fix(amazonq): Render first response before receiving all paginated inline completion results (#7663)

## Problem
Previous the pagination API call was blocking and it does not render
until all pagination API calls are done.
Without a force refresh of the inline ghost text, paginated response
will not be rendered.

## Solution
1. Keep doing paginated API call in the background.
2. Refresh the ghost text of inline completion when user press Left or
Right key to add paginated responses by calling VS Code native commands.


Note that we can do such refresh below whenever the pagination session
finish but that will result in inline completion flickering, hence it is
better to do it on demand when Left or Right is pressed.

```
await vscode.commands.executeCommand('editor.action.inlineSuggest.hide')
 await vscode.commands.executeCommand('editor.action.inlineSuggest.trigger')
```

---

- Treat all work as PUBLIC. Private `feature/x` branches will not be
squash-merged at release time.
- Your code changes must meet the guidelines in
[CONTRIBUTING.md](https://github.com/aws/aws-toolkit-vscode/blob/master/CONTRIBUTING.md#guidelines).
- License: I confirm that my contribution is made under the terms of the
Apache 2.0 license.

Author: leigaol

packages/amazonq/.changes/next-release/Bug Fix-91380b87-5955-4c15-b762-31e7f1c71575.json

@@ -0,0 +1,4 @@
+{
+	"type": "Bug Fix",
+	"description": "Render first response before receiving all paginated inline completion results"
+}

packages/amazonq/package.json

@@ -917,12 +917,12 @@
             },
             {
                 "key": "right",
-                "command": "editor.action.inlineSuggest.showNext",
+                "command": "aws.amazonq.showNext",
                 "when": "inlineSuggestionVisible && !editorReadonly && aws.codewhisperer.connected"
             },
             {
                 "key": "left",
-                "command": "editor.action.inlineSuggest.showPrevious",
+                "command": "aws.amazonq.showPrev",
                 "when": "inlineSuggestionVisible && !editorReadonly && aws.codewhisperer.connected"
             },
             {

packages/amazonq/src/app/inline/recommendationService.ts

@@ -2,7 +2,6 @@
  * Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
  * SPDX-License-Identifier: Apache-2.0
  */
-
 import {
     InlineCompletionListWithReferences,
     InlineCompletionWithReferencesParams,
@@ -80,7 +79,7 @@ export class RecommendationService {
                     nextToken: request.partialResultToken,
                 },
             })
-            let result: InlineCompletionListWithReferences = await languageClient.sendRequest(
+            const result: InlineCompletionListWithReferences = await languageClient.sendRequest(
                 inlineCompletionWithReferencesRequestType.method,
                 request,
                 token
@@ -120,18 +119,10 @@ export class RecommendationService {
                 getLogger().info(
                     'Suggestion type is COMPLETIONS. Start fetching for more items if partialResultToken exists.'
                 )
-                try {
-                    while (result.partialResultToken) {
-                        const paginatedRequest = { ...request, partialResultToken: result.partialResultToken }
-                        result = await languageClient.sendRequest(
-                            inlineCompletionWithReferencesRequestType.method,
-                            paginatedRequest,
-                            token
-                        )
-                        this.sessionManager.updateSessionSuggestions(result.items)
-                    }
-                } catch (error) {
-                    languageClient.warn(`Error when getting suggestions: ${error}`)
+                if (result.partialResultToken) {
+                    this.processRemainingRequests(languageClient, request, result, token).catch((error) => {
+                        languageClient.warn(`Error when getting suggestions: ${error}`)
+                    })
                 }
             } else {
                 // Skip fetching for more items if the suggesion is EDITS. If it is EDITS suggestion, only fetching for more
@@ -140,11 +131,6 @@ export class RecommendationService {
                 getLogger().info('Suggestion type is EDITS. Skip fetching for more items.')
                 this.sessionManager.updateActiveEditsStreakToken(result.partialResultToken)
             }
-
-            // Close session and finalize telemetry regardless of pagination path
-            this.sessionManager.closeSession()
-            TelemetryHelper.instance.setAllPaginationEndTime()
-            options.emitTelemetry && TelemetryHelper.instance.tryRecordClientComponentLatency()
         } catch (error: any) {
             getLogger().error('Error getting recommendations: %O', error)
             // bearer token expired
@@ -167,4 +153,31 @@ export class RecommendationService {
             }
         }
     }
+
+    private async processRemainingRequests(
+        languageClient: LanguageClient,
+        initialRequest: InlineCompletionWithReferencesParams,
+        firstResult: InlineCompletionListWithReferences,
+        token: CancellationToken
+    ): Promise<void> {
+        let nextToken = firstResult.partialResultToken
+        while (nextToken) {
+            const request = { ...initialRequest, partialResultToken: nextToken }
+
+            const result: InlineCompletionListWithReferences = await languageClient.sendRequest(
+                inlineCompletionWithReferencesRequestType.method,
+                request,
+                token
+            )
+            this.sessionManager.updateSessionSuggestions(result.items)
+            nextToken = result.partialResultToken
+        }
+
+        this.sessionManager.closeSession()
+
+        // refresh inline completion items to render paginated responses
+        // All pagination requests completed
+        TelemetryHelper.instance.setAllPaginationEndTime()
+        TelemetryHelper.instance.tryRecordClientComponentLatency()
+    }
 }

packages/amazonq/src/app/inline/sessionManager.ts

@@ -24,6 +24,7 @@ export interface CodeWhispererSession {
 export class SessionManager {
     private activeSession?: CodeWhispererSession
     private _acceptedSuggestionCount: number = 0
+    private _refreshedSessions = new Set<string>()
 
     constructor() {}
 
@@ -86,4 +87,20 @@ export class SessionManager {
     public clear() {
         this.activeSession = undefined
     }
+
+    // re-render the session ghost text to display paginated responses once per completed session
+    public async maybeRefreshSessionUx() {
+        if (
+            this.activeSession &&
+            !this.activeSession.isRequestInProgress &&
+            !this._refreshedSessions.has(this.activeSession.sessionId)
+        ) {
+            await vscode.commands.executeCommand('editor.action.inlineSuggest.hide')
+            await vscode.commands.executeCommand('editor.action.inlineSuggest.trigger')
+            if (this._refreshedSessions.size > 1000) {
+                this._refreshedSessions.clear()
+            }
+            this._refreshedSessions.add(this.activeSession.sessionId)
+        }
+    }
 }

packages/amazonq/src/lsp/client.ts

@@ -265,6 +265,14 @@ async function onLanguageServerReady(
 
     toDispose.push(
         inlineManager,
+        Commands.register('aws.amazonq.showPrev', async () => {
+            await sessionManager.maybeRefreshSessionUx()
+            await vscode.commands.executeCommand('editor.action.inlineSuggest.showPrevious')
+        }),
+        Commands.register('aws.amazonq.showNext', async () => {
+            await sessionManager.maybeRefreshSessionUx()
+            await vscode.commands.executeCommand('editor.action.inlineSuggest.showNext')
+        }),
         Commands.register({ id: 'aws.amazonq.invokeInlineCompletion', autoconnect: true }, async () => {
             await vscode.commands.executeCommand('editor.action.inlineSuggest.trigger')
         }),