Add TextExtractor.Extract overloads for custom extraction results

[KevM]

It was requested to change the format of the current shape of the TextExtractionResults. Rather than break backwards compatibility I've introduced new overloads which allow the user to specify their own extraction results assemblers.

public class CustomResult
{
    public string Text { get; set; }
    public IDictionary<string, string[]> Metadata { get; set; }
}

public static CustomResult CreateCustomResult(string text, Metadata metadata)
{
    var metaDataDictionary = metadata.names().ToDictionary(name => name, metadata.getValues);

    return new CustomResult
    {
        Metadata = metaDataDictionary,
        Text = text,
    };
}

[Test]
public void should_extract_author_list_from_pdf()
{
    var textExtractionResult = new TextExtractor().Extract("file_with_authors.pdf", CreateCustomResult);

    textExtractionResult.Metadata["meta:author"].Should().ContainInOrder("Fred Jones, M. D.", "Donald Evans D. M.");
}

This will close #117 when it is merged.

[KevM]

Also closes #115. I needed to get the CI updated for our brave new world of 2018.

[KevM]

@bouletator let me know if this is good for you?

@TechnikEmpire care to also do a code review?

[KevM]

If no one complains I'll pull in this PR later today.