[ome-devel] exporting images from old version of OMERO

[Josh Moore]

On Fri, Sep 30, 2016 at 11:57 AM, Phil Chapman
<Phil.Chapman at cruk.manchester.ac.uk> wrote:
> Hi,

Hi Phil,


> Our institute has the commercial Columbus software which happens to use OMERO 4.4.11.10 as its image repository, and for this reason I was hoping that this group might be able to help.

a quick response from our side though suggestions from the community
are welcome.


> We are in the process of generating a large compound screening dataset (in the region of 100 384-well plates) and to analyse this data we want to make use of our 1000-core linux cluster that we typically use for NGS analysis rather than the 32 core cluster that Columbus is hosted on (we have estimated that analysing our dataset would take 3 days on this!!).  I have done some test exports from Columbus into the OME-TIFF format (using Columbus Helper), and have analysed data using the R EBImage package and this all works just fine.  However, since Columbus Helper is only available for Windows and Mac, I want to develop a programmatic way to do this transfer straight to the linux cluster.

You're primarily interested in the binary image data? If so,
`bin/omero export` may be a starting point in OMERO 4.4:

  https://github.com/openmicroscopy/openmicroscopy/blob/v.4.4.11/components/tools/OmeroPy/src/omero/plugins/export.py

which exports on an image-by-image basis. (There have been a number of
file leakage fixes since 4.4 which may or may not pertain. Keep an eye
on the open file handles for the server).


> I can connect to the OMERO repository just fine with OMERO.insight and view the data structure and see images, but the batch export functionality and any processing or exporting doesn’t work because it can’t find a processor to use – presumably it’s set up in such a way that OMERO only stores images rather than processes them?

This sounds like the "export and zip together" functionality for a
whole plate. Likely you don't particularly need the functionality of
batch exporting when working from a cluster (nor want the added burden
of storing the zips).


> Does anyone have any suggestions given that this is quite an old version of OMERO, and that the OMERO server doesn’t seem to have any compute resources.  If someone could point me in the right direction then that would help greatly!  I’m hoping there’s some sort of Python or Java API I can make use of maybe?

This is probably as far as we can go in supporting OMERO 4.4 on the
mailing lists. In addition, you may want to contact PerkinElmer
support before trying anything too significant. If you continue to
have issues, though, you're welcome to contact
info at glencoesoftware.com about dedicated support.


> Thanks in advance,
>
> Phil Chapman

All the best,
~Josh.