Constructor
new DataManager(args)
Parameters:
| Name | Type | Description |
|---|---|---|
args |
DataManagerConstructorArgs.<M, E> |
- Source:
Members
dataOperationsOverrideBehavior
- Source:
dataOperationsRecords
Using stack so later can debug order of requests
- Source:
deleteDataPipeline :DeleteDataPipelineWorker.<M>
Type:
- Source:
loadNewDataPipeline :LoadNewDataPipelineWorker.<M>
Type:
- Source:
masterWorkingModel
- Source:
serverSideDataLoadPipeline
- Deprecated:
- - new hydration logic
- Source:
serverSideOptions
- Source:
updateDataPipeline :UpdateDataPipelineWorker.<M>
Type:
- Source:
uploadDataPipeline :UploadDataPipelineWorker.<M>
Type:
- Source:
viewManagersPendingInit
- Source:
(static) _ARRAY_SELF_TYPE
- Source:
(static) _CANCELLED_DATA_OP
- Source:
(static) _MODEL_ROOT_SCOPE
- Source:
(static) _NESTED_SCOPE_KEY_SPLITTER
- Source:
(static) _SCOPED_ARRAY_LITERAL
- Source:
(static) _SERVER_SIDE_DATA_ATTRS
- Deprecated:
- Yes
- Source:
(static) _SERVER_SIDE_PASSED
- Source:
Methods
bulkCreateModels()
- Source:
commitBulkModels()
- Source:
commitModel()
CHECK WARNING AT COMMIT
So, if working on whole model, to update well, we need an algo change
Need to go nest deep. Based on committed
Change values only given explicitly in temp
recursive on keys going one level deep, node (no child), go through all keys for it,
then return for the parent to finish.
So, depth-first search?*
- Source:
comparator()
- Source:
createAndCommitModel()
- Source:
createModel()
Creates a new model and commits new data to it
- Source:
dataLength()
- Source:
deleteCompleteModel(modelId)
Parameters:
| Name | Type | Description |
|---|---|---|
modelId |
string |
- Source:
deleteData()
- Source:
flushAllData()
- Source:
flushModelTemp()
- Source:
flushScopedData()
Passing MODEL_ROOT for modelId and scope will flush all data
Else, a modelId MUST have a valid scope
Cancels ALL data operations for the given scope
- Source:
getAllViewManagersForScope(scope)
Parameters:
| Name | Type | Description |
|---|---|---|
scope |
- Source:
getDe_MappedOperationScope(mappedScope)
Demaps a scope. mappedDataId goes LAST in mapping, after _FOR_MAPPED_ keyword.
Parameters:
| Name | Type | Description |
|---|---|---|
mappedScope |
string |
- Source:
getMappedDataOperationScope(scope, mappedDataId) → {string}
Parameters:
| Name | Type | Description |
|---|---|---|
scope |
ReqScope | |
mappedDataId |
string |
- Source:
Returns:
- Type
- string
getModel()
Returns copy of stored model (COMMITTED)
Copy to avoid external manipulation not using data manager channels
- Source:
getModelId()
- Source:
getModelInIndex()
Change to getModelInPosition to rhyme with queue?
- Source:
getNewDataScopedToRequest()
- Source:
getOrderedArrayIndicesForMappedDataId(scope, mappedDataId)
Parameters:
| Name | Type | Description |
|---|---|---|
scope |
||
mappedDataId |
string |
- Source:
getScopedModel()
- Source:
getScopedModelFromRef()
ACCESS THIS IF AND ONLY IF YOU'VE REDUCED YOUR SCOPE TO THE DATA SPACE
- Source:
getValidDataOperationsStack(recordsStamp, buildNewopt)
Parameters:
| Name | Type | Attributes | Description |
|---|---|---|---|
recordsStamp |
string | ||
buildNew |
boolean |
<optional> |
- Source:
Returns:
getViewManager()
- Source:
hasData()
- Source:
informObserversOfMutationState(scope, cb)
Parameters:
| Name | Type | Description |
|---|---|---|
scope |
||
cb |
- Source:
(async) initDataManagerServerSide()
Called to initialize server-side data for the data manager
- Source:
initViewManagersInWait()
- Deprecated:
- Yes
- Source:
loadData()
- Source:
(async) loadServerSideData() → {Promise.<void>}
- Deprecated:
- PUT IN PIPELINE THAT CAN BE CANCELLED
- Source:
Returns:
- Type
- Promise.<void>
mergeScopedDataToModel()
CONFIRM THE ALGO
- Source:
onPostDataOperation()
- Source:
overwriteModel()
Overwrite temp in model to new value
- Source:
preProcessDataOperation()
FINISH CANCELLATIONS
REJECT IF NEW OPERATION PARENT OR CHILD IN SCOPE TO RUNNING (SO LET INTERNAL ONES FINISH FIRST AS PER MODEL OBJECT)
As per data operations override behavior
- Source:
recursiveValueReference(scope, prevKey, orderedArrayIndices, referencedModel, mainModel, stopAtNodeforRef) → {ValueTypeOfNested.<MainModel, ReqScope>|ValueTypeOfArrayOnly.<M, ReqScope>}
Use this to recursively reference a model and its type.
