limberer

Limberer: Flexible document generation based on WeasyPrint, mustache templates, and Pandoc.

limberer is a utility for assembling markdown into documents.

Usage

$ limberer create projname
$ cd projname
$ limberer build report.toml
$ open projname.pdf

Features

• Markdown-based
• Consistent builds
• Automatic table of contents generation
• Mustache template hardening (disable __ access and lambda evaluation)
• WeasyPrint hardening (restricts file access)
• Source code snippet syntax highlighting
• 2-column layouts
• Footnote support
• Image/figure support
• Markdown within HTML tables
• Flexible

Installation

Prerequisites

$ sudo apt-get install pandoc

Note: If your distro has an older version of pandoc (e.g. 2.9.x), get it from https://github.com/jgm/pandoc/releases/.

$ wget https://github.com/jgm/pandoc/releases/download/<ver>/pandoc-<...>.deb
$ sudo dpkg -i ./pandoc-*.deb

Install

$ pip3 install --user limberer

From Source

$ git clone https://github.com/ChaosData/limberer && cd limberer
$ python3 -m pip install --user --upgrade pip setuptools
$ python3 -m pip install --user .

Packaging

$ python3 -m pip install --user wheel build
$ python3 -m build --sdist --wheel .
$ python3 -m pip install --user dist/limberer-*.whl

Cleaning

$ rm -rf ./build ./dist ./src/limberer.egg-info ./src/limberer/__pycache__

Guide

limberer is primarily about structuring documents through the use of "sections," which are ordered in a document's <project>.toml file.

These sections are based on HTML Mustache templates. For the most part, the section template will be used to write document content. For such type = "section" sections, the content will be sourced from a (currently) Markdown file based on the section name value (sections/<name>.md). By default, sections will be rendered against the template/section.html template, but the template used can be changed via an alt = "othertemplate" section list setting.

Project TOML Example

title = "Example Document"
subheading = "..."
#globaloption=value
#...

authors = [
  { name = "Example Person", email = "[e-mail]" },
  ...
]

sections = [
  { type = "cover" },
  { type = "toc" },
  { name = "example", type = "section" },
  { name = "example2", type = "section", sectionoption = "value" },
  ...
]

Additional or overriding settings or Mustache template variables can be configured by passing additional TOML file paths into the limberer build command:

$ limberer build report.toml stats.toml

Section Templates

Out of the box, limberer comes with some initial section templates:

• cover: A title page section.
• toc: A table of contents section.
• section: The underlying template for custom sections.

Additionally, limberer supports the following template-like pseudo-sections:

• appendix_start: Subsequent sections will be treated as appendices.
• appendix_end: Disables the above setting; subsequent sections are not treated as appendices. The appendix counter will not be cleared.

Custom Sections

Custom sections ("sections") are a mix of Markdown (and, in some cases, HTML), with support for Mustache expressions (using the TOML configuration), where most document content is written. A section can begin with a block of Markdown metadata:

---
toc_header_level: 2
columns: true
title: Example3
classes: foo bar baz
---

The metadata options are merged with the section entry options. Currently supported options are the following:

• toc_header_level: Header level limit to use for table of contents entry generation.
• columns: Whether or not to use the 2-column format for the section.
• title: Section title for 2-column format.
• classes: List of HTML class names to apply to the section (<article>)
• end_footnotes: Footnotes will be placed at the end of the section instead of at/near the site of placement.

Additionally, the following options may be passed in a section entry within the project TOML:

• conf = "path/to/config.toml": This specifies a path to a TOML file that is loaded for the context of the section. This is useful for loading generated data into a section containing Mustache templating, or for reusing a section that is itself a "config template" combining Mustache and Markdown.
• cont = true: This specifies that the section's Markdown should be directly concatenated instead of being handled as a full section. This is useful for combining "config templates" together, or, more simply, just managing large single sections that are not intended to be divided across multiple <article> elements within the underlying HTML.

