Collection configuration

1. Explicitly configure a path for every collection#

CloudCannon no longer detects collections during your build step.

Previously, CloudCannon could infer the path for each collection during the Site build step (e.g., Jekyll Sites automatically assume your posts collection is in the _posts folder). Additionally, some build integrations could discover collections at build time (e.g., posts, page, and drafts on Jekyll Sites). CloudCannon no longer uses this behavior.

Please ensure that each collection in collections_config has the path key defined. The below code shows how to configure a path explicitly:

collections_config:
  blog:
    path: content/blog

{
  "collections_config": {
    "blog": {
      "path": "content/blog"
    }
  }
}

2. Replace parse_branch_index with a glob#

CloudCannon no longer uses parse_branch_index: true to add index files from one collection to a pages collection. Instead, you can define a file glob to exclude your index files from their initial collection.

Most Hugo sites will have a pages collection defined with a path of content. Pages excluded from any collections that are subdirectories will be assigned to this pages collection.

By default, all content files within a specific path are included in the associated collection, unless they also fall under a more specific path. CloudCannon now uses files globs to customize which files belong to a collection and are visible in the collection browser. You can add a negative glob to the collections_config.*.glob key by prefixing your file name with the ! character.

The below code shows how to configure the glob key to exclude the _index.md file for your collection:

collections_config:
  blog:
    path: content/blog
    glob:
      - '!_index.md'

{
  "collections_config": {
    "blog": {
      "path": "content/blog",
      "glob": [
        "!_index.md"
      ]
    }
  }
}

The above example ensures that all files within the blog folder are included in the blog collection, with the exception of _index.md.

3. Replace collection filters with globs#

CloudCannon no longer uses the _unlisted or collections_config.*.filter keys to filter which files to show in a collection browser. All collections now show only collection-like files that explicitly belong to a collection, as per the old filter.base: strict configuration (i.e., will not show collection-like files from a different collection that are stored at the same collection path).

CloudCannon now uses file globs to determine which files belong to a collection and are visible in the collection browser. A file glob is a pattern used to identify matching files. The collections_config.*.glob key is an Array. Each entry in the array is a pattern that controls which files are visible in CloudCannon's collection browser. The glob pattern is relative to your collection path and can include negative values by adding the ! character as a prefix.

The below code shows how to configure the glob key for your collection:

collections_config:
  blog:
    path: content/blog
    glob:
      - '!**/*.yml'

{
  "collections_config": {
    "blog": {
      "path": "content/blog",
      "glob": [
        "!**/*.yml"
      ]
    }
  }
}

The above example ensures that all YAML files within the blog folder and all subfolders are not included in the blog collection.

4. Explicitly configure your data collection#

CloudCannon no longer uses build integrations to set the following keys in collections_config.data:

1. disable_url
2. disable_file_actions
3. disable_add_folder
4. disable_add

You must explicitly set these keys.

5. Configure collection groups#

CloudCannon no longer groups collections in your Site Navigation by default. If you want to retain your original grouping, add the collection_groups key to your CloudCannon configuration file.

The below code shows how to configure the collection_groups key:

collection_groups:
  - heading: Pages
    collections:
      - pages
  - heading: Blogging
    collections:
      - posts
      - drafts
  - heading: Data
    collections:
      - data

{
  "collection_groups": [
    {
      "heading": "Pages",
      "collections": [
        "pages"
      ]
    },
    {
      "heading": "Blogging",
      "collections": [
        "posts",
        "drafts"
      ]
    },
    {
      "heading": "Data",
      "collections": [
        "data"
      ]
    }
  ]
}

6. Filter _defaults.md with a glob or schema#

If you still use the deprecated _defaults.md file, CloudCannon will no longer filter this file from your collection by default. Please replace these files with the Schema feature, or use a collection glob (i.e., collections_config.*.glob) to filter the _defaults.md file from your collection browser.

7. Defunct configuration options#

CloudCannon no longer uses some configuration options in the Unified Configuration file format. Although leaving these keys in your Site files will not harm your Site, they also will not function. If you want to maintain a clean configuration file, we recommend removing defunct configuration options.

We recommend updating the following keys to their new counterparts:

1. collections_config.*.singular_key

Please use _inputs.*.options.values rather than relying on naming conventions.

2. collections_config.*.output

Collections now default to being output if CloudCannon can determine an output URL. If you need to configure output: false, please use the collections_config.*.disable_url key set to true.

We recommend deleting the following keys:

1. collections_config.*.parser
2. collections_config_override