Developed and built in Visual Studio 2017. Open evan_schiewe_entegra_challenge.sln and build the solution. 2 NuGet packages (IKVM and Standford.NLP.CoreNLP) should be pulled on the first build.
After building or downloading the release, evan_schiewe_entegra_challenge.exe can be run in a console without arguments. Test cases are included in Program.cs.
The parser can be used in other applications by using BusinessCard and adding BusinessCardParser.cs and ContactInfo.cs to your project. Currently, the name parsing assumes that the english.all.3class.distsim.crf.ser.gz model file is in the working directory.
- Because Stanford CoreNLP's Named Entity Recognition is used to extract names, instantiating a BusinessCardParser takes approximately 5 seconds while the model is loaded. After this, actual parsing performance is quite fast.
SLF4J: Failed to load...: This is a known issue with the Stanford.NLP.NET package, but thankfully doesn't get in the way. Removing the auto-generatedslf4j-simple.dllandslf4j-api.dllsilence this error.