Tables

Tables can be written using the pipe-delimited GitHub-flavored Markdown convention, or with HTML tables.

| First Header  | Second Header |
| ------------- | ------------- |
| Content Cell  | Content Cell  |
| Content Cell  | Content Cell  |

<table>
  <thead>
    <tr>
      <th style="width: 20%">a</th>
      <th style="width: 40%">b</th>
      <th>c</th>
      <th>d</th>
    </tr>
  </thead>
  <tbody>
    <tr><td>**Bold**</td><td>_italic_</td>
<td>
```
code block
```
</td>
<td>
* list
* of
* things
</td>
    </tr>
    <tr><td></td><td></td><td></td><td></td></tr>
    <tr><td></td><td></td><td></td><td></td></tr>
    <tr><td></td><td></td><td></td><td></td></tr>
  </tbody>
</table>

Images

A simple image can be embedded with custom CSS to place/style it:

![](./images/test1.jpg){style="width: 30%; margin: auto; display: block;"}

However, for the most part, figures are a better way to embed images:

![An example image](./images/test1.jpg){style="width: 30%; border: 1px solid red;"}

The figures themselves can be styled with the figstyle and figclass options:

![Another example image](./images/test1.jpg){style="width: 1.5in" figstyle="color: red; width: 35%;" figclass="aaa bbb"}

Additionally, there is support for a side-by-side image-and-text in the default layout/theme using a little HTML:

<div class="two-col-fig">
![This is on the left](./images/test1.jpg)
<div style="width: 40%">
## A Header

Lorem ipsum dolor sit amet...
</div>
</div>

The order of these can also be swapped:

<div class="two-col-fig">
<div style="width: 40%">
## A Header

Lorem ipsum dolor sit amet...
</div>
![This is on the right](./images/test1.jpg){style="width: 1.5in"}
</div>

Cross-References

Xrefs can be made through links to an element's id. By default, headings have autogenerated IDs based on the heading text (e.g. # Examplely Example will have an id of {section}-examplely-example. However, this can be overridden as follows:

## AAA{#aaah2}

Note: The {section}- prefix will be applied to all id values other than of the sections themselves.

To xref, any link to #<id> will suffice, but to style xrefs, a few options are available in the default style/layout.

• .xref: A simple unstyled segment of text.

  * [xref to Example2.AAA](#example2-aaah2){.xref}
  * <a class="xref" href="#example2-aaah2">xref to Example2.AAA</a>
  

• .xrefn: This will autopopulate the target's text content.
  Note: Be careful not to use this on elements containing large quantities of text via child nodes.

  * <a class="xrefn" href="#example2-aaah2"></a>
  * [](#example2-aaah2){.xrefn}
  

• .xrefpg: This will add a " on page XX".

  * <a class="xrefn xrefpg" href="#example2-aaah2"></a>
  * [](#example2-aaah2){.xrefn .xrefpg}
  

Code Blocks

```js { lines="true" start="99" highlight="1,5" filename="HelloWorld.js" }
let j = await fetch("https://wat.wat", {
  "headers": {
    "x-test": "foo"
  }
}).then((res)=>res.json());
```
<p class="caption">Example Snippet of Code</p>

```js
let j = await fetch("https://wat.wat", {
  "headers": {
    "x-test": "foo"
  }
}).then((res)=>res.json());
```
<figure class="caption"><figcaption>Example Snippet of Code with a figure prefix</figcaption></figure>

The following settings can be configured in the project TOML:

• highlight (defaults to "monokai")
• highlight_plaintext (defaults to "solarized-dark")
• highlight_font (defaults to "'DejaVu Sans Mono', monospace")
• highlight_style (defaults to "border-radius: 2px; overflow-x: auto;")
• highlight_padding (defaults to "padding: 0.5rem 1rem 0.5rem 1rem;")
• highlight_padding_lines (defaults to `"padding: 0.25rem 0.5rem 0.25rem 0.5rem;")

Breaks

The following can be added to force a break.

<div class="pagebreak"></div>

<div class="columnbreak"></div>

Footnotes

Footnotes should mostly work as expected, but can fit poorly and may be better shifted to another page.

Hello world.[^test]

[^test]: <https://github.com/ChaosData/limberer>

Theming/Styling

By default, limberer comes with some core styling and section templates. It is expected that users will customize their document templates beyond what is provided. As such, the limberer create command supports a -t <path> option to generate a new project from a given template project directory.

Generally speaking, the styling you want to use is up to you. However, for convenience, the CSS is organized as core, style, and custom. The intent is for the assets/core.css to cover the main layout and functioning of the document, the assets/style.css to cover group theming for consistency, and custom/custom.css to be for any per-document/project styling.

Todo List

• Support for custom Mustache-generated CSS
• Better footnotes
• Draft builds
• Partial builds of individual sections
• Support for non-Markdown sections

FAQ

Why?

For a litany of reasons, but if I had to go out on a limb and pick one, it would be that LaTeX is a great typesetter, but a terrible build system.

What!?

Greetz to asch, tanner, agrant, jblatz, and dthiel. <3