Not as straightforward especially when working with arrays
TO MAKE WORK EASIER, MANAGER HOLDS REFERENCE TO WHERE CHANGE SHOULD HAPPEN. CHANGES THAT TO NEW VALUE.
BETTER SOLUTION TBF
IF OBJECT, MERGE. IF LITERAL, OVERWRITE. YES.
Parameters:
| Name | Type | Description |
|---|---|---|
scope |
ReqScope | |
prevKey |
string | |
orderedArrayIndices |
ViewManagerOrderedArrayIndices | This provided by view managers. Can infer its parent for index of its model id to create right order. So, keeping child list which can also be used to invoke a new build for recyclable lists |
referencedModel |
ValueTypeOfNested.<MainModel, ReqScope> | |
mainModel |
MainModel | |
stopAtNodeforRef |
boolean |
- Source:
Returns:
//Returning value type of array too, if deconstructed
- Type
- ValueTypeOfNested.<MainModel, ReqScope> | ValueTypeOfArrayOnly.<M, ReqScope>
reduceModelToProperties()
If it return undefined, this property doesn't exist in target model
- Source:
requestDataUpload(modelID, options, scope, newData, mappedDataId, reqAddr, requestedMutation, overrideNetworkInterface, overrideNetworkInterfaceScopeopt)
Parameters:
| Name | Type | Attributes | Description |
|---|---|---|---|
modelID |
string | ||
options |
SendDataOptions | ||
scope |
ReqScope | ||
newData |
ValueTypeOfNested.<M, ReqScope> | ||
mappedDataId |
string | ||
reqAddr |
string | ||
requestedMutation |
DataManagerMutations | ||
overrideNetworkInterface |
|||
overrideNetworkInterfaceScope |
NestedParentKeysOf.<M, ReqScope> |
<optional> |
- Source:
Returns:
(async) runDataMutation()
A helper method to ensure all mutations follow a basic or given flow for mutation execution.
Help homogenize future updates, and better trace mutation errors, on a global scope
FOR SCOPE: Always ensure it is the original scope to ensure data integrity passes are done well
Now including mappedDataId per scope to cover for individual array changes, and make them asynchronous
- Source:
- To Do:
-
- ABOVE HAS A PROBLEM cause we can't trace a mapped data id to an individual view manager to avoid side effects. However, I have a solution. For later (use uniqueIds - auto generated) Because of the override scope and how it affects data access, onDataLoadPostProcess can allow you to sort of commit extra data to this scope. However, code disallows this by using original scope (before, found it out by bug where MODEL_ROOT scope temp was null. Thus merge to old failed.) Now, algo STRICTLY commit to original scope. So, allow free use of override to avoid writing same API options severally. For extra data you want to commit that comes from server, silent update once
setDataDeleteAPI()
For data deletes
- Source:
setDataLoadAPI()
For data loads
- Source:
setDataUpdateAPI()
For data updates
- Source:
setDataUploadAPI()
For data uploads
- Source:
setDataWatcher()
- Source:
setScopedAPIDataOpInterfaceObserver()
- Source:
setScopedAPIOptions()
- Source:
setViewManager()
Children scope for view also updated. YESSS
Okay, no. Might be separating some concerns. Fire only relevant scope.
- Source:
silentUpdateModel()
Directly commits - but ensure integrity despite being non-network?. So, run well
- Source:
spawnPartialShellModel()
Creates a partial shell model of the target reference model, fulfilling the scope from root
Added targetReferenceModel because spread operator is breaking with nested objects
- Source:
updateModel()
overrideNetworkInterface allows developer to pass data in more specific scope, but have it balooned to
the one provided network interface scope. Just makes life easier.
- Source:
updateWatchers(mutation, scope, newData, oldData)
Parameters:
| Name | Type | Description |
|---|---|---|
mutation |
DataManagerMutations | |
scope |
S | |
newData |
||
oldData |
- Source:
uploadDataInModel()
- Source:
uploadNewData()
Uploading data NOT in model using the upload data API address - READ MORE
Update spec to loadData, such that for scope lower than MODEL_ROOT, we don't create a new model.
instead, we'll just update the existing
- Source:
valueBasedRecursiveObjectMerge(oldModel, newModel, orderedArrayIndices, afterFirstRunopt)
ONLY CALL FOR MATCHING OBJECT TYPES, WHERE YOU'RE USING SPREAD OPERATOR TO MERGE VALUES AND PROPERTIES
Merges new model into old model
Parameters:
| Name | Type | Attributes | Description |
|---|---|---|---|
oldModel |
object | ||
newModel |
object | ||
orderedArrayIndices |
QueueInstance.<number> | ||
afterFirstRun |
boolean |
<optional> |
- Source: