zola/docs/content/documentation/templates/overview.md
Hanno Braun 6d6df45f23
Make shortcodes work in markdown filter (#1358)
* Move `load_tera` to `templates`

I don't know if this is a good place for it, conceptually. I'm moving it
there because I need to use it from `templates`, and `templates` can't
depend on `site`, because there's already a dependency in the opposite
direction.

* Load templates in `markdown` filter

This enables the `markdown` filter to handle shortcodes, as long as
those shortcodes don't access any context variables.

Addresses #1350

* Update documentation of `markdown` filter

* Only load templates for `markdown` filter once

* Clarify `markdown` filter documentation

This is a lightly edited version of what @southerntofu suggested.
2021-02-19 20:51:08 +01:00

13 KiB

+++ title = "Overview" weight = 10 +++

Zola uses the Tera template engine, which is very similar to Jinja2, Liquid and Twig.

As this documentation will only talk about how templates work in Zola, please read the Tera template documentation if you want to learn more about it first.

All templates live in the templates directory. If you are not sure what variables are available in a template, you can place {{ __tera_context }} in the template to print the whole context.

A few variables are available on all templates except feeds and the sitemap:

  • config: the configuration without any modifications
  • current_path: the path (full URL without base_url) of the current page, always starting with a /
  • current_url: the full URL for the current page
  • lang: the language for the current page

Config variables can be accessed like config.variable, in HTML for example with {{ config.base_url }}. The 404 template does not get current_path and current_url (this information cannot be determined).

Standard templates

By default, Zola will look for three templates: index.html, which is applied to the site homepage; section.html, which is applied to all sections (any HTML page generated by creating a directory within your content directory); and page.html, which is applied to all pages (any HTML page generated by creating an .md file within your content directory).

The homepage is always a section (regardless of whether it contains other pages). Thus, the index.html and section.html templates both have access to the section variables. The page.html template has access to the page variables. The page and section variables are described in more detail in the next section.

Built-in templates

Zola comes with four built-in templates: atom.xml and rss.xml (described in Feeds), sitemap.xml (described in Sitemap), and robots.txt (described in Robots.txt). Additionally, themes can add their own templates, which will be applied if not overridden. You can override built-in or theme templates by creating a template with the same name in the correct path. For example, you can override the Atom template by creating a templates/atom.xml file.

Custom templates

In addition to the standard index.html, section.html and page.html templates, you may also create custom templates by creating an .html file in the templates directory. These custom templates will not be used by default. Instead, a custom template will only be used if you apply it by setting the template front-matter variable to the path for that template (or if you include it in another template that is applied). For example, if you created a custom template for your site's About page called about.html, you could apply it to your about.md page by including the following front matter in your about.md page:

+++
title = "About Us"
template = "about.html"
+++

Custom templates are not required to live at the root of your templates directory. For example, product_pages/with_pictures.html is a valid template.

Built-in filters

Zola adds a few filters in addition to those already present in Tera.

markdown

Converts the given variable to HTML using Markdown. Please note that shortcodes evaluated by this filter cannot access the current rendering context. config will be available, but accessing section or page (among others) from a shortcode called within the markdown filter will prevent your site from building. See this discussion.

By default, the filter will wrap all text in a paragraph. To disable this behaviour, you can pass true to the inline argument:

{{ some_text | markdown(inline=true) }}

You do not need to use this filter with page.content or section.content, the content is already rendered.

base64_encode

Encode the variable to base64.

base64_decode

Decode the variable from base64.

Built-in global functions

Zola adds a few global functions to those in Tera to make it easier to develop complex sites.

get_page

Takes a path to an .md file and returns the associated page.

{% set page = get_page(path="blog/page2.md") %}

get_section

Takes a path to an _index.md file and returns the associated section.

{% set section = get_section(path="blog/_index.md") %}

If you only need the metadata of the section, you can pass metadata_only=true to the function:

{% set section = get_section(path="blog/_index.md", metadata_only=true) %}

get_url

Gets the permalink for the given path. If the path starts with @/, it will be treated as an internal link like the ones used in Markdown, starting from the root content directory.

{% set url = get_url(path="@/blog/_index.md") %}

It accepts an optional parameter lang in order to compute a language-aware URL in multilingual websites. Assuming config.base_url is "http://example.com", the following snippet will:

  • return "http://example.com/blog/" if config.default_language is "en"
  • return "http://example.com/en/blog/" if config.default_language is not "en" and "en" appears in config.languages
  • fail otherwise, with the error message "'en' is not an authorized language (check config.languages)."
{% set url = get_url(path="@/blog/_index.md", lang="en") %}

This can also be used to get the permalinks for static assets, for example if we want to link to the file that is located at static/css/app.css:

{{/* get_url(path="css/app.css") */}}

By default, assets will not have a trailing slash. You can force one by passing trailing_slash=true to the get_url function. An example is:

{{/* get_url(path="css/app.css", trailing_slash=true) */}}

In the case of non-internal links, you can also add a cachebust of the format ?h=<sha256> at the end of a URL by passing cachebust=true to the get_url function.

get_file_hash

Gets the hash digest for a static file. Supported hashes are SHA-256, SHA-384 (default) and SHA-512. Requires path. The sha_type key is optional and must be one of 256, 384 or 512.

