RLE Workflow (submission package creation and processing steps)

6-08-2010 RLE Progress Reports

DC files for items in each Progress Report supplied by Rob Wolfe's metadata team and bitstreams consisting of PDF's and .tiff files were scanned and produced by Document Services.

Steps to create and submit Research Library For Electronic SIPs:

1. Create new collection and administrative and permissions for  current Progress Report using dspace.mit.edu web interface.  As it stands now, creating the collection and assigning permissions is probably the most time-consuming part of the process.  I looked into possibility for automating the collection creation process, but unfortunately there are no dspace utilities that do quite what we need at this time.

2. Run ruby script, rle.rb , to clean SIP directory:  'rle.rb -c'  ( clean up files from preceding collection load )

3. Create new SIPs for specified collection: 'rle.rb -d 081'  (create new SIPs for Report 081)

4. Check  hard drive /Volumes/MyBook/RLE/sip directory to make sure new content is there

5. secure copy newly created SIPs to dspace /home/carlj/upload directory:  scp -R -c * dspace.mit.edu:upload/

6.  Login to dspace.mit.edu

7.  On dspace.mit.edu run shell script, rle_import.sh, in test mode:

              /home/carlj/bin/rle_import.sh -u 55434 -t

8. Look a the metadata as it scrolls by in test mode, make sure year and report number are correct

9. Run rle_import.sh scrip† in 'real' mode: e.g. /home/carlj/bin/rle_import.sh -u 55434

        - Wait for ItemImport to finish (depends on how many items to load in each report)

        - If ItemImport stops with an error – usually because of the assetstore file creation issues (see Alex Brennan for more information) -- then 'resume' the load with rle_resume.sh script one or more times: /home/carlj/bin/rle_resume.sh -u 55434

10. Refresh current RLE collection page and check to see that the new items look OK; collection name is spelled correctly, correct year, etc.   Check at least one item (click and navigate through brief and full metadata display screens, view a .pdf).   For example, Progress Report No. 081 (1966): http://dspace.mit.edu/handle/1721.1/55434

11.  Clean up dspace.mit.edu /home/carlj/upload directory:  /home/carlj/bin/rle_import.sh -k

Summing up, it takes approximately 15 (no resume operations) to  20+ (if we need to run rle_resume.sh) minutes per collection. I think a reasonable average, then, of 20 minutes per report is about right.