Hi Rowan, I'm assuming in your example a DSpace item is a document bitstream plus multiple datastream bitstreams. You can have subdirectories in the itemimport dir structure, you just need to point to the files via the "contents" file. For example: item1/doc1 item1/mydir1/dataset1 item1/mydir1/dataset2 item1/mydir2/dataset1
and the contents file would be: doc1<tab>bundle:ORIGINAL mydir1/dataset1<tab>bundle:ORIGINAL mydir1/dataset2<tab>bundle:ORIGINAL mydir2/dataset1<tab>bundle:ORIGINAL Note in your example below you have spaces in dir names, not sure whether they'll cause a parsing problem or not. Note also the entire path will appear in the "Files in this Item" box on the item's metadata display page. Also, if you want the doc to appear first in the filelist on the metadata page, list it last in the contents file. Whether the dir names/paths should be replicated in metadata in some way might also be worth thinking about (e.g. maybe it's useful for a Manakin theme to take advantage of, may help in discovery, etc) Scott. > Message: 2 > Date: Mon, 8 Oct 2007 19:30:29 +1000 > From: "Rowan Brownlee" <[EMAIL PROTECTED]> > Subject: [Dspace-tech] importing a collection of files with identical > names (from different directories) > To: <dspace-tech@lists.sourceforge.net> > Message-ID: > <[EMAIL PROTECTED]> > Content-Type: text/plain; charset="iso-8859-1" > > I have a collection of document files sitting on a file server. Each document > is associated with a set of data files. The data files live in > subdirectories of the parent document file. Many of a document's data files > have identical names and are only differentiated by their position within the > directory structure. > > I wrote a python script which reads metadata from a spreadsheet, grabs the > associated document and data files and creates simple archive directories > suitable for import using itemimport. > > My understanding is that itemimport directories cannot contain > subdirectories. (i.e. I can't replicate the fileserver's document/data tree > in the dspace simple archive format). Is that correct? > > For any given document, I can't copy all of the data files to one directory. > The file names are identical so they end up overwriting each other. There's > also the problem of loss of contextual information. (For any data file, a > great deal of identifying information is contained in the name of its > directory. e.g. directory name "High rainfall maps", data file name > "Co_surface.jpg" / directory name "Quiescent maps", data file name > "Co_surface.jpg") > > I'm thinking that I'll need to prefix each of the data file names with their > original directory names, so as to retain the contextual information. I'm > fairly new to dspace so I'm wondering - are there other options? Eg. can I > associate a single metadata record with easily navigable related sets of data > files in a way that would make sense to a user approaching the default dspace > interface? > > How do others deal with this sort of issue? > > thanks > rowan > > Rowan Brownlee > Digital Project Analyst > University of Sydney Library > NSW 2006 Australia > Phone: +61 2 9036 6450 > Fax: +61 2 9351 3689 > http://escholarship.usyd.edu.au/dpa/ > > > -------------- next part -------------- > An HTML attachment was scrubbed... > > ------------------------------ > > ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ DSpace-tech mailing list DSpace-tech@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dspace-tech
