Publishing Failed with Markdown Docs where Formula is in use

[dcwise]

Hi there, We run into a problem when publishing Markdown documents using the Oxygen-enabled DITA OT. Specifically, in case there are formula in use as shown below,publishing always fails. Use Case 1

功能

$$ y = \frac{e^{x} - e^{-x}}{e^{x} + e^{-x}} $$

Use Case 2

功能

$$ Y[n_i][c_j][yh_m][ywn] = \max^{kh-1}{h=0} \max^{kw-1}_{w=0} X[n_i][c_j][sh yh_m + h][sw yw_n + w] $$

二维最大池化正向传播过程。根据输入、池化核、步长（stride）等参数计算输出特征图的大小，NCHW数据格式可用以上公式表示。

Please can you advise how to make this embeded Latex work from within Markdown documents so that the DITA OT enabled publishing will work without a problem?

[dcwise]

Further to my last message, It was this line thrown by the DITA OT process as follows: "no render configured for vladsch.flexmark.superscript.Superscript" that led to the publishing failure.

[jelovirt]

LwDITA Markdown doesn't support LaTeX extension for input or output. That LaTeX uses a ^ character that is interpreted as a superscript delimiter. LwDITA Markdown support superscript, but maybe file is parsed as MDITA that doesn't.

[jelovirt]

It's very unlikely that support for LaTeX will be added, because the underlying Markdown parser doesn't have an extension for it, and we would need to convert LaTeX into MathML.

[raducoravu]

@jelovirt maybe the processing could attempt to just extract the Latex code and generate a DITA "foreign" element containing it with a certain outputclass attribute, in this way a simple publishing plugin like this one could process it further in the publishing stages: https://blog.oxygenxml.com/topics/adding-latex-equation-support-dita.html Looking at the wiki for flexmark: https://github.com/vsch/flexmark-java/wiki/Extensions they seem to have some support for this in GitHub flavored markdown:

Math, inline via $``$ or as fenced code with math info string requiring inclusion of Katex in the rendered HTML page.

[raducoravu]

@jelovirt any feedback for my previous remark? How about if we would use the "GitLabExtension" when parsing with flexmark?

[dcwise]

GitLab actually does a better job when parsing a Latex enabled expression.

On 7/3/23, Radu Coravu @.***> wrote:

@jelovirt any feedback for my previous remark? How about if we would use the "GitLabExtension" when parsing with flexmark?

-- Reply to this email directly or view it on GitHub: https://github.com/jelovirt/org.lwdita/issues/181#issuecomment-1617451832 You are receiving this because you authored the thread.

Message ID: @.***>

-- Keep an Exacting Eye for Detail

[raducoravu]

GitHub also has support for embedding Latex equations in markdown: https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/writing-mathematical-expressions