[ome-users] searching bug

[Will Moore]

Hi Martin,

 I thought I'd try to explore the "Metadata Search" you described below, so I've written a script that takes a few search parameters, builds a database query and then tags the resulting images. Script is here https://github.com/will-moore/scripts/blob/metadata_search/omero/search_scripts/Metadata_Search.py

See screenshots below for the script search form and the results.

You can search for any combination of search fields. When the script has finished, refresh your Tags view and you'll see a 
"Search Results <time-stamp>" Tag with the resulting images. NB: In the right panel you'll notice that the description of the Tag
has a summary of what you searched for. 

Have a play with it and let us know what you think (E.g. what other search fields do we need?).

Once we have something we're happy with, we can implement it in the clients "properly".

 Cheers,

   Will. 
 






On 20 Feb 2013, at 10:48, Martin Spitaler wrote:

> Thanks, Will. 
> 
>  Ideas for missing search features: First, some traditional functions:
> IMAGE TYPE, e.g. time series, stack, multi-channel...; maybe even specifying a specific range (number of channels or time points)
> METADATA SEARCH, e.g. a specific microscope, lens or excitation laser; would also help with the above (e.g. search for "confocal images with the 2 channels GFP+DAPI"
> ADVANCED SEARCH, i.e. the user can search for different things in different fields, PUBMED-style
> And then an idea that came from one of our users, saying the typical situation is having an image and wanting to find related ones, but not remembering how to search for them (I seem to remember this has already been discussed at OME meetings):
> SIMILAR IMAGES: some of this functionality would be covered by METADATA and IMAGE TYPE search above, but an additional feature that sounds doable is mean and standard deviation of intensities
>  I know this last idea opens a whole new world of complexity, and also workload for servers if calculated automatically (although once done, it would only be two more numbers per frame to store), but maybe one or a few of the scripting aficionados could help with this? Would it be possible to run a script on a dataset calculating such values and adding them to the metadata? Long-term, the ideal situation would be to have a tick box on data import to do this calculation if wanted, and a script to do it later if not ticked on import.
> 
>  Obviously there is no limit to ideas, in an ideal world (now still dream land), analysis results could be linked to images and allow searches like "images with >100 cells" or "cells with 'roundness>1.5'. And in a world where photographic cameras take pictures when people smile, it's not inconceivable to have basic analysis like counting nuclei on import. But that's something after version 4.5 (and 5) - maybe someone in the community wants to have a play?
> 
> Good work,
> 
> Martin
> -- 
> Martin Spitaler, PhD
> FILM - Facility for Imaging by Light Microscopy
> - Facility Manager -
> Sir Alexander Fleming Building, desk 401
> Imperial College London / South Kensington
> Exhibition Road
> London SW7 2AZ
> UK
> 
> Tel. +44-(0)20-759-42023
> E-mail m.spitaler at imperial.ac.uk
> Website: http://imperial.ac.uk/imagingfacility
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openmicroscopy.org.uk/pipermail/ome-users/attachments/20130220/c8f1e710/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screen shot 2013-02-20 at 23.24.01.png
Type: image/png
Size: 77123 bytes
Desc: not available
URL: <http://lists.openmicroscopy.org.uk/pipermail/ome-users/attachments/20130220/c8f1e710/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screen shot 2013-02-20 at 23.31.00.png
Type: image/png
Size: 160814 bytes
Desc: not available
URL: <http://lists.openmicroscopy.org.uk/pipermail/ome-users/attachments/20130220/c8f1e710/attachment-0001.png>