Contentful provides a content infrastructure for digital teams to power content in websites, apps, and devices. Unlike a CMS, Contentful was built to integrate with the modern software stack. It offers a central hub for structured content, powerful management and delivery APIs, and a customizable web app that enable developers and content creators to ship digital products faster.
Contentful Middleman is a Middleman extension to use the Middleman static site generator together with Contentful. It is powered by the Contentful Ruby Gem.
Experience the power of Middleman while staying sane as a developer by letting end-users edit content in a web-based interface.
The main release works for Middleman v4 starting on version
4.0.0
.If you are using Middleman v3, use older releases. Latest stable release is
3.0.0
Add the following line to the Gemfile of your Middleman project:
gem "contentful_middleman"
Then as usual, run:
bundle install
Run middleman contentful
in your terminal. This will fetch entries for the configured
spaces and content types and put the resulting data in the
local data folder as yaml files.
The contentful
command has a --rebuild
option which will trigger a rebuild of your site only if there were changes between the last
and the current import.
To configure the extension, add the following configuration block to Middleman's config.rb:
activate :contentful do |f|
f.space = SPACE
f.access_token = ACCESS_TOKEN
f.cda_query = QUERY
f.content_types = CONTENT_TYPES_MAPPINGS
end
Parameter | Description |
---|---|
space | Hash with an user choosen name for the space as key and the space id as value. |
access_token | Contentful Delivery API access token. |
cda_query | Hash describing query configuration. See contentful.rb for more info (look for filter options there). Note that by default only 100 entries will be fetched, this can be configured to up to 1000 entries using the limit option. Example: f.cda_query = { limit: 1000 } . |
client_options | Hash describing client configuration. See contentful.rb for more info. This option should commonly be used to change Rate Limit Management, Include Resolution, Logging and Proxies. |
content_types | Hash describing the mapping applied to entries of the imported content types. |
default_locale | String with the value for the default locale for your space. Defaults to 'en-US' . |
use_preview_api | Boolean to toggle the used API. Set it to false to use cdn.contentful.com (default value). Set it to true to use preview.contentful.com . More info in the documentation |
all_entries | Boolean to toggle multiple requests to the API for getting over 1000 entries. This uses a naive approach and can get rate limited. When using this, have in mind adding an order in your :cda_query . Default order is order: 'sys.createdAt' . |
all_entries_page_size | Integer amount of items per page for :all_entries requests, allowing for smaller page sizes on content heavy requests. |
rebuild_on_webhook | Boolean to toggle Webhook server. Server will run in port 5678, and will be expecting to receive Contentful Webhook calls on /receive . |
webhook_timeout | Integer (in seconds) for wait time after Webhook received for rebuilding. Only used if :rebuild_on_webhook is true. Defaults to 300 seconds. |
webhook_controller | Class for handling Webhook response, defaults to ::ContentfulMiddleman::WebhookHandler . |
rich_text_mappings | Hash with 'nodeType' => RendererClass pairs determining overrides for the RichTextRenderer library configuration. |
base_path | String with path to your Middleman Application, defaults to current directory. Path is relative to your current location. |
destination | String with path within your base path under which to store the output yaml files. Defaults to data . |
use_sync | If true, the gem will try to use the contentful sync feature for all given content_types . |
You can activate the extension multiple times to import entries from different spaces.
The extension will transform every fetched entry before storing it as a yaml file in the local data folder. If a custom mapper is not specified a default one will be used.
The default mapper will map fields, assets and linked entries.
You can create your own mappers if you need so. The only requirements for a class to behave as a
mapper are an initializer and a map(context, entry)
instance method.
The initializer takes two parameters:
- A
Contentful::Array
of all entries for the current content model - A
Middleman::Configuration::ConfigurationManager
object containing the Contentful configuration options set inconfig.rb
See BackrefMapper for an example use of entries.
See the Base mapper for an example use of options.
The map
method takes two parameters:
- A context object. All properties set on this object will be written to the yaml file
- An entry
Following is an example of such custom mapper:
class MyAwesomeMapper
def initialize(entries, options)
@entries = entries
@options = options
end
def map(context, entry)
context.slug = entry.title.parameterize
#... more transformations
end
end
If you don't want to map all the fields by hand inherit from the Base mappper:
class MyAwesomeMapper < ContentfulMiddleman::Mapper::Base
def map(context, entry)
super
# After calling super the context object
# will have a property for every field in the
# entry
end
end
There's also an example back-reference mapper in the examples directory for adding back-references onto entries that are linked to by other entries.
If you want to process a Content Type with multiple mappers, you can use the Composite Design Pattern. The Mapper code should look something similar to the following.
