Encoding and recoding

[pwalsh]

See openspending/openspending#897

We need the ability to recode a file to utf-8, creating this file as a side effect of a recode processor.

Internally, Tabulator is recoding as utf-8 anyway, as goodtables is.

I'm not sure exactly the steps required: maybe there are alternatives to having a dedicated processor to produce the output file, but at a high-level view it seems reasonable.

I guess the spec looks something like this:

• input file has some encoding
• recode processor can create an output file, at a passed location, in any desired encoding, with utf-8 as default.

[roll]

WILL BE FIXED in #36

[roll]

Based on #36 discussion closing for now because it's definitely will be covered by write mode. Also added as write mode issue todo task.