{{/* get_file_hash(path="js/app.js", sha_type=256) */}}

This can be used to implement subresource integrity. Do note that subresource integrity is typically used when using external scripts, which get_file_hash does not support.

<script src="{{/* get_url(path="js/app.js") */}}"
        integrity="sha384-{{/* get_file_hash(path="js/app.js", sha_type=384) */}}"></script>

Whenever hashing files, whether using get_file_hash or get_url(..., cachebust=true), the file is searched for in three places: static/, content/ and the output path (so e.g. compiled SASS can be hashed, too.)

get_image_metadata

Gets metadata for an image. This supports common formats like JPEG, PNG, as well as SVG. Currently, the only supported keys are width and height.

  {% set meta = get_image_metadata(path="...") %}
  Our image is {{ meta.width }}x{{ meta.height }}

get_taxonomy_url

Gets the permalink for the taxonomy item found.

{% set url = get_taxonomy_url(kind="categories", name=page.taxonomies.category, lang=page.lang) %}

name will almost always come from a variable but in case you want to do it manually, the value should be the same as the one in the front matter, not the slugified version.

lang (optional) default to config.default_language in config.toml

get_taxonomy

Gets the whole taxonomy of a specific kind.

{% set categories = get_taxonomy(kind="categories") %}

The type of the output is:

kind: TaxonomyConfig;
items: Array<TaxonomyTerm>;

See the Taxonomies documentation for a full documentation of those types.

load_data

Loads data from a file or URL. Supported file types include toml, json, csv and bibtex. Any other file type will be loaded as plain text.

The path argument specifies the path to the data file relative to your base directory, where your config.toml is. As a security precaution, if this file is outside the main site directory, your site will fail to build.

{% set data = load_data(path="content/blog/story/data.toml") %}

The optional required boolean argument can be set to false so that missing data (HTTP error or local file not found) does not produce an error, but returns a null value instead. However, permission issues with a local file and invalid data that could not be parsed to the requested data format will still produce an error even with required=false.

The snippet below outputs the HTML from a Wikipedia page, or "No data found" if the page was not reachable, or did not return a successful HTTP code:

{% set data = load_data(path="https://en.wikipedia.org/wiki/Commune_of_Paris", required=false) %}
{% if data %}{{ data | safe }}{% else %}No data found{% endif %}

The optional format argument allows you to specify and override which data type is contained within the file specified in the path argument. Valid entries are toml, json, csv, bibtex or plain. If the format argument isn't specified, then the path extension is used.

{% set data = load_data(path="content/blog/story/data.txt", format="json") %}

Use the plain format for when your file has a toml/json/csv extension but you want to load it as plain text.

For toml and json, the data is loaded into a structure matching the original data file; however, for csv there is no native notion of such a structure. Instead, the data is separated into a data structure containing headers and records. See the example below to see how this works.

In the template:

{% set data = load_data(path="content/blog/story/data.csv") %}

In the content/blog/story/data.csv file:

Number, Title
1,Gutenberg
2,Printing

The equivalent json value of the parsed data would be stored in the data variable in the template:

{
    "headers": ["Number", "Title"],
    "records": [
        ["1", "Gutenberg"],
        ["2", "Printing"]
    ],
}

The bibtex format loads data into a structure matching the format used by the nom-bibtex crate. The following is an example of data in bibtex format:

@preamble{"A bibtex preamble" # " this is."}

@Comment{
    Here is a comment.
}

Another comment!

@string(name = "Vincent Prouillet")
@string(github = "https://github.com/getzola/zola")

@misc {my_citation_key,
    author= name,
    title = "Zola",
    note = "github: " # github
}                                                    }

The following is the json-equivalent format of the produced bibtex data structure:

{
    "preambles": ["A bibtex preamble this is."],
    "comments": ["Here is a comment.", "Another comment!"],
    "variables": {
        "name": "Vincent Prouillet",
        "github": "https://github.com/getzola/zola"
    },
    "bibliographies": [
        {
            "entry_type": "misc",
            "citation_key": "my_citation_key",
            "tags": {
                "author": "Vincent Prouillet",
                "title": "Zola",
                "note": "github: https://github.com/getzola/zola"
            }
        }
    ]
}

Finally, the bibtex data can be accessed from the template as follows:

{% set tags = data.bibliographies[0].tags %}
This was generated using {{ tags.title }}, authored by {{ tags.author }}.

Remote content

Instead of using a file, you can load data from a remote URL. This can be done by specifying a url parameter to load_data rather than path.

{% set response = load_data(url="https://api.github.com/repos/getzola/zola") %}
{{ response }}

By default, the response body will be returned with no parsing. This can be changed by using the format argument as below.

{% set response = load_data(url="https://api.github.com/repos/getzola/zola", format="json") %}
{{ response }}

Data caching

Data file loading and remote requests are cached in memory during the build, so multiple requests aren't made to the same endpoint. URLs are cached based on the URL, and data files are cached based on the file modified time. The format is also taken into account when caching, so a request will be sent twice if it's loaded with two different formats.

trans

Gets the translation of the given key, for the default_language or the language given

{{/* trans(key="title") */}}
{{/* trans(key="title", lang="fr") */}}

resize_image

Resizes an image file. Please refer to Content / Image Processing for complete documentation.