This package is a simple loader to load data from a PocketBase database into Astro using the Astro Loader API introduced in Astro 5.

In your content configuration file, you can use the pocketbaseLoader function to use your PocketBase database as a data source.

import { pocketbaseLoader } from "astro-loader-pocketbase";
import { defineCollection } from "astro:content";

const blog = defineCollection({
  loader: pocketbaseLoader({
    url: "https://<your-pocketbase-url>",
    collectionName: "<collection-in-pocketbase>"
  })
});

export const collections = { blog };

Remember that due to the nature Astros Content Layer lifecycle, the loader will only fetch entries at build time, even when using on-demand rendering. If you want to update your deployed site with new entries, you need to rebuild it.

_{When running the dev server, you can trigger a reload by using s + enter.}

Since PocketBase 0.23.0, the updated field is not mandatory anymore. This means that the loader can't automatically detect when an entry has been modified. To enable incremental builds, you need to provide the name of a field in your collection that stores the last update date of an entry.

const blog = defineCollection({
  loader: pocketbaseLoader({
    ...options,
    updatedField: "<field-in-collection>"
  })
});

When this field is provided, the loader will only fetch entries that have been modified since the last build. Ideally, this field should be of type autodate and have the value "Update" or "Create/Update" in the PocketBase dashboard. This ensures that the field is automatically updated when an entry is modified.

After generating the schema (see below), the loader will automatically parse the content of the entries (e.g. transform ISO dates to Date objects, coerce numbers, etc.).

You can also specify a field or an array of fields to use as content. This content will then be used when calling the render function of Astros content collections.

const blog = defineCollection({
  loader: pocketbaseLoader({
    ...options,
    contentFields: "<field-in-collection>"
  })
});

Since the goal of the render function is to render the content as HTML, it's recommended to use a field of type editor (rich text) in PocketBase as content.

If you specify an array of fields, the loader will wrap the content of these fields in a <section> and concatenate them; each section will have an id attribute matching the content field name.

PocketBase can store images and files for each entry in a collection. While the API only returns the filenames of these images and files, the loader will out of the box transform these filenames to the actual URLs of the files. This doesn't mean that the files are downloaded during the build process. But you can directly use these URLs in your Astro components to display images or link to the files.

By default, the loader will use the id field of the collection as the unique identifier. If you want to use another field as the id, e.g. a slug of the title, you can specify this field via the idField option.

const blog = defineCollection({
  loader: pocketbaseLoader({
    ...options,
    idField: "<field-in-collection>"
  })
});

Please note that the id should be unique for every entry in the collection. The loader will also automatically convert the value into a slug to be easily used in URLs. It's recommended to use e.g. the title of the entry to be easily searchable and readable. Do not use e.g. rich text fields as ids.

By default the loader will fetch all entries in the specified collection. If you want to restrict the entries to a specific subset, you can use the filter option.

const blog = defineCollection({
  loader: pocketbaseLoader({
    ...options,
    filter: "<filter>"
  })
});

For example, if you want to only fetch entries that are released but not deleted, you can use "release >= @now && deleted = false". This filter will be added to the PocketBase API request and will only fetch entries that match the filter. This is in addition to the built-in filtering of the loader, which handles the incremental builds using the updated field. For more information on how to use filters, check out the PocketBase documentation.

By default, the loader fetches all fields for each entry in your PocketBase collection. If you want to optimize data loading or restrict the fields available in your content collection, you can use the fields option.

const blog = defineCollection({
  loader: pocketbaseLoader({
    ...options,
    fields: ["title", "summary", "coverImage"]
  })
});

This parameter will be added to the PocketBase API request and will only return these fields for each entry. Additional system fields like id, collectionName and collectionId, as well as any fields specified for idField, updatedField or contentFields will be added automatically. For further details on field selection, see the PocketBase API documentation.

The loader can automatically generate types for your collection. This is useful for type checking and autocompletion in your editor. These types can be generated in two ways:

To use the live remote schema, you need to provide superuser credentials for the PocketBase instance.

const blog = defineCollection({
  loader: pocketbaseLoader({
    ...options,
    superuserCredentials: {
      email: "<superuser-email>",
      password: "<superuser-password>",
      // or
      impersonateToken: "<superuser-impersonate-token>"
    }
  })
});

It's recommended to use an impersonate token (API token) instead of the email and password, as this is more secure and can be easily revoked.

