Libgutenberg

- test_orm now tests the orm
- removed stray savepoint that caused file addition to fail
- added tests for ORM add/delete file
- added a setter for dc.project_gutenberg_id to populate dc.canonical_url and dc.is_format_of
- added warning for bad filetypes
- updated logger syntax
- accommodate output dirs to be named cache1, cache2, etc for saving in db
- delint

- added GutenbergDublinCore.canonical_url to support https canonical urls distinct from the http URIs we're keeping so as not to break RDF links.
- updated RDFa style in html head elements from archaic to less archaic (DOCTYPE and schema declarations)
- use dublin core elements instead of dcterms where appropriate for modern metadata.

- fixed a spelling mistake
- added an is_audiobook method
- added a back relationship
- quelled scalar_subquery coercion warnings

This version is a significant change from previous version, adding an object relational mapping for the Project Gutenberg database and an ORM-based "Dublin Core" object, "DublinCoreObject" which replicates the "GutenbergDatabaseDublinCore" for most uses. because it is "lazy" - only making queries when needed - it may be much faster in many use cases, and about 20% slower when all attributes are accessed. This release is considered "alpha" with regard to the new objects, but should be production worthy and need no changes when the ORM based objects are not used.
- add Models.py and DublinCoreMapping modules.
- test suite greatly expanded, includes a comparative load test
-

- add SVG to mediatypes

- fix GutenbergGlobals.archive_dir with a special case for single digit book numbers