• This is a bit of a sticking point, especially when working with large datasets
• For example, if you want to edit fields for an already selected image – you'll need to go into the plugin and then have to manually find that image (which can be laborious when sifting through thousands of assets)
• A future update will provide the ability to 'jump' straight to a selected asset
• However, exposing plugin fields directly on the desk (e.g. via a custom input component) is currently outside the scope of this project

• This is currently due to Sanity studio's file picker component taking precedence over window drag and drop events
• For now, you'll need to manually press the 'upload' button if you want to add images whilst in a selecting context

• Any images downloaded in the plugin are those already processed by Sanity without any image transformations applied
• Please note these are not the original uploaded images: they will likely have a smaller file size and will be stripped of any EXIF data.
• Currently, it's not possible in Sanity to grab these original image assets within the studio - but this may change in future!

• Currently, opt.media.tags on assets aren't accessible via GraphQL. This is because opt is a custom object used by this plugin and not part of Sanity's asset schema.

• This plugin will read and write directly on the asset document itself. This will either a document of type sanity.imageAsset or sanity.fileAsset
• This is analagous to setting values globally across all instances of these assets
• This is in contrast to using the fields property when defining your document schema (on both image and file objects). Values that you define in the fields property can be considered 'local', or bound to the the document where that asset is linked.
• In other words, if you want to set a caption for an image and have that change between different documents – customise the fields property in your document schema's file/image field
• If you want to set values you can query in all instances of that asset (alternate text being a good example), consider setting those in the plugin

The following GROQ query will return an image with additional asset text fields as well as an array of tag names.

Note that tags are namespaced within opt.media and tag names are accessed via the current property (as they're defined as slugs on the tag.media document schema).

*[_id == 'my-document-id'] {
  image {
    asset->{
      _ref,
      _type,
      altText,
      description,
      "tags": opt.media.tags[]->name.current,
      title
    }
  }
}

• ISO, aperture, focal length, exposure time and original date are displayed
• By default, Sanity won't automatically extract EXIF data unless you explicitly tell it to
• Manually tell Sanity to process EXIF metadata by updating your image field options accordingly
• Note that all images uploaded directly within the plugin will include all metadata by default

• This plugin defines the document type media.tag
• All tags are stored as weak references and being a third-party plugin, are stored in the namespaced object opt.media
• This behaviour differs from asset fields such as title, description and altText which are stored directly on the asset as they're part of Sanity's defined asset schema

• If you're not using a custom desk, Sanity attaches custom schema defined by third party plugins to your desk. This is currently the default behaviour
• However, you can override this behaviour by defining your own custom desk with Sanity's structure builder and simply omit the media.tag document type in your definition

• You can create, rename and delete tags from directly within the plugin itself
• It is strongly recommended that you manually delete tags directly from within the plugin – doing so will ensure that (weak) references are removed from any linked assets
• Alternatively, you can delete tags either from the desk (if you're not using a custom desk) or via Sanity's API – just be mindful that any assets previously assigned to deleted tags will have 'hanging' weak references. This won't cause serious issues, but it may cause some false positives when searching. (E.g. a search for 'all assets where tags is not empty' will yield assets that have references to tags that no longer exist)

• Batch mutations are carried out via Sanity transactions. These transactions are atomic, meaning that if one deletion fails (often because it's referenced elsewhere), then all mutations in the transaction will fail and no changes will occur
• To get around this, simply make sure that all assets you've marked for deletion are not referenced – this can be easily accomplished by using a search facet to only show assets which are not in use

• As a rule of thumb, when uploading when accessing the plugin as a tool (e.g. if you've acceessed it via the studio menu), it will look at any incoming files' MIME type. All files of type image/* will be uploaded as sanity.imageAsset whilst everything else will be treated as sanity.fileAsset
• If you upload when using the plugin in a file selection context, these be uploaded as sanity.fileAsset regardless of their MIME type. This is probably not what you want, since images uploaded as files won't have associated metadata nor will they work in Sanity's image pipeline.