Add optional encoding argument to set output character encoding

[SimonBrazell]

What's this PR do?

Adds an optional encoding argument to #text, #html and .read for setting the output character encoding, which is passed to Tika as the --encoding option. This value is validated against Ruby's Encoding.name_list, raising an ArgumentError if it isn't included in the list.

Why is it needed?

So we can set the output encoding via Henkei / Tika instead of having to do it ourselves afterwards.

Where should the reviewer start?

• lib/henkei.rb:222

How should this be manually tested?

henkei = Henkei.new 'sample.pages'
utf_8_text = henkei.text(encoding: 'UTF-8')
utf_8_text.encoding 
=> #<Encoding:UTF-8>

[abrom]

capture2 uses write at this stage. It is also wrapped in a begin/rescue block catching the Errno::EPIPE error.

Looking at the documentation/source they do slightly different things.. including puts writing an extra newline

[abrom]

and it also sets up the read thread before writing to the input pipe

[SimonBrazell]

@abrom I modified the Open3.popen2 call to more closely match the capture2 source.

https://github.com/ruby/open3/blob/b8909222051b4103a19eba19506727faece252e7/lib/open3.rb#L775

[abrom]

Nice one @SimonBrazell 👍