A Few Notes on Modifying Koha for Use as a Journal Catalog

A barebones Web page with barebones content.

With better formatting and more content to be added later.

Post MAC-MLA note: thanks to all who looked at my poster. The PDF of the poster is linked below. It's only one megabyte, but the page size is 3x4 feet.

I posted this to the Koha-Community list a while back. Most of the responses indicated that people were interested in the RIS to MARC mapping, and in seeing screen shots of what the modified OPAC display looks like, so I've uploaded those first. If you want to skip the original very long post and go directly to the downloads, click here.

A couple of months ago, I brought up the topic of using Koha as a catalog of journal articles. I've been working on this ever since, and I finally have our system up and running. Unfortunately, it's behind a firewall so I can't give you a link, but I can tell you about it and post screen shots if there's any interest.

I work in a teaching hospital, and a lot of our residents and staff publish articles, most of which are indexed in Medline. For years, I've been collecting citations and putting them together on a web page. Several years ago, we started Authors Day, where we recognize the literary efforts of our associates. We also put together a bibliography that includes the Medline citations as well as conference abstracts, poster presentations, books (editor, chapter author, or full book)--basically anything we know about, and not limited to medical literature. Last year we started using Reference Manager to track citations. I wanted to see if I could use Koha to make the publications more accessible within the hospital.

My conclusion: yes. It takes a lot of work to set up, and it takes more steps than I'd like to convert Medline records to (my modified version of) MARC, but the end product is a lot more versatile than the bare Medline citations with local authors' names highlighted, which is what we had.

A lot of the constraints I faced are specific to our situation and type of library, so YMMV. In particular:

  1. Most of the citations are from Medline.
  2. Most articles have a lot of co-authors, not all of whom are local.
  3. The citations had to be formatted for Koha and Reference Manager.
  4. I wanted to make authors' names uniform; Medline citations don't have authority control.
  5. I wanted to include the authors' departments.

We're using Koha 3.12, running Ubuntu.

There are two main parts to setting this up: modifying Koha templates to display the information I wanted, and modifying citations to fit Koha.

Modifying citations: I use Ovid to export the Medline citations in RIS format. PubMed and Ebsco will also export tagged citations--anything with field tags will do. RIS is a format used by Reference Manager, EndNote, Zotero, and a few others. The main (hard!) part was coming up with an appropriate MARC tag to match the field tag. Since I envision this being for institution-wide use only, I didn't worry too much about straying slightly from MARC standards. One of my colleagues is entering conference presentations, etc., into Reference Manager; as soon as I can get my head above water, we'll convert those tags to MARC, too.

Modifying Koha: I modified the OPACXSLTDetailsDisplay and OPACXSLTResultsDisplay templates to show just the fields I wanted. I didn't need to display item/barcode information, so I took out the item display at the bottom. I put in the PMID (every article in Medline has a unique identifier; I used MARC 099), MeSH (MEdical Subject Heading, 650), department (I used 656), publication type (clinical trial, editorial, randomized controlled trial, etc., 657), abstract (520), and a few others.

General Koha Setup: Right now, I have only one hospital's authors in the database, so that's the only branch. If this catches on, we can add others. For Item Types I used Journal Article, Book (Chapter Author), Book (Editor), Book (Author), Poster, and a few others. That's about all I had to do.

Adding Citations: As I said, more steps than I'd like. Here's an outline.

1. Determine which citations are local authors. I've already written too much, so I'm going to leave this part out. Let's just assume I end up with a few citations to add, with all the fields I want to include in both Reference Manager and Koha. I've also added links to the full text of the articles, if available.

2. Standardize the authors' names and add their departments. Medline/RIS has two fields for the authors' names: standard AU (Lastname FirstIntitalSecondInitial) and full FA (Lastname, Firstname MiddleInitial). I've already taken out non-local authors from the FA list. I use Replace Text (free from www.ecobyte.com) to search for the name as it appears in the FA field and replace it with the name and department (on a separate line, with a separate tag). If Replace Text doesn't find a match, it puts NOT FOUND by the author's name. I search through the results for NOT FOUND and either add the author and department, or add the variation of the author's name. For example, Dr. William J Rubin of the Department of Hematology might be listed in the Medline FA field as
Rubin, William J
or Rubin, Wm James
or Rubin, Billy

and by editing the Replace Text file, all of those names will be converted to
Rubin, William J
Hematology

complete with RIS field tags for name and department ready to upload. Adding new names and variations to the Replace Text file takes a while, but I think it will become faster as I catch all the missing authors and name variations.

3. Import the edited file into Reference Manager. We can use this database to print bibliographies by name, department, etc. It's not necessary if all you need is a catalog.

4. Convert RIS tags to MARC. Replace Text again.

5. Remove duplicate department tags. In step 2 above, I replaced each name with the author's name and department. However, a lot of authors are from the same department. Uploading a MARC record with three instances of Hematology will result in a catalog display with three instances of Hematology as well. I want each department mentioned only once, so I have to go through each citation and take out duplicates. There's probably an easier way to do this.

6. Compile the citations into a .mrc file with MarcEdit.

7. Upload into Koha.

A few thoughts:

The most important part is to think everything out beforehand. No, this wasn't meant to be surprising news. I thought out most things, but not everything. I discovered that modifing 360 records one-by-one takes a while. No, that's not meant to be surprising either.

Was it worth the time and effort to develop? I think it was. We needed a database of local authors anyway, and the steps to convert from RIS to Koha don't add much more to the process. Setting up a separate instance of Koha, modifying it to display what I wanted, and developing the RIS to MARC tags took a lot of time, but once that's done, it's fairly simple.

Why modify Koha instead of using some other system? I didn't want our users to have to learn another system. Neither did I. Koha is easy to use, and we're using it for our library catalog already.

Will anyone use it? I hope so. :-)

If anyone has made it this far and would like to hear more, I'll be glad to provide more information. Thoughts, improvements, etc., are all welcome. Please e-mail kohauser at phred dot us. (Address corrected on 26 December 2014.)

The Downloads: