dataset provides a way to manage your JSON documents. This can be done on your local disk using a technique called a pairtree. Optionally can also be done by storing the JSON Objects in a SQL database table.
The pairtree layout is described by “collection.json” and “keymaps.json” documents. These are located in the root folder of the collection.
A JSON document may also have one or more “attachments”. These are stored in their own pairtree under the “attachments” sub-directory of the collection. Attachments are supported for both pairtree storage and SQL storage of JSON objects in a collection.
Pairtrees ensures sequence of characters in a object’s key will not collide with and is legal on common file systems. E.g. storing the document “hello-world.json” with the attachment “smiles.png” in a collection named “C.ds” would result in paths like
`C.ds/pairtree/he/ll/o-/wo/rl/d/hello-world.json` and
`C.ds/attachments/he/ll/o-/wo/rl/d/smiles.png".Attachments are experimental and how they are handled may change in the future.
In a collection’s root directory you will find two or three JSON documents.
If you’re using SQLite3 as your JSON document storage engine you can choose to include your SQLite3 database file in the directory too, if so “collection.db” is a good name. That way if you zip up your collection and share it with a friend the JSON documents will travel appropriately.
The directory layout looks like: