> 468,470c471
> >>>> <       String name = "Well " + (well + 1) + ", Field " + (field +1) + " (Spot " +
> >>>> <         (i + 1) + ")";
> >>>> <       store.setImageName(name, i);
> >>>> ---
> >>>>>     store.setImageName(names[i], i);
> >> 
>