Then you can attach as many Custom Mappers as you want to that one.
class CompositeMapper < ContentfulMiddleman::Mapper::Base
@@mappers = []
def self.mappers
@@mappers
end
def map(context, entry)
super
mappers.each do |m|
m.map(context, entry)
end
end
end
Then in your config.rb file:
CompositeMapper.mappers << YourMapper.new
CompositeMapper.mappers << OtherMapper.new
activate :contentful do |f|
'... your config here ...'
f.content_types = {content_type_name_you_want_to_map: {mapper: CompositeMapper, id: 'content_type_id'}}
end
NOTE: This kind of Composite Mapper is static, therefore if you want to have multiple combinations of mappers for multiple entries, you'd need to write code a bit differently.
To render rich text in your views, you can use the rich_text
view helper.
An example using erb
:
<% data.my_space.my_type.each do |_, entry| %>
<%= rich_text(entry.rich_field) %>
<% end %>
This will output the generated HTML generated by the RichTextRenderer
library.
When using rich text, if you're planning to embed entries, then you need to create your custom renderer for them. You can read how create your own renderer classes here.
To configure the mappings, you need to add them in your activate
block like follows:
activate :contentful do |f|
# ... all the regular config ...
f.rich_text_mappings = { 'embedded-entry-block' => MyCustomRenderer }
end
You can also add renderers for all other types of nodes if you want to have more granular control over the rendering.
In case you have multiple activated extensions, and have different mapping configurations for them. You can specify which extension instance you want to pull the configuration from when using the helper.
The helper receives an additional optional parameter for the extension instance. By default it is 0
, indicating the first activated extension.
The instances are sequentially numbered in order of activation, starting from 0.
So, if for example you have 2 active instances with different configuration, to use the second instance configuration, you should call the helper as: rich_text(entry.rich_field, 1)
.
activate :contentful do |f|
f.space = {partners: 'space-id'}
f.access_token = 'some_access_token'
f.cda_query = { content_type: 'content-type-id', include: 1 }
f.content_types = { partner: 'content-type-id'}
end
The above configuration does the following:
- Sets the alias
partners
to the space with id some-id - Sets the alias
partner
to the content type with id content-type-id - Uses the default mapper to transform
partner
entries into yaml files (no mapper specified for thepartner
content type)
Entries fetched using this configuration will be stored as yaml files in data/partners/partner/ENTRY_ID.yaml
.
class Mapper
def map(context, entry)
context.title = "#{entry.title}-title"
#...
end
end
activate :contentful do |f|
f.space = {partners: 'space-id'}
f.access_token = 'some_access_token'
f.cda_query = { content_type: '1EVL9Bl48Euu28QEOa44ai', include: 1 }
f.content_types = { partner: {mapper: Mapper, id: 'content-type-id'}}
end
The above configuration is the same as the previous one only that this time we are setting a custom mapper
for the entries belonging to the partner
content type.
activate :contentful do |f|
f.space = {parters: 'space-id'}
f.access_token = 'some_access_token'
f.use_sync = true
# this is necessary for sync, otherwise it would store entries in a fashion
# that each entry has their first 10 child nesting stored with them
#f.client_options = {max_include_resolution_depth: 0, default_locale: 'en'}
f.content_types = {
local_content_type_name: 'cfContentTypeName',
another_content_type: 'anotherContentfulType',
# ...
}
end
Using the sync feature changes the folder structure in the data/
directory
so that within each folder for the content types, all entries matching this on
the server side get stored, one entry per file.
It would look roughly like this:
data/
partners/
local_content_type_name/
some_random_id.yaml
another_id.yaml
- ...
another_content_type/
related_entry_id.yaml
- ...
In order to also have a fast lookup, the file partners-repository.yaml
is
also saved in the data/
directory, keeping track of all cached entries, their
respective content types, and meta information regarding the sync.
For a fresh start, simply delete the repository file and the folder named like
the space, in this example case partners/
.
Sometimes, it is also necessary to look up single entries by their entry ID.
This is also supported in templates, config and in the entire application scope
through the repository
keyword.
Because middleman_contenful supports using multiple spaces at once, each space
can be accessed through dot-notation, resulting in repository.partners
.
From there on out, one can do the following:
# accessing all entries for a content type:
repository.partners.local_content_type_name.each do |entry|
puts entry.meta.id
puts entry.meta.content_type
puts entry.some_field_name
end
# querying a single entry with its id:
repository.partners.entry('some_random_id') # => yields the local_content_type_name.some_random_id item
If an entry has sub-entries in this case, these are populated on access, such that one can traverse the entry hierarchy just like usual. Contentful habitually also delivers links to entries that are not published yet, even on the production api endpoint. This leads to interesting behavior when accessing an entry that has not-published children, since the production api will not make those drafted articles accessible on sync. As a result, while building the static site, this can result in messages indicating that a certain entry was not found. This is expected behavior and a reliable indicator that these entries are not published yet.
