Clock that keeps track of logical timestamps of the current editing session and logical clocks of all known peers.
OptionalonbeforepatchCallback called before every applyPatch call.
OptionalonbeforeresetCallback called before model isi reset using the .reset() method.
OptionalonpatchCallback called after every applyPatch call.
OptionalonresetCallback called after model has been reset using the .reset() method.
Root of the JSON document is implemented as Last Write Wins Register, so that the JSON document does not necessarily need to be an object. The JSON document can be any JSON value.
Static ReadonlysidGenerates a random session ID. Use this method to generate a session ID for a new user. Store the session ID in the user's browser or device once and reuse it for all editing sessions of that user.
Generating a new session ID for each editing session will work, however, that is not recommended. If a user generates a new session ID for each editing session, the session clock table will grow indefinitely.
Generates a random session ID up to 53 bits in size, skips first 0xFFFF values, keeping them reserved for future extensions.
Random session ID.
Experimental strictly typed node retrieval API using proxy objects. Automatically resolves nested "val" nodes.
Experimental node retrieval API using proxy objects. Returns a strictly typed proxy wrapper around the value of the root node.
Applies a batch of patches to the document.
A batch, i.e. an array of patches.
Works like applyPatch, but is intended to be used by the local client
for locally generated patches. It checks if the model clock is ahead of
the patch clock and rebases the patch if necessary.
A patch to apply to the document.
Applies a single patch to the document. All mutations to the model must go through this method. (With the only exception of local changes through API, which have an alternative path.)
A patch to apply to the document.
Use this method to generate a random session ID for an existing document. It checks for the uniqueness of the session ID given the current peers in the document. This reduces the chance of collision substantially.
A random session ID that is not used by any peer in the current document.
Strictly types the model and sets the default value of the model, if the document is empty.
The schema to set for this model.
Strictly typed model.
Changes the session ID of the model. By modifying the attached clock vector of the model. Be careful when changing the session ID of the model, as this is an advanced operation.
Use the Model.load method to load a model with the the right session ID, instead of changing the session ID of the model. When in doubt, use the Model.fork method to create a new model with the right session ID.
The new session ID to set for the model.
Serialize this model using "binary" structural encoding.
This model encoded in octets.
Static ReadonlycreateCreate a new JSON CRDT model. If a schema is provided, the model is strictly typed and the default value of the model is set to the default value of the schema.
By default, the model is created with a random session ID and is using
a logical clock. It is also possible to create a model which uses a server
clock by providing the session ID SESSION.SERVER (1).
Create a basic model, without schema and default value:
const model = Model.create();
Create a strictly typed model with a schema and default value:
const schema = s.obj({
ticker: s.con<string>('BODEN'),
name: s.str('Jeo Boden'),
tags: s.arr(
s.str('token'),
),
});
const model = Model.create(schema);
const patch = model.api.flush();
Create a model with a custom session ID for your logical clock:
const schema = s.str('');
const sid = 123456789;
const model = Model.create(schema, sid);
const patch = model.api.flush();
The session ID must be at least 65,536 or higher, see JSON CRDT Patch specification.
To create a model with a server clock, use the SESSION.SERVER, which is
equal to 1:
const model = Model.create(undefined, SESSION.SERVER);
// or
const model = Model.create(undefined, 1);
Finally, you can create a model with your clock vector:
const clock = new ClockVector(123456789, 1);
const model = Model.create(undefined, clock);
Optionalschema: SThe schema (typing and default value) to set for this model.
When a schema is provided, the model is strictly typed and the default
value of the model is set to the value of the schema. Also, you MUST
call model.api.flush() immediately after creating the model to clear
the change buffer of the patch that was created during the initialization
of the model.
Session ID to use for local operations. Defaults to a random session ID generated by Model.sid.
A strictly typed model.
Static ReadonlyfromDecodes a model from a "binary" structural encoding.
Use Model.load instead, if you want to set the session ID of the model and the right schema for the model, during the de-serialization.
Binary blob of a model encoded using "binary" structural encoding.
An instance of a model.
StaticfromInstantiates a model from a collection of patches. The patches are applied to the model in the order they are provided. The session ID of the model is set to the session ID of the first patch.
A collection of initial patches to apply to the model.
A model with the patches applied.
Static ReadonlyloadUn-serializes a model from "binary" structural encoding. The session ID of
the model is set to the provided session ID sid, or the default session
ID of the un-serialized model is used.
Binary blob of a model encoded using "binary" structural encoding.
Optionalsid: numberSession ID to set for the model.
Optionalschema: SAn instance of a model.
Static ReadonlywithCreate a CRDT model which uses server clock. In this model a central server timestamps each operation with a sequence number. Each timestamp consists simply of a sequence number, which was assigned by a server. In this model all operations are approved, persisted and re-distributed to all clients by a central server. The session ID part of the clock is the same for all clients, i.e. SESSION.SERVER.
Optionalschema: SLatest known server sequence number.
CRDT model.
In instance of Model class represents the underlying data structure, i.e. model, of the JSON CRDT document.