Reading styles

Parse the text

ParseTextWithAnsiStyles returns an AnsiText: the text without its escape sequences, and the runs of text that share the same style:

var text = "\x1b[1;38;5;208mWarning\x1b[0m and \x1b[4;34mInfo\x1b[0m";
var parsed = AnsiTextProcessor.ParseTextWithAnsiStyles(text);

Assert.Equal("Warning and Info", parsed.Text);

var runs = parsed.Runs.Select(run => $"{run.Start}-{run.End}: Bold={run.Style.Bold}, Underline={run.Style.Underline}");
Assert.Equal(
    [
        "0-7: Bold=True, Underline=False",
        "7-12: Bold=False, Underline=False",
        "12-16: Bold=False, Underline=True",
    ],
    runs);

• Text is the text without the escape sequences.
• Runs is a list of AnsiTextRun. Each run has a Start (inclusive) and an End (exclusive) index in Text, and an AnsiStyle. The runs follow each other, without gap or overlap, and cover the whole text. A run starts each time the style changes; an empty text has no run.
• The text that has no style is a run whose style is AnsiStyle.None.
• ParseTextWithAnsiStyles throws an ArgumentNullException for null.

Styles

AnsiStyle is a record with Bold, Italic, Underline and Inverse, and a Foreground and a Background color, null for the default color of the terminal. Two styles are equal when all their values are equal.

The SGR codes it reads are:

The extended colors are also read in their colon form, such as 38:5:208 or 38:2::255:128:0. Codes that are not in the table, such as blinking or faint text, are ignored, and a parameter that cannot be read does not change the style.

Colors

AnsiColor is a record. Its Kind tells how to read it:

• AnsiColorKind.Indexed: IndexedValue is an index in the 256-color palette. Indexes 0-15 are the standard colors (black, red, green, yellow, blue, magenta, cyan, white, then their bright variants), whose exact RGB values depend on the terminal theme.
• AnsiColorKind.Rgb: Red, Green and Blue are the components of a 24-bit color.

AnsiColor.FromIndexed(index) and AnsiColor.FromRgb(red, green, blue) create colors to compare with:

var parsed = AnsiTextProcessor.ParseTextWithAnsiStyles(
    "\u001b[31mred\u001b[0m \u001b[38;2;255;128;0morange\u001b[0m \u001b[1;97;44mwhite on blue\u001b[0m");

Assert.Equal("red orange white on blue", parsed.Text);
Assert.HasCount(5, parsed.Runs);

// 30-37 and 90-97 are the 16 standard colors: indexes 0-7 and 8-15
var red = parsed.Runs[0];
Assert.Equal("red", parsed.Text[red.Start..red.End]);
Assert.Equal(AnsiTextProcessor.AnsiColor.FromIndexed(1), red.Style.Foreground);

// The space between two styled words is a run without style
Assert.Equal(AnsiTextProcessor.AnsiStyle.None, parsed.Runs[1].Style);

// 38;2;r;g;b is a 24-bit color
var orange = parsed.Runs[2];
Assert.Equal(AnsiTextProcessor.AnsiColor.FromRgb(255, 128, 0), orange.Style.Foreground);
Assert.Equal(AnsiTextProcessor.AnsiColorKind.Rgb, orange.Style.Foreground?.Kind);

var whiteOnBlue = parsed.Runs[4].Style;
Assert.True(whiteOnBlue.Bold);
Assert.Equal(AnsiTextProcessor.AnsiColor.FromIndexed(15), whiteOnBlue.Foreground);
Assert.Equal(AnsiTextProcessor.AnsiColor.FromIndexed(4), whiteOnBlue.Background);

Convert to HTML

The runs map directly to HTML elements. This function writes bold, underlined and RGB-colored text; extend it with the other attributes and a palette for the indexed colors:

static string ToHtml(string ansiText)
{
    var parsed = AnsiTextProcessor.ParseTextWithAnsiStyles(ansiText);
    var html = new StringBuilder();
    foreach (var run in parsed.Runs)
    {
        var text = WebUtility.HtmlEncode(parsed.Text[run.Start..run.End]);
        var style = run.Style;
        if (style.Bold)
            text = "<strong>" + text + "</strong>";

        if (style.Underline)
            text = "<u>" + text + "</u>";

        if (style.Foreground is { Kind: AnsiTextProcessor.AnsiColorKind.Rgb } color)
            text = "<span style=\"color:#" + Convert.ToHexStringLower([color.Red, color.Green, color.Blue]) + "\">" + text + "</span>";

        html.Append(text);
    }

    return html.ToString();
}

Assert.Equal(
    "<strong>Error:</strong> <span style=\"color:#ff8000\">disk &lt;C:&gt; is full</span>",
    ToHtml("\u001b[1mError:\u001b[0m \u001b[38;2;255;128;0mdisk <C:> is full\u001b[0m"));

Encode the text of each run, as WebUtility.HtmlEncode does here: the output of a process can contain < and &.