As with the regular fetching mechanics, calling middleman contentful
will
either start a new sync if a) the stored token for the next sync is invalid
or b) no repository with the name of the space is found. It will then perform an
initial sync, which can take some time if a lot of entries are in the namespace.
Any further invocation will then only perform updates yielded by the contentful
server's diffing mechanism.
To not bloat the update logic unnecessarily, updates to new entries will simply
override the whole file of that entry.
Also note that assets currently are not supported to be synced to the local environment, since it is expected to have them referenced to the external contentful content delivery network.
Middleman will load all the yaml files stored in the local data folder. This lets you use all the imported data into your templates.
Consider that we have data stored under data/partners/partner
. Then in our templates we could use that data like
this:
<h1>Partners</h1>
<ol>
<% data.partners.partner.each do |id, partner| %>
<li><%= partner["name"] %></li>
<% end %>
</ol>
If you want to use markdown in your content types you manually have to render this to markdown. Depending on the markdown library you need to transform the data. For Kramdown this would be:
<%= Kramdown::Document.new(data).to_html %>
If you have localized entries, and want to display content for multiple locales.
You can now include locale: '*'
in your CDA query.
Then you have the following methods of accessing locales:
- Manual access
You can access your localized fields by fetching the locale directly from the data
<h1>Partners</h1>
<ol>
<% data.partners.partner.each do |id, partner| %>
<li><%= partner["name"]['en-US'] %></li>
<% end %>
</ol>
- Entry Helper
You can also map an specific locale for all entry fields using localize_entry
<h1>Partners</h1>
<ol>
<% data.partners.partner.each do |id, partner| %>
<% localized_partner = localize_entry(partner, 'es') %>
<li><%= localized_partner["name"] %></li>
<% end %>
</ol>
- Generic Field Helper
The localize
helper will map an specific locale to a field of your entry
<h1>Partners</h1>
<ol>
<% data.partners.partner.each do |id, partner| %>
<li>Value Field: <%= localize(partner, 'name', 'en-US') %></li>
<li>Array Field: <%= localize(partner, 'phones', 'es') %></li>
<% end %>
</ol>
- Specific Field Type Helper
Or, you can use localize_value
or localize_array
if you want more granularity.
This method is discouraged, as
localize
achieves the same goal and is a field-type agnostic wrapper of these methods.
<h1>Partners</h1>
<ol>
<% data.partners.partner.each do |id, partner| %>
<li>Value Field: <%= localize_value(partner['name'], 'en-US') %></li>
<li>Array Field: <%= localize_array(partner['phones'], 'es') %></li>
<% end %>
</ol>
If your fields are not localized, the value of the field will be returned.
In case of the field being localized but no value being set for a given entry, it will use
a fallback locale, by default is en-US
but can be specified as an additional
parameter in all the mentioned calls.
You can use the #with_preview
helper to try your Preview API content without having to
generate the entire data
structures.
This generates a new Preview Contentful Client and has a cache that will store your objects in memory until they are considered to need refresh.
It can be used like a Contentful Client:
<% with_preview(space: 'cfexampleapi', access_token: 'b4c0n73n7fu1') do |preview| %>
<% entry = preview.entry('nyancat') %>
<p>Name: <%= entry.name %></p>
<% end %>
If you want to clear the cache to force a refresh:
<% with_preview(space: 'cfexampleapi', access_token: 'b4c0n73n7fu1') do |preview| %>
<% preview.clear_cache %>
<% end %>
- Every preview client will be cached by Space/Access Token combination
- Only
entry
,entries
,asset
andassets
will be cached - Every call will be cached by it's query parameters and ID (if ID is applicable)
- Each call will be considered, by default, stale after 3 tries or 2 hours
- Cache can be cleared by calling
#clear_cache
, this applies per preview client
You can configure :tries
and :expires_in
in the #with_preview
call like this:
<% with_preview(
space: 'cfexampleapi',
access_token: 'b4c0n73n7fu1',
tries: 20, # Set Tries to 20 before stale
expires_in: ContentfulMiddleman::Tools::PreviewProxy.minutes(5) # Set Expiration to 5 minutes
) do |preview| %>
<!-- do your stuff -->
<% end %>
For platform specific issues, please look into the DEPLOYING document. This document is expected to grow with user contributions. Feel free to add your own discoveries to that file by issuing a Pull Request.