reorganize by lang-code

[JarbasAl]

i have been recording a few samples for "hey computer", the thing is that in portuguese the "hey" can be lots of different things, things fit much more nicely together if we organize the repo like this

lang-code/wake_word works with word variations much more naturally than wake_word/lang-code

using "hey computer" has an example, in general you would say "ei computador", but in northern portugal you are more likely to say "ó computador" while in brazil you might say "oi computador", this is not a matter of accent, those are different words, and the data needs to be labelled better to take this into account

[el-tocino]

Based on your example above, and other thought, closing this. Neither your change or the current method handles those, but a lang-wakeword or wakeword-lang format would.

[JarbasAl]

Based on your example above, and other thought, closing this. Neither your change or the current method handles those, but a lang-wakeword or wakeword-lang format would.

i don't understand, what this does is lang-wakeword, my example shows what it is solving with the hey_computer word, lang-wakeword allows to do it properly

[JarbasAl]

wakeword_lang would not make much sense i think,

hey_computer
   - pt
      - 3 folder here for variations
   - en
       - hey computer

lang_wakeword (this PR) looks better IMHO

pt
  - variation 1
  - variation 2
  - varation 3
en
 - hey computer