I'm using Jekyll to create a new blog. It uses Liquid underneath.
Jekyll defines certain "variables": site, content, page, post and paginator. These "variables" have several "members". For instance, post.date will return the date of a post, while post.url will return its url.
My question is: can I access a variable's member using another variable as the member name?
See the following example:
{% if my_condition %}
{% assign name = 'date' %}
{% else %}
{% assign name = 'url' %}
{% endif %}
I have a variable called name which is either 'date' or 'url'.
How can I make the liquid equivalent of post[name] in ruby?
The only way I've found is using a for loop to iterate over all the pairs (key-value) of post. Beware! It is quite horrible:
{% for property in post %}
{% if property[0] == name %}
{{ property[1] }}
{% endif %}
{% endfor %}
Argh! I hope there is a better way.
Thanks.
I don't know what I was thinking.
post[name] is a perfectly valid liquid construction. So the for-if code above can be replaced by this:
{{ post[name] }}
I thought that I tried this, but apparently I didn't. D'oh!
Liquid admits even fancier constructs; the following one is syntactically correct, and will return the expected value if post, element, categories, etc are correctly defined:
{{ post[element.id].categories[1].name }}
I am greatly surprised with Liquid. Will definitively continue investigating.
Actually, this did not work for me. I tried a bunch of different combinations and what finally worked was
<!-- object myObj.test has the string value "this is a test" -->
{% assign x = 'test' %}
{{ myObj.[x] }}
Output:
this is a test
Related
This works perfectly fine:
{% capture foo %}{% include test.html %}{% endcapture %}
I want to do this:
frontmatter.md:
---
baaz: test.html
layout: layout.html
---
layout.html:
{% capture foo %}{% include {{ page.baaz }} %}{% endcapture %}
But when I do I'm given this error:
"Liquid Exception: Invalid syntax for include tag. File contains invalid characters or sequences: Valid syntax: {% include file.ext param='value' param2='value' %}"
I've seen this addressed in several other questions, with the most recent explanation I've found being this:
"...dynamic filename paths can't be added due to the fact that the included files are calculated and added at the compilation phase and not at run time phase. And compilation phase means dynamic paths aren't yet recognized."
But that source is nearly two years old. Does anyone have a solution to this yet? Or a workaround that would allow me to include a file defined as a variable in frontmatter?
You can try {% include page.baaz %}
Edit : after some investigations, it appears that your syntax is correct, and that the error fires only when page.baaz is not present.
This ends up in an include tag which looks like this for liquid :
{% include %}
In order to avoid this error on certain pages/post with no baaz set, you can use a condition.
{% if page.baaz %}
{% capture foo %}{% include {{ page.baaz }} %}{% endcapture %}
{% endif %}
I just came to this case recently. I assume the syntax works as expected. See sample and result.
{% include {{ page.baaz }} %}
However in your case it might be the page name could not be put in a variable as the error stated:
Error: Invalid syntax for include tag:
File contains invalid characters or sequences
Valid syntax:
***% include file.ext param='value' param2='value' %***
So to come out from the problem I would suggest you to inventory all file names and choose it:
{% case page.baaz %}
{% when 'test.html' %}
{% capture foo %}{% include test.html %}{% endcapture %}
{% when 'othertest.html' %}
{% capture foo %}{% include othertest.html %}{% endcapture %}
{% else %}
This is not a test
{% endcase %}
I had a similar issue... I have found a very usable work-around. Allow me to share my experience and solution. I hope it helps you to find a suitable solution for your problem.
What I wanted to build
I wanted to make a page with multiple sections. The sections should be reusable, be able to contain includes and they should be easy to manage in the CloudCannon CMS.
What I came up with
I ended up using the following front matter:
---
title: Lorem ipsum
description: Lorem ipsum
image: /img/default.jpg
section_blocks:
- section: sectionwithinclude
- section: anothersection
- section: andyetanothersection
---
... and the following tempate:
{% for item in page.section_blocks %}
{% for section in site.sections %}
{% if item.section == section.slug %}
<div class="section {{ item.section }}">
{{ section.content }}
</div>
{% endif %}
{% endfor %}
{% endfor %}
Within the _sections folder/collection I have a file called sectionwithinclude.md that looks like this:
---
---
{% include mycustominclude.html %}
Why this is great
When you edit your page, CloudCannon will show the section_blocks as an array with reorder buttons. Additionally, CloudCannon will automagically recognize section as a collection and show the options in a dropdown. Therefore adding a section is a matter of adding an empty item to the array, selecting a section from the dropdown and reordering it with the array buttons. On the same time, the inline editing option of CloudCannon still works. So management of text can be WYSIWYG, while block management can be done in the front matter array.
Super easy and powerful for (you and) your editors.
PS. You might find out that you will have some 'scope' issues, because page no longer relates to the actual page, but to the section. To solve this you can/should alter the loop in the template. You can let the loop manage the include instead of the section.
