Do not try to use ${type}_stable_id tables that do not any longer exist.

  Reformat.

croaking when '--(db)host' isn't passed, adding -D to the list of options recognized for --dbname (after MySQL), improving readability of the dba_opts data structure

Adding the method is_chromosome to the Slice object. This kind of method is not nice but does deal with a lot of issues with seq_regions being attached to chromosome coordinate systems which are infact unassembled regions.

Header function should not be responsbile for the addition of a > for the FASTA header. That's the serailisers responsiblity

Replacing the 6 coordinate methods with versions which can operate on transcripts which have not been persisted to a database. The DBID is still used for caching purposes

Switch to using $subsequence_slice than $self for the start and end values. If we don't then we will try to retrieve all repeat features for an entire chromosome.

Simplification of the AUTOLOAD since there were blocks of code which could never be called. Removed these unused blocks.

Escape _ in the input parameters as otherwise we spend a lot of time doing very expensive LIKE operations

(rather than 'not_biotypes'), for cleaner looking configuration.

  If the configuration parameter 'not_biotypes' is present, exclude all
  genes whose biotype is listed (but include all other genes).

  If the configuration parameter 'biotypes' is present, include all
  genes whose biotype is listed (as usual).

  It is an error to define both options in the configuration.

Only do block retrieval of repeats when we are caching our hits. Otherwise we're bringing back more data for no reason

added method fetch_all_by_display_label and changed fetch_by_display_label to return a gene on the reference slice if there is one