Build a bookworm from the openly available Chronicling America data
This is a repo that contains code for building a Bookworm specifically on the Chronicling America
set from the library of congress. (Once it's fully complete), running make will download gigabytes of data,
install a bookworm instance, and parse and load it in. This will take a very long term: maybe, like, a week or so?
It will also work with the LOC data to build in test suites.
This could be considered the largest bookworm test suite: 6 million documents of fairly significant length. But it's not polite to download all those LOC data dumps more than once, everyone.