It is possible to expand a variable into a post Front-Matter?
I use a series of items for links into my template, like:
related_links:
- text: foo
link: bar
But sometimes I need to refer to other posts into my site. Normally I would use just /bar into link, but this is also used as shownotes into a podcast, I want to expand the link for http://example.com/bar. But using {% post_url YYYY-MM-DD-bar %} results into:
Error: could not read file [REDACTED]: (<unknown>): found character that cannot start any token while scanning for the next token at line 33 column 12
Any tips?
If you use
related_links:
- text: foo
link: {% post_url YYYY-MM-DD-bar %}
you will get an error because the { will start a flow style mapping and % cannot start a token in YAML. You have to put the whole scalar in (double) quotes:
related_links:
- text: foo
link: "{% post_url YYYY-MM-DD-bar %}"
What you want to do does not work because of Jekyll's pipeline:
parse YAML front matter
process rest of document with Liquid
parse result of previous step with Markdown
{% post_url YYYY-MM-DD-bar %} is a Liquid command. As you see, Liquid processes just the part of your file below the YAML front matter. Therefore, no Liquid replacement takes place there.
You can instead write something like this below the front matter:
{% assign link = post_url YYYY-MM-DD-bar %}
And then use {{link}} elsewhere. If you have multiple links, things get hacky. Something like this might work, but I am not enough of a Liquid user to know for sure:
{% capture nl %}
{% endcapture %}
{% capture rawlinks %}
{% post_url YYYY-MM-DD-bar %}
{% post_url YYYY-MM-DD-bar %}
{% endcapture %}
{% assign links = rawlinks | split nl %}
You can then specify indexes in your YAML front matter:
related_links:
- text: foo
linkindex: 0
And finally, somewhere in your document:
{{ links[related_link.linkindex] }}
YMMV if this level of uglyness is justified for your use-case.
I found a more elegant at my own:
{% assign real_link=link.link %}
{% assign link_start = real_link | slice: 0 %}
{% if link_start == "/" %}{% assign real_link = real_link | prepend: site.url %}{% endif %}
As I start al my locak links with a / to warrant that the link will be related with the root in the site, that was a better way for me.
I have created an empty array in Jekyll but I am not able to push data.
Here it is described how to do but it does not seem to work on my machine. I use Jekyll 2.5.3
{% assign selected = site.array %}
{% for success in site.data.success %}
{% assign selected = selected | push: success.id %}
{% endfor %}
site.array is defined into my _config.yml as
site.array: []
I am sure that success.id exists
UPDATE 1
I changed in _config.yml
array: []
Still no luck.
Why define a site.something in _config.yml ?
Any variable defined here will be accessed with the site prefix, but it's not up to you to add this prefix, Jekyll takes care of it. eg: {{ site.title }} refers to site: My awesome title. It's automatic.
Just define : emptyArray: [], and access it with site.emptyArray.
With site.array: [], you're defining a configuration variable that will never be site.site.array because a yaml key is supposed to be a string. Here, the only way to access your variable, is to write site["site.array"].
Note: If you can, upgrade to a newer version of Jekyll if you want to be future-proof.
I'm not sure if Jekyll accepts an empty array, "out of the blue".
So, try to use an if statement before the for loop:
{% if site.array %}
{% assign selected = site.array %}
{% for success in site.data.success %} {% assign selected = selected | push: success.id %}
{% endfor %}
{% endif %}
And keep just array: in your _config.yml. Any variable defined in this file will be called by site.variable.
I suggest you to try with some values in the array before leaving it empty, to make sure the code works.
Hope to have helped!
I would like to display all session variables with twig but I dont know exactly how:
{% for variable in app.session %}
session[ {{variable clef}}] = {{variable value}}
{% endfor %}
thank you for your help
You can use this, it will work for other associative arrays as well:
{% for key, value in app.session %}
session['{{ key }}'] = {{ value }}
{% endfor %}
It's also explained in the documentation: Iterating over Keys and Values.
You could also use twig's debugging function, dump. This will work even if the value of a key is an object or array.
{{ dump(app.session.all) }}
Any .html template for django-registration module works fine with {% blocktrans %} and {% trans %} template blocks. With {% load i18n %} in place, of course.
But I cannot make use of i18n tags in activation_email.txt and activation_email_subject.txt templates. Strings marked for translation just don't appear in .po file after makemessages.
Also, when wrapping a text with {% blocktrans %}{% endblocktrans %}, all variables such as {{ site.domain }} and {{ site.name }} are not processed.
Can you suggest what I am doing wrong?
That was my bad, I just improperly used makemessages. By default it processes only .html files.
In my case
django-admin.py makemessages -a -e html,txt
does all the work.
As for variables, {% blocktrans %}{% endblocktrans %} cannot process variables inside object, so we have to retrieve them before translation:
{% blocktrans with site.name as site_name and site.domain as site_domain %}
Good examples of templates for django-registration are given here.