[ome-devel] OMEIS Content header validity

Thomas Kuo thekuo at ece.ucsb.edu
Tue Feb 28 19:45:23 GMT 2006


I am using the url of an image ".../cgi-bin/omeis?Method=ReadFile&FileID=622&fake=" to "download" the image for an ImageJ applet, so this is done through OMEIS's ReadFile.

ImageJ's built-in URL reader reads the end of the URL to determine the file type, which is invalid with the above URL.  So I am looking at ways to use Java and ImageJ to do this correctly.  Among the ideas, I can read the filename from the header and determine the type, as you suggest, or use Java's included functions to extract the file type.

However, I was wondering if it made sense for OMEIS or something else within OME to determine the file type beyond the generic octet-stream (such as image/tiff, etc.) and send that as the content-type header.

~Thomas


Ilya Goldberg wrote:
> When you say "download images", are you referring to OriginalFiles 
> (the OME object which corresponds to the File OMEIS object)?
> I think raw pixel data (GetPixel*) and the ReadFile OMEIS methods are 
> the only ones that return application/octet-stream.  If you ask OMEIS 
> for a rendered image, it returns the correct MIME type.  This is 
> because application/octet-stream is the right thing for pixel data, 
> and OMEIS doesn't know anything about file types (they're all 
> application/octet-stream as far as OMEIS knows).
>
> It would be somewhat of a challenge to do this right.  Possibly with 
> libmagic, possibly with the unix command "file".  A very poor way of 
> doing this is how its done in Windows - by looking at the file 
> extension (the name of the file is stored by OMEIS).
>
> Can you give some details what you want to achieve by this?  Afterall, 
> even though it downloads as an application/octet-stream, the OS will 
> surely know what kind of file it is once you try to do something with it.
> -Ilya
>
>
> On Feb 28, 2006, at 12:51 PM, Thomas Kuo wrote:
>
>> Hello,
>>
>> I've been trying to download images from the OME database; however, the
>> content-type or MIME type is currently set to application/octet-stream
>> regardless of type.  I think it should be set according to the data that
>> is present, i.e. image/tiff for tiff images, etc.  This may be fairly
>> easily done using a program like libmagic, that extracts the type from
>> the header information.
>>
>> ~Thomas
>>
>> --Thomas Kuo
>> Graduate Student
>> Electrical & Computer Engineering Department
>> University of California Santa Barbara
>> Email: thekuo at umail.ucsb.edu
>>
>> _______________________________________________
>> ome-devel mailing list
>> ome-devel at lists.openmicroscopy.org.uk
>> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel
>>


-- 
Thomas Kuo
Graduate Student
Electrical & Computer Engineering Department
University of California Santa Barbara
Email: thekuo at umail.ucsb.edu



More information about the ome-devel mailing list