Join multi-line book titles with semicolon + <space>

[sethwoodworth]

For example:
GITenberg/Slave-Narratives--a-Folk-History-of-Slavery-in-the-United-StatesFlorida-Narratives_12297#1

[tfmorris]

Some versions of the script apparently included the carriage return as an HTML encoded in titles and repo names, so this needs to be cleaned up too.

[eshellman]

the problem noted by @tfmorris resulted in doubly-loaded texts. The current format_title method https://github.com/gitenberg-dev/gitberg/blob/master/gitenberg/book.py#L103 fixes this by (1) using a real xml parser to get titles from rdf and (2) uses unicodedata to intelligently asciify the title for the repo name. legacy repo names haven't been changed.