Google's Natural Language API, Content Classification, in C#

#csharp #naturallanguage #classifytext #api

I've written about this before, in the context of ASP.NET. Here's a link to the revised project on Github from which a DLL may be derived.

We started off with classifying Html and went on to classifying Text. Thus we have

Classify()
ClassifyHtml()
ClassifyText() which emit JSON.

Recently we added XML serialization and the functions

ClassifyHtmlAsXml()
ClassifyTextAsXml()

Both serialization methods emit an Error/Cargo tuple. If Error is null, then there's no error and Cargo contains the result of the classification. If Error is a string, then there is an error (in Error) and Cargo is null/indefinite.

Below is from a REPL session in one of our javascript-on-steroids tools, demonstrating the outputs of the DLL:

>attach("Classification.DLL","C")
Attached C
Microsoft.ClearScript.VoidResult

>C.Classification.Classifier.ClassifyHtml(CS.System.IO.File.ReadAllText("https_www.sil.org.html"))
{"Error":null,"Cargo":[{"Name":"/Reference/Language Resources","Confidence":0.6}]}

>C.Classification.Classifier.ClassifyHtmlAsXml(CS.System.IO.File.ReadAllText("https_www.sil.org.html"))
<?xml version="1.0" encoding="utf-16"?>
<Classification xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema">
  <Cargo>
    <Cargo>
      <Name>/Reference/Language Resources</Name>
      <Confidence>0.6</Confidence>
    </Cargo>
  </Cargo>
</Classification>