[ome-devel] MetaVue 7.7.9.0 timestamps not read correctly

[Christoffer Aberg]

On 30 March 2015 at 15:44, Melissa Linkert <melissa at glencoesoftware.com> wrote:
> Hi Christoffer,
>
>> First of all, apologise if this has been addressed before, but I couldn't
>> find anything reported on this topic prior.
>>
>> I am trying to use the BioFormats library to read the timestamps from a
>> file produced by MetaVue version 7.7.0.0. I am (currently) using the
>> BioFormats Importer in Fiji for this purpose, but that should not be source
>> of (at least not both of) my problems.
>
> Thank you for reporting these problems.
>
>> 1. The first issue I am facing is that a time-stamp which should be
>> "2015-03-12 16:43:45:753" is returned as "1426178625753" when viewing the
>> "Original Metadata". I am not sure if that is an ImageJ/Fiji issue or stems
>> from the BioFormats library. In the actual TIF file, it is correctly
>> written as "20150312 16:43:45.753" which is also what I can see when
>> opening the file in the original MetaVue software.
>
> Is the "1426178625753" value recorded in the "timestamp" field?  If so,
> that is expected; the nicely formatted date should be present in an
> "acquisition-time-local" and/or "DateTime" entry.

Yes. My main problem was that I don't know how to read it. I have
since realised that it's simply "Unix time" with added ms.

>> 2. My second, and more important, issue is that the milliseconds part of
>> the time-stamps are incorrectly read. This is due to the strange way the
>> time-stamps are stored in the file, whereas the BioFormats library attempts
>> to read them in a more "sane" way.
>>
>> The problem is the following: While the date-time March 12 2015
>> 16:43:45:753 is stored in the file as "2015-03-12 16:43:45:753", the
>> date-time March 12 2015 16:43:46:003 is stored as "2015-03-12 16:43:45:03".
>> That is, notice that "003" is just stored as "03". I have confirmed this
>> with several different time-points, looking at what is written in the file
>> and comparing with what I can see in the original MetaVue software.
>>
>> The solution is, thus, simply to pad the millisecond field with zeros from
>> the left, rather than from the right. This should be a very simple
>> correction and I would do it myself, were it not for that I do not write
>> Java and I am afraid to break something which is otherwise working
>> (especially given that this does not seem to have been reported before).
>
> A ticket for this is now open on our issue tracking system:
>
> http://trac.openmicroscopy.org/ome/ticket/12810
>
> You will be automatically notified of updates to the ticket; if you
> prefer not to receive updates, please let us know.

That's fine. Please let me know if I can be of more help. As I said, I
would be able to supply small test files for unittests which expose
the bug etc.

Thanks!

Christoffer

>
> Regards,
> -Melissa
>
> On Thu, Mar 26, 2015 at 05:30:42PM +0100, Christoffer Aberg wrote:
>> Dear all,
>>
>> First of all, apologise if this has been addressed before, but I couldn't
>> find anything reported on this topic prior.
>>
>> I am trying to use the BioFormats library to read the timestamps from a
>> file produced by MetaVue version 7.7.0.0. I am (currently) using the
>> BioFormats Importer in Fiji for this purpose, but that should not be source
>> of (at least not both of) my problems.
>>
>> 1. The first issue I am facing is that a time-stamp which should be
>> "2015-03-12 16:43:45:753" is returned as "1426178625753" when viewing the
>> "Original Metadata". I am not sure if that is an ImageJ/Fiji issue or stems
>> from the BioFormats library. In the actual TIF file, it is correctly
>> written as "20150312 16:43:45.753" which is also what I can see when
>> opening the file in the original MetaVue software.
>>
>> 2. My second, and more important, issue is that the milliseconds part of
>> the time-stamps are incorrectly read. This is due to the strange way the
>> time-stamps are stored in the file, whereas the BioFormats library attempts
>> to read them in a more "sane" way.
>>
>> The problem is the following: While the date-time March 12 2015
>> 16:43:45:753 is stored in the file as "2015-03-12 16:43:45:753", the
>> date-time March 12 2015 16:43:46:003 is stored as "2015-03-12 16:43:45:03".
>> That is, notice that "003" is just stored as "03". I have confirmed this
>> with several different time-points, looking at what is written in the file
>> and comparing with what I can see in the original MetaVue software.
>>
>> The solution is, thus, simply to pad the millisecond field with zeros from
>> the left, rather than from the right. This should be a very simple
>> correction and I would do it myself, were it not for that I do not write
>> Java and I am afraid to break something which is otherwise working
>> (especially given that this does not seem to have been reported before).
>>
>> I am more than willing to help out with further tests, should they be
>> needed, and also to supply test image files with known metadata to improve
>> the fidelity with which the library reads these files.
>>
>> Best,
>> Christoffer
>
>> _______________________________________________
>> ome-devel mailing list
>> ome-devel at lists.openmicroscopy.org.uk
>> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel
>