Supply missing language identifiers to code blocks

[guardrex]

[guardrex]

Update: Azure ML is quite a wonderful technology; however like all statistical approaches, it's only as powerful as the data fed into it. Setting it up and evaluating various classification models went well. The problem is that training the models with the existing code block data is insufficient for producing any decent predictive model that we can use.

I'm going to Plan B.

[mairaw]

is this done @guardrex?