Here's the note I got from Fitz about his scripts (my starting point).
Hi everyone,
I had to do some digging on through my backups, but I've resurrected the scripts I used to convert and ingest items into the AIMS application.
I've put the on github over here --> https://github.com/cfitz/aims_scripts
These probably need some explaining....
For the first set of objects that we used (the Gould files), I received a directory of files that Peter had run the FTK toolkit over and had also run a program (I think it was call Avantstar Transit?) that generated HTML files from source files. So, I had to make few scripts that worked with this data in order to get them into fedora.
They are:
html_to_csv.rb : This takes the FTK Bookmark html files that are exported. These bookmark files have technical metadata, as well as descriptive metadata following a scheme that Peter had set up in the FTK application. This script aggrigates all the bookmark HTML files and adds them to a CSV file.
reorg_directory.rb : This script takes the CSV file and makes new directories for each of the objects to be imported into Fedora. It also copies all the source files and Transit converted HTML file for each object into their respective directory.
convert_objects.rb: This script takes the Transit HTML file, converts it to a postscript file, convert the postscript file into a PDF, and multiple per-page jp2000 and text files. This script will require ImagicMagick (with jasper + jp2000 libraries installed), ps2pdf, and perl (for the html2ps script).
aims_ingestor.rb: This script ingest files in the directories into Fedora. Metadata from the CSV file is added into Fedora's XML. Originally, the data model was set so the source object gets it's own fedora object, while the text/jp2000s/pdf/ect files get added as datastreams in a additional "child" fedora object. This script assume that it's being run in the script directory of a rails application that has specific hydra models defined (AimsDocument).
That's about it. I've run these using Ruby REE and they seem to still work fine. The convert objects can take awhile (a few hours), especially if you're using a machine without much processing power.
Let me know if there are any problems or if you have any questions!
best,chris.
FYI - to stanford folks, there's a copy of the files Peter gave me lyberadmin@salt-dev:~/Chris_08_27/ . The file structure I think is an artifact of the Transit application.
