JSON, XML and HTML values

JSON, XML or HTML embedded in a value is hard to read and compare on one line. Formatters rewrite it: they apply to JsonNode, JsonDocument and JsonElement values, to XmlNode and XElement values, and to the content of HTTP messages, according to its media type.

JSON and XML

var options = new HumanReadableSerializerOptions()
    .AddJsonFormatter(new JsonFormatterOptions { OrderProperties = true })
    .AddXmlFormatter(new XmlFormatterOptions { OrderAttributes = true });

var value = new
{
    Json = JsonNode.Parse("""{"name":"Ada","id":42}"""),
    Xml = XElement.Parse("""<user name="Ada" id="42"><role>admin</role></user>"""),
};

Assert.Equal("""
    Json:
      {
        "id": 42,
        "name": "Ada"
      }
    Xml:
      <user id="42" name="Ada">
        <role>admin</role>
      </user>
    """, HumanReadableSerializer.Serialize(value, options));

Without a formatter, the value is written as it is. A value that cannot be parsed, such as invalid JSON, is written as it is too.

FormatAsStandardObject writes the JSON as the serializer writes objects, without braces and quotes:

var options = new HumanReadableSerializerOptions()
    .AddJsonFormatter(new JsonFormatterOptions { FormatAsStandardObject = true });

var value = new { Json = JsonNode.Parse("""{"name":"Ada","roles":["admin","dev"]}""") };

Assert.Equal("""
    Json:
      name: Ada
      roles:
        - admin
        - dev
    """, HumanReadableSerializer.Serialize(value, options));

HTML

The Meziantou.Framework.HumanReadableSerializer.Html package adds a formatter for text/html values:

dotnet package add Meziantou.Framework.HumanReadableSerializer.Html

using var content = new StringContent(
    """<p id='greeting' class=intro nonce=r4nd0m>Hello<br/>world &amp; all</p>""",
    encoding: null,
    "text/html");

var options = new HumanReadableSerializerOptions()
    .AddHtmlFormatter(new HtmlFormatterOptions
    {
        OrderAttributes = true,
        RedactContentSecurityPolicyNonce = true,
    });

Assert.Equal("""
    Headers:
      Content-Type: text/html; charset=utf-8
    Value: <p class="intro" id="greeting" nonce="[redacted]">Hello<br>world &amp; all</p>
    """, HumanReadableSerializer.Serialize(content, options));

The value is parsed by the HTML parser of Meziantou.Framework.WebPlatform, which follows the HTML specification, and serialized back, so the output is normalized: attribute values are quoted, character references are decoded and escaped again, and <br/> becomes <br>. Two HTML values that differ only by these details give the same text.

A value that contains a doctype or an html, head or body tag is parsed as a document, and keeps these elements:

using var content = new StringContent("""<!DOCTYPE html><html lang=en><body><a href="/docs">Docs</a></body></html>""", encoding: null, "text/html");

var options = new HumanReadableSerializerOptions()
    .AddHtmlFormatter(new HtmlFormatterOptions { AttributeQuote = HtmlAttributeQuote.SimpleQuote });

// A document keeps its doctype, html, head and body elements
Assert.Equal("""
    Headers:
      Content-Type: text/html; charset=utf-8
    Value: <!DOCTYPE html><html lang='en'><head></head><body><a href='/docs'>Docs</a></body></html>
    """, HumanReadableSerializer.Serialize(content, options));

Any other value is parsed as the content of a body element, and written without the elements the parser adds around it. A value that cannot be formatted is written as it is.

Other media types

AddFormatter(mediaType, formatter) registers a ValueFormatter for any media type. A formatter overrides Format(writer, value, options) and writes the value with the HumanReadableTextWriter. The ValueFormatter constants name the media types the serializer recognizes: JsonMediaTypeName, XmlMediaTypeName, HtmlMediaTypeName, WwwFormUrlEncodedMediaTypeName, CssMediaTypeName and JavascriptMediaTypeName.