This directory contains all the SQL fragments needed to create a
Harvester database, and to upgrade it through its various versions.
Note that this is a database describing the harvesting jobs only: the
actual harvested data goes into a Solr database which is configured in
the separate but related "lui-solr" package.

* initial-schema.sql
	The script that created the original version of the
	database. The live databases have been updated from this
	beginning by the scripts in the v*.* directories (see below).

* v2.2, v2.3, v2.4, etc.
	Sets of scripts which have been used to cumulatively upgrade
	the original database (as described by localindices.sql) to
	its present form.

	Each of these directories in general contains several SQL
	files, each named for the date when it was added. In general,
	to upgrade to any given version of the schema from the
	previous one, it is necessary to run each of these files in
	order.

	The first two of these schema-upgrade directories, and only
	those two (v2.2 and v2.3) also include an "alter" script
	called alter_<VERSION>.sql. For v2.2, that script contains a
	duplicate of one of the two datestamped files but not the
	other, plus some other changes. For v2.3, the alter script is
	the concatenation of the two datestamped scripts. A best guess
	would be that v2.2 was supposed to work in the same way, but
	changes were made to the alter script. In any case, subsequent
	updates have not provided a single "alter" script at all, and
	this is a simpler and less error-prone approach.

* schema.v2.8.sql
	For convenience, a script to create version of the databases
	composed of localindices.sql plus all of the updates, bringing
	it up to v2.8 of the schema (which at the time of writing is
	the current version).

* schema.v2.8-with-sample-data.sql
	A modified version of schema.v2.8.sql that has been enhanced
	by the addition of INSERT statements to create a skeleton
	database. (Similar records are included in the original
	localindices.sql file, but were somehow omitted from
	schema.v2.8.sql. It's not clear how or why.)

* samples
	Snapshop dumps of complete harvester databases from live
	installations. These can be loaded into local development
	systems (but note that the schema changes through time, so
	older dumps may not work on newer systems).

* obsolete
	Due to erratic record-keeping in the early days of this
	project, the exact purpose of some of these fragments is not
	known, has had to be recovered by archaeology. Those fragments
	are all in the "obsolete" directory, and the README therein
	explains what is known.

The upshot of all this is that there are three ways to make a
database. The simplest and most reliable is probably using one of the
snapshots in the "samples" directory. Also simple is to use
schema.v2.8-with-sample-data.sql, but it's not completely clear how
that file came about and whether it's a 100% match for what is in
production. Finally, the least convenient approach, but perhaps the
most honest, is to run localindices.sql and all of the updates in the
v*.* directories.