Under the hood, the loader will use the PocketBase API to fetch the schema of your collection and generate types with Zod based on that schema.

If you don't want to provide superuser credentials (e.g. in a public repository), you can also provide the schema manually via a JSON file.

const blog = defineCollection({
  loader: pocketbaseLoader({
    ...options,
    localSchema: "<path-to-schema-file>"
  })
});

In PocketBase you can export the schema of the whole database to a pb_schema.json file. If you provide the path to this file, the loader will use this schema to generate the types locally.

When superuser credentials are provided, the loader will always use the remote schema since it's more up-to-date.

If you don't want to use the automatic type generation, you can also provide your own schema manually. This manual schema will always override the automatic type generation.

By default PocketBase reports number and boolean fields as not required, even though the API will always return 0 and false respectively if no value is set. This means that the loader will add undefined to the type of these fields. If you want to enforce that these fields are always present, you can set the improveTypes option to true.

const blog = defineCollection({
  loader: pocketbaseLoader({
    ...options,
    improveTypes: true
  })
});

This will remove undefined from the type of these fields and mark them as required.

For more information on how to enable and use live content collections, please refer to the Astro documentation.

The options for this packages loader are similar to the regular PocketBase loader:

const blogLive = defineLiveCollection({
  loader: experimentalPocketbaseLiveLoader({
    url: "https://<your-pocketbase-url>",
    collectionName: "<collection-in-pocketbase>"
  })
});

export const collections = { blogLive };

You can also specify additional options for the live content loader, such as filters and content fields. For access to private collections and hidden fields, you need to provide superuser credentials, also similar to the regular PocketBase loader.

To access a single entry you can use the default getLiveEntry function:

// Get a single entry by its id
const entry = await getLiveEntry("blogLive", { id: "<entry-id>" });

In here you need to specify the id of the entry you want to access. In contrast to the regular content loader, the live content loader can not use a custom id field, so this needs to be the primary id used by PocketBase.

To access a collection of entries, you can use the default getLiveCollection function:

// Get a whole collection of entries
const entries = await getLiveCollection("blogLive");

// Get a filtered and paginated collection of entries
const entries2 = await getLiveCollection("blogLive", {
  filters: "release >= @now && deleted = false",
  sort: "-created,id",
  page: 1,
  perPage: 10
});

By default, the loader will fetch all entries in the collection. But you can also specify additional options, such as filters (will be added to the global filters), sorting, and pagination.

The loader automatically adds cache hints for each entry / collection. By default only the tags hint is used, which includes the collection name and the id of the entry.

If you also want to use the lastModified hint, just tell the loader which field to use for this date:

const blogLive = defineLiveCollection({
  loader: experimentalPocketbaseLiveLoader({
    ...options,
    updatedField: "<field-in-collection>"
  })
});

Live content loaders do not (yet 🤞🏼) support zod schemas and thus schema generation and entry transformation. This means that options like custom ids or image and file transformations are not available. Dates will also be treated as plain ISO strings instead of date objects.

If you want an accompanying type for your live content, you can use the experimental.liveTypesOnly option on the regular loader. This will cause it to skip the data fetching and only generate types, fitting the live content structure.

const blogTypes = defineCollection({
  loader: pocketbaseLoader({
    ...options,
    experimental: {
      liveTypesOnly: true
    }
  })
});

const blogLive = defineLiveCollection({
  loader:
    experimentalPocketbaseLiveLoader<CollectionEntry<"blogTypes">["data"]>(
      options
    )
});

If you want to access a private collection or hidden fields, you also need to provide superuser credentials. Otherwise, you need to make the collection public in the PocketBase dashboard.

Generally, it's not recommended to use private collections, especially when users should be able to see images or other files stored in the collection.

PocketBase can store arbitrary JSON data in a json field. While this data is checked to be valid JSON, there's no schema attached or enforced. Theoretically, every entry in a collection can have a different structure in such a field. This is why by default the loader will treat these fields as unknown and will not generate types for them.

If you have your own schema for these fields, you can provide them via the jsonSchemas option. The keys of this record should be the field names of your json fields, while the values are Zod schemas. This way, the loader will also generate types for these fields and validate the data against these schemas. So be sure that the schema is correct, up-to-date and all entries in the collection adhere to it.