This repo should build metadata for a HathiTrust bookworm from original MARC records. It builds off of the pymarc module to pull metadata likely to be useful in analysis. This includes:
- Dates
- Author information
- Classification information
- Titles
- Language information
- Holding library information
- Scanner information.
The heart of the module is the new BRecord class in bookwormMARC/bookwormMARC.py.
This extends the existing pymarc.Record class with several new methods analogous to
the original Record.title() that just parses out the first title field.
Certain MARC fields, like 100 (creator) have so much useful information that rather than overload the main
class, I've created a few new classes (Author for field 100; F008 for field 008) that can return a little dictionary
suitable for tacking onto the main record.
At a first pass, I'm opting to create methods for the first author, date, etc.; although MARC allows for multiple fields in all of these, Bookworm queries frequently make more sense with a single location. This will likely require some refactoring later.
There are a number of particularities around Hathi MARC records that I (Ben) don't yet fully understand. In particular, this relies a lot on the 974 field.
The files in bookwormMARC should be generalizable classes for reading in MARC records, although they may have hathi-specific methods. I want this code to be portable to at least the Medical Heritage library.
Done right, this could be a quick and easy way to take any library catalog and build a Bookworm around at least the metadata. That would be nice to have.