org.mulesoft.als.suggestions.Core class is the entry point to the suggestions.
init() method returning void/unit future must be called before starting the work with the system.
After that user should call the following method for unit text:
def prepareText(text:String, offset:Int, syntax:Syntax):String
This method modifies the potentially broken text to make it readable by the parser.
Prepared/modified text should be passed as a unit text to other suggestions methods unless said otherwise.
org.mulesoft.als.suggestions.CompletionProvider is dedicated to return actual suggestions.
class CompletionProvider {
def withConfig(cfg:ICompletionConfig):CompletionProvider
def suggest: Future[Seq[ISuggestion]]
def suggest(filterByPrefix:Boolean): Future[Seq[ISuggestion]]
...
The initial task is to construct the proper provider. Usual code looks like this:
val completionConfig = new CompletionConfig()
.withEditorStateProvider(editorStateProvider)
.withAstProvider(astProvider)
.withFsProvider(platformFSProvider)
.withOriginalContent(unmodifiedContent)
CompletionProvider().withConfig(completionConfig)
Editor state provider provides unit text:
trait IEditorStateProvider {
def getText: String
def getPath: String
def getBaseName: String
def getOffset: Int
}
This text must be first prepared by prepareText method.
AST provider provides the pre-parsed high-level AST:
trait IASTProvider {
def getASTRoot: IHighLevelNode
def getSelectedNode: Option[IParseResult]
def language:Vendor
def syntax: Syntax
}
FS provider should provide the AST:
trait IFSProvider {
def content(fullPath: String): Future[String]
def dirName(fullPath: String): String
def existsAsync(path: String): Future[Boolean]
def resolve(contextPath: String, relativePath: String): Option[String]
def isDirectory(fullPath: String): Boolean
def readDirAsync(path: String): Future[Seq[String]]
def isDirectoryAsync(path: String): Future[Boolean]
}
And withOriginalContent should contain unit contents before it was pre-patched to fix grammar errors.
The result is obtained by suggest method, it returns a list of:
trait ISuggestion {
def text: String
def description: String
def displayText: String
def prefix: String
def category: String
}
Where text is actual suggestion test, displayText is what is displayed to the user, and prefix is what should be replaced by suggestion text.
description and category provide additional suggestion data for the users.
Besides scala interface, suggestions module also contains JS proxy, which simplifies usage from JS tooling.
This interface assumes that suggestions module launches code parsing and text patching internally.
Those who wish to control this process can create their own JS proxy basing on this one.
Initialization should be performed first by calling
def init(fsProvider: IFSProvider): js.Promise[Unit]
FSProvider instance must provide file system information:
trait IFSProvider extends js.Object {
def contentAsync(fullPath: String): Promise[String] = js.native
def dirName(fullPath: String): String = js.native
def name(fullPath: String): String = js.native
def existsAsync(path: String): Promise[Boolean] = js.native
def resolve(contextPath: String, relativePath: String): Option[String] = js.native
def isDirectory(fullPath: String): Boolean = js.native
def readDirAsync(path: String): Promise[js.Array[String]] = js.native
def isDirectoryAsync(path: String): Promise[Boolean] = js.native
def content(fullPath: String): String
def exists(fullPath: String): Boolean
def readDir(fullPath: String): js.Array[String]
def separatorChar(): String
}
After initialization is finished (promise is resolved), suggestions can be obtained by calling:
def suggest(language: String, url: String, position: Int): js.Promise[js.Array[ISuggestion]]
Here an array of suggestions is returned from the promise:
class ISuggestion (
val text: String,
val description: String,
val displayText: String,
val prefix: String,
val category: String
) extends js.Object
{
}
Where text is actual suggestion test, displayText is what is displayed to the user, and prefix is what should be replaced by suggestion text.
description and category provide additional suggestion data for the users.