HTML Text Filters

Plugins that deal with HTML content, can opt-in to the filtering framework. This helps to improve HTML output consistently everywhere.

Bundled filters

This package has a two standard filters included:

Filter types

A filter may be included in one of these settings:

While some filters could be used in both settings, there is a semantic difference.

Pre-filters

Any changes made by pre-filters affect the original text. These changes are visible in the WYSIWYG editor after saving. Thus, the pre-filter should be idempotent; it should be able to run multiple times over the same content. Typical use cases of a pre-filter are:

  • Validate HTML
  • Sanitize HTML (using bleach)
  • Replace "``regular quotes”`` with curly “smart” quotes.

Post-filters

The changes made by post-filters are not stored in the original text, and won’t be visible in the WYSIWYG editor. This allows a free-form manipulation of the text, for example to:

  • Add soft-hyphens in the code for better line breaking.
  • Improve typography, such as avoiding text widows, highlighting ampersands, etc.. (using django-typogrify).
  • Highlight specific words.
  • Parse “short codes” - if you really must do so. Please consider short codes a last resort. It’s recommended to create new plugins instead for complex integrations.

Since post-filters never change the original text, any filter function can be safely included as post-filter. When there is an unwanted side-effect, simply remove the post-filter and resave the text.

Creating filters

Each filter is a plain Python function that receives 2 parameters:

  • The ContentItem model, and
  • The HTML text it can update.

For example, see the smartypants filter:

from smartypants import smartypants

def smartypants_filter(contentitem, html):
    return smartypants(html)

Since the original ContentItem model is provided, a filter may read fields such as contentitem.language_code, contentitem.placeholder or contentitem.parent to have context.

The filters may also raise a ValidationError to report any errors in the text that should be corrected by the end-user.

Supporting filters in custom plugins

The bundled text plugin already uses the filters out of the box. When creating a custom plugin that includes a HTML/WYSIWYG-field, overwrite the full_clean() method in the model:

from django.db import models
from fluent_contents.extensions import PluginHtmlField
from fluent_contents.models import ContentItem
from fluent_contents.utils.filters import apply_filters


class WysiwygItem(ContentItem):
    html = PluginHtmlField("Text")
    html_final = models.TextField(editable=False, blank=True, null=True)

    def full_clean(self, *args, **kwargs):
        super(TextItem, self).full_clean(*args, **kwargs)
        self.html, self.html_final = apply_filters(self, self.html, field_name='html')

The PluginHtmlField already provides the standard cleanup and sanitation checks that the FLUENT_TEXT_CLEAN_HTML and FLUENT_TEXT_SANITIZE_HTML settings enable.