schemas/vegbien.sql: taxonpath: Renamed canon_id to canon_taxonpath_id to clarify that this is a recursive fkey. The convention is that a recursive fkey includes the table name plus a descriptive prefix.

schemas/filter_ERD.csv: Don't filter out fkeys from taxonpath to itself

schemas/vegbien.sql: taxonpath: Added canon_id for the canonical (scrubbed) taxonpath determined by TNRS

schemas/vegbien.sql: taxonpath: taxonpath_unique_within_datasource_by_name unique index: Added otherranks, so that ranks without a named column will be used in uniquely identifying the taxonpath

sql.py: DbConn.col_info(): Parse array types as sql_gen.ArrayType

sql_gen.py: EnsureNotNull: Support ArrayType types

strings.py: remove_prefix(), remove_suffix(): Added require param to raise aan exception if the string does not have the given prefix/suffix

sql.py: DbConn.col_info(): Moved parsing of user-defined datatypes to Python code, so that parsing for other composite types which also requires both data_type and udt_name can easily be added

sql_gen.py: Added ArrayType

schemas/vegbien.sql: Scope taxonpath instead of taxon with taxonscope, because a morphospecies name is specific to a datasource entity, so it should go in the datasource-specific taxonpath table instead of the datasource-general taxon table