/// <summary>
/// Parse the content, and detect all real lines breaks.
/// In fact, within markdown it is possible to write a unique paragraph, with line breaks.
/// If there is not a double space before the line break, it's the same paragraph, and the line break is not interpreted as is.
///
/// So in this parsing method, cut the content by "block" with same nature.
/// 1| Detect the paragraph type by the start of the line
/// 2| Process the content by its type
/// 3| Then remove the processed content, and continue to parse the next content
///
/// A block is detected by a <see cref="ParaPattern"/> which have an associated regex.
///
/// </summary>
/// <param name="engine">Describe an openXML object by code</param>
/// <param name="mdText">The markdown text to parse</param>
internal static void Parse(Md2MlEngine engine, string mdText)
{
while (!string.IsNullOrEmpty(mdText))
{
var firstLine = GetFirstLine(mdText);
var matchedPattern = PatternMatcher.GetMarkdownMatch(firstLine);
Paragraph para;
(int counter, string textBlock)rebuildText = default;
switch (matchedPattern.Key)
{
case ParaPattern.InfiniteHeading:
string titleChars = matchedPattern.Value.Groups[1].Value;
int titleLvl = titleChars.Count(c => c == '#');
titleLvl = titleLvl <= 9 ? titleLvl : 9;
para = engine.CreateParagraph(new ParaProperties()
{
StyleName = string.Concat("Heading", titleLvl.ToString())
});
engine.WriteText(para, matchedPattern.Value.Groups[2].Value);
mdText = DeleteLines(mdText);
continue;
case ParaPattern.Image:
var link = matchedPattern.Value.Groups[2].Value;
if (link.StartsWith("http://") || link.StartsWith("https://"))
{
engine.AddImage(new System.Net.WebClient().OpenRead(link));
}
else
{
engine.AddImage(ConvertRelativeToAbsolutePath(link, engine.GetFileDirectory()));
}
mdText = DeleteLines(mdText);
continue;
case ParaPattern.Table:
//case ParaPattern.TableHeaderSeparation:
rebuildText = BuildWithoutBreakingLines(mdText, ParaPattern.Table);
ProcessTable(engine, rebuildText.textBlock);
mdText = DeleteLines(mdText, rebuildText.counter);
continue;
case ParaPattern.OrderedList:
case ParaPattern.UnorderedList:
// TODO : Ordered list can contain unordered items inside and vice versa
rebuildText = BuildWithoutBreakingLines(mdText, ParaPattern.OrderedList);
ProcessBullets(engine, rebuildText.textBlock);
mdText = DeleteLines(mdText, rebuildText.counter);
continue;
case ParaPattern.CodeBlock:
// TODO : Improve the rendering - not a priority for my needs
rebuildText = BuildWithoutBreakingLines(mdText, ParaPattern.CodeBlock);
para = engine.CreateParagraph(new ParaProperties()
{
StyleName = DocStyles.CodeBlock.ToDescriptionString()
});
FormatText(engine, para, rebuildText.textBlock, new StyleProperties());
mdText = DeleteLines(mdText, rebuildText.counter);
continue;
case ParaPattern.Quote:
// Markdown supports nested quotes, but word does not
// So whatever, put the "nested" paragraph in the same quote
rebuildText = BuildWithoutBreakingLines(mdText, ParaPattern.Quote);
para = engine.CreateParagraph(new ParaProperties()
{
StyleName = DocStyles.Quote.ToDescriptionString()
});
foreach (var text in Regex.Split(rebuildText.textBlock, "\r\n|\r|\n"))
{
para.AppendChild(new Break());
FormatText(engine, para, text, new StyleProperties());
}
mdText = DeleteLines(mdText, rebuildText.counter);
continue;
case ParaPattern.AnyChar:
default:
rebuildText = BuildWithoutBreakingLines(mdText, matchedPattern.Key);
para = engine.CreateParagraph();
FormatText(engine, para, rebuildText.textBlock, new StyleProperties());
// engine.WriteText(para, text.textBlock);
mdText = DeleteLines(mdText, rebuildText.counter);
continue;
}
}
}
/// <summary>
/// Format a paragraph which could contains some text styles, Bold, Italics, Images and so on...
/// Split the markdown for each pattern found. Then append correctly that text or image into the same paragraph.
/// </summary>
/// <param name="core">The openXML object with a document, a body and a paragraph</param>
/// <param name="paragraph">The Paragraph object previously created</param>
/// <param name="markdown">The string to be processed</param>
/// <param name="fontProperties">Style properties to apply to the text</param>
internal static void FormatText(Md2MlEngine core, Paragraph paragraph, string markdown,
StyleProperties fontProperties)
{
var hasPattern = PatternMatcher.HasPatterns(markdown);
while (hasPattern)
{
var s = PatternMatcher.GetPatternsAndNonPatternText(markdown);
var newFontProperties = new StyleProperties();
switch (s.Key)
{
case StylePattern.BoldAndItalic:
newFontProperties.Bold = true;
newFontProperties.Italic = true;
FormatText(core, paragraph, s.Value[0], new StyleProperties());
FormatText(core, paragraph, s.Value[1], newFontProperties);
FormatText(core, paragraph, FramePendingString(s.Value, "***"), new StyleProperties());
break;
case StylePattern.Bold:
newFontProperties.Bold = true;
FormatText(core, paragraph, s.Value[0], new StyleProperties());
FormatText(core, paragraph, s.Value[1], newFontProperties);
FormatText(core, paragraph, FramePendingString(s.Value, "**"), new StyleProperties());
break;
case StylePattern.Italic:
newFontProperties.Italic = true;
FormatText(core, paragraph, s.Value[0], new StyleProperties());
FormatText(core, paragraph, s.Value[1], newFontProperties);
FormatText(core, paragraph, FramePendingString(s.Value, "*"), new StyleProperties());
break;
case StylePattern.MonospaceOrCode:
newFontProperties.StyleName = DocStyles.CodeReference.ToDescriptionString();
FormatText(core, paragraph, s.Value[0], new StyleProperties());
FormatText(core, paragraph, s.Value[1], newFontProperties);
FormatText(core, paragraph, FramePendingString(s.Value, "`"), new StyleProperties());
break;
case StylePattern.Strikethrough:
newFontProperties.Strikeout = true;
FormatText(core, paragraph, s.Value[0], new StyleProperties());
FormatText(core, paragraph, s.Value[1], newFontProperties);
FormatText(core, paragraph, FramePendingString(s.Value, "~~"), new StyleProperties());
break;
case StylePattern.Image:
var regex = PatternMatcher.GetStyleMatch(s.Value[1]);
FormatText(core, paragraph, s.Value[0], new StyleProperties());
core.AddImage(ConvertRelativeToAbsolutePath(regex.Value.Groups[2].Value, core.GetFileDirectory()), paragraph);
FormatText(core, paragraph, FramePendingString(s.Value, ""), new StyleProperties());
break;
case StylePattern.Underline:
newFontProperties.Underline = UnderlineValues.Single;
FormatText(core, paragraph, s.Value[0], new StyleProperties());
FormatText(core, paragraph, s.Value[1], newFontProperties);
FormatText(core, paragraph, FramePendingString(s.Value, "__"), new StyleProperties());
break;
}
return;
}
core.WriteText(paragraph, markdown, fontProperties);
}