[ome-devel] Archiving Utility for Image Files

[T.J. Macura]

Hey Mike,

Good job on this cool new function. It is much appreciated because it  
is useful and we didn't have to develop it.

http://cvs.openmicroscopy.org.uk/horde/chora/diff.php?f=/OME/src/C/ 
omeis/omeis.c&r1=1.78&r2=1.79&ty=h&num=10

A very small point. Your commit to omeis.c changes alot of white-space  
around. When I look at the CVS web page, it makes it hard to find the  
meat. I always do cvs diff before I commit to check that every line  
changed is valid. This is an easy way to find vestigial debug  
statements.

I am not sure exactly what this function does, are you planning to  
write some web documentation about your current method?
http://www.openmicroscopy.org/api/omeis/methods.html

This is probably a file method. i am not sure about the name  
"getDownloadAllURL". Ilya is the final arbitrator but I would prefer to  
call it "GetFiles" (like GetPixels).

I think a cool test would be to write a script to randomly (a random  
number of random files) download zipped-up files, unzip them, import  
them back into OMEIS and see if a new linked FileID is created (that is  
basically a SHA1 digest test that the unzipped files are the same as  
the original files). Then DeleteFile the new files.

This is a very thorough test and will check many OMEIS components   
(Upload File, DeleteFile, and the clever symbolic-linking that goes  
behind the scenes).

I am very willing to collaborate on this test with a minimum  
involvement of your time.

Tom


On Nov 17, 2005, at 2:12 AM, Michael Hlavenka wrote:

> Hello,
>
> I corrected the space problem with archive.c (I thought I caught that,  
> but guess I forgot to add it in or something). I tested it with the  
> test images that were given and it worked fine. I'll throw a few more  
> tests at it tonight, but so far, it's been working fine.
>
> I also took your advice and put getDownloadAllURL in  
> OME/Image/Server.pm . That seems like a better spot for it.
>
> Let me know if it's still not doing what we want it to do. Thank you  
> for your input.
>
> ~Mike
>
> Josiah Johnston wrote:
>
>> Great job. That's been a long standing bug.
>> Chris or Ilya or Tomasz will have to comment on the changes to omeis.  
>> I haven't messed with it for a long while.
>> I found one bug in the system.
>> I imported the HTD images from the samples available on  
>> http://openmicroscopy.org/TestImages/
>> I looked at the first image, which was named "010B1-011B1 Morph Well  
>> P01 Site 2", then clicked "download them all at once"
>> I got an error.
>> I tracked the error to problems archive.c was having with spaces in  
>> the file name. As a work-around, I changed your getDownloadAllURL  
>> method to substitute spaces with underscores.
>>
>> After that, it seemed to work fine.
>>
>> I didn't notice any breakages, but I don't have a system set up to  
>> torture test omeis. I'm not sure OME/Web.pm is the best place for a  
>> function specific to images. I'm thinking either OME/Image/Server.pm  
>> or OME/Tasks/ImageManager.pm would be more appropriate. That's a  
>> minor point, all in all.
>>
>> Thanks for the contribution!
>>
>> -Josiah
>
>
> _______________________________________________
> ome-devel mailing list
> ome-devel at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel
>