Thumbnail Viewing -- Is this possible

[thaneshk]

Suppose we are using imagemagick to convert an image to a thumbnail, is this possible.

Could the image be converted and stored directly into the repository and file system, and in otherwords how do we manipulate the system so that we dont have to create thumbnail, then store it in repository and then delete the temporary file that was in the local file system and not the alfresco file system.

See my point?? If not let me be abit clear.

When you want to create a thumbnail u use imconvert in the fiollowing way – imcovert image.jpeg -option image_thumbnail,jpeg

so now we have two files image.jpeg and image_thumbnail.jpeg.

My question is whether this image_thumbnail.jpeg can be directly stored into the alfresco file system and repository without the need of having to create a copy, upload it to alfresco system and then delete the temprary file created

See my point???

[kevinr]

The Alfresco "Rules" functionality can be used to automatically generate thumbnails and add them to the repository. A couple of temporary files are created when imagemagick runs, but they are cleaned up by Alfresco.

You can currently use Templates to display those thumbnails as a dashboard view.

We do not yet have a dedicated Thumbnails view, but as i say it is easy enough to use Alfresco to generate them.

Here is an example using rules to generate thumbnails:
http://www.alfresco.org/forums/viewtopic.php?p=4301

Thanks,

Kevin

[derek]

Hi,

I see your point.  The answer is "unfortunately not".  Consider the nature of a ContentTransformer.  It has to transform content from a ReadableByteChannel and output it to a WritableByteChannel.  Some transformers, especially those that run in a process outside of our system, cannot deal directly with streams.  This is true of the ImageMagick convert executable that we use.  Other transformers (ImageMagick and OpenOffice) require files with well-known extensions.  But we only have streams and are therefore compelled to create temporary files.

So why not just pass files in and out of the ContentTransformer?  Because the content storage system is designed to be extendable.  We can bolt on a database store, a tape-drive store, a zip-file store and any other number of remote content stores.  The lowest common denominator is an IO channel.

Regards