Add logging

README.md

@@ -125,6 +125,14 @@ The models above will require enough RAM to run them correctly, you should have
 
 ## Release Notes
 
+### 0.1.3
+
+Added logging output for the extension for troubleshooting purposes. Improved error handling and user feedback for invalidated configurations.
+
+<p align="center" width="100%">
+    <img width="20%" src="./docs/LoggingOutput.png">
+</p>
+
 ### 0.1.2
 
 -   Add two new settings:

package.json

@@ -2,7 +2,7 @@
 	"name": "wing-man",
 	"displayName": "Wingman-AI",
 	"description": "Wingman - AI powered assistant to help you write your best code, we won't leave you hanging.",
-	"version": "0.1.2",
+	"version": "0.1.3",
 	"publisher": "WingMan",
 	"license": "MIT",
 	"author": {

src/events/eventEmitter.ts

@@ -1,6 +1,8 @@
 import * as vscode from "vscode";
 
 class EventEmitter {
+	public _onFatalError: vscode.EventEmitter<void> =
+		new vscode.EventEmitter<void>();
 	public _onQueryStart: vscode.EventEmitter<void> =
 		new vscode.EventEmitter<void>();
 	public _onQueryComplete: vscode.EventEmitter<void> =

src/providers/loggingProvider.ts

@@ -0,0 +1,24 @@
+import * as vscode from "vscode";
+
+class LoggingProvider {
+	dbgChannel: vscode.OutputChannel;
+
+	constructor() {
+		this.dbgChannel = vscode.window.createOutputChannel("Wingman");
+	}
+
+	public logInfo(message: string): void {
+		this.dbgChannel.appendLine(
+			`${new Date().toLocaleString()} - [info] ${message}`
+		);
+	}
+
+	public logError(message: string): void {
+		this.dbgChannel.appendLine(
+			`${new Date().toLocaleString()} - [error] ${message}`
+		);
+	}
+}
+
+const loggingProvider = new LoggingProvider();
+export { loggingProvider };

src/providers/statusBarProvider.ts

@@ -3,6 +3,10 @@ import { eventEmitter } from "../events/eventEmitter";
 
 export class ActivityStatusBar {
 	activityStatusBarItem: vscode.StatusBarItem;
+	isInErrorState: boolean = false;
+
+	public readonly onFatalError: vscode.Event<void> =
+		eventEmitter._onFatalError.event;
 
 	public readonly onQueryStart: vscode.Event<void> =
 		eventEmitter._onQueryStart.event;
@@ -26,14 +30,25 @@ export class ActivityStatusBar {
 		this.onQueryComplete(() => {
 			this.TogglePending(false);
 		});
+
+		this.onFatalError(() => {
+			this.ToggleError();
+		});
 	}
 
 	public TogglePending(pending: boolean) {
+		if (this.isInErrorState) return;
+
 		this.activityStatusBarItem.text = `${
 			pending ? "$(sync~spin)" : "$(wingman-logo)"
 		} Wingman`;
 	}
 
+	public ToggleError() {
+		this.isInErrorState = true;
+		this.activityStatusBarItem.text = "$(testing-error-icon) Wingman";
+	}
+
 	dispose() {
 		this.activityStatusBarItem?.dispose();
 	}

src/service/base.ts

@@ -3,6 +3,7 @@ import { Ollama } from "./ollama/ollama";
 import { Settings } from "../types/Settings";
 import { HuggingFace } from "./huggingface/huggingface";
 import { AIModel } from "../types/Models";
+import { loggingProvider } from "../providers/loggingProvider";
 
 export function GetProviderFromSettings(): AIProvider {
 	const config = vscode.workspace.getConfiguration("Wingman");
@@ -12,7 +13,7 @@ export function GetProviderFromSettings(): AIProvider {
 		?.toLocaleLowerCase()
 		.trim();
 
-	console.log(`AI Provider: ${aiProvider} found.`);
+	loggingProvider.logInfo(`AI Provider: ${aiProvider} found.`);
 
 	if (aiProvider === "huggingface") {
 		return new HuggingFace();

src/service/huggingface/huggingface.ts

@@ -4,6 +4,8 @@ import { Settings, defaultMaxTokens } from "../../types/Settings";
 import { HuggingFaceAIModel } from "../../types/Models";
 import { CodeLlama } from "./models/codellama";
 import { Mistral } from "./models/mistral";
+import { loggingProvider } from "../../providers/loggingProvider";
+import { eventEmitter } from "../../events/eventEmitter";
 
 type HuggingFaceRequest = {
 	inputs: string;
@@ -37,47 +39,57 @@ export class HuggingFace implements AIProvider {
 		const huggingFaceConfig =
 			config.get<Settings["huggingface"]>("HuggingFace");
 
-		console.log("HuggingFace settings loaded: ", huggingFaceConfig);
+		loggingProvider.logInfo(
+			`HuggingFace settings loaded: ${JSON.stringify(huggingFaceConfig)}`
+		);
 
-		if (huggingFaceConfig) {
-			this.settings = huggingFaceConfig;
+		if (!huggingFaceConfig) {
+			this.handleError("Unable to log HuggingFace configuration.");
+			return;
+		}
 
-			if (!this.settings.apiKey.trim()) {
-				vscode.window.showErrorMessage(
-					"Hugging Face API key is required."
-				);
-				throw new Error("Missing Hugging Face API key.");
-			}
+		this.settings = huggingFaceConfig!;
 
-			this.chatModel = this.getChatModel(this.settings.chatModel);
-			this.codeModel = this.getCodeModel(this.settings.codeModel);
+		if (!this.settings.apiKey.trim()) {
+			const errorMsg = "Hugging Face API key is required.";
+			vscode.window.showErrorMessage(errorMsg);
+			loggingProvider.logInfo(errorMsg);
+			throw new Error(errorMsg);
 		}
+
+		this.chatModel = this.getChatModel(this.settings.chatModel);
+		this.codeModel = this.getCodeModel(this.settings.codeModel);
 	}
 
-	private getCodeModel(codeModel: string): HuggingFaceAIModel {
+	private handleError(message: string) {
+		vscode.window.showErrorMessage(message);
+		loggingProvider.logError(message);
+		eventEmitter._onFatalError.fire();
+		throw new Error(message);
+	}
+
+	private getCodeModel(codeModel: string): HuggingFaceAIModel | undefined {
 		if (codeModel.includes("codellama")) {
 			return new CodeLlama();
 		} else if (codeModel.includes("mistral")) {
 			return new Mistral();
-		} else {
-			vscode.window.showErrorMessage(
-				"Invalid code model name, currently code supports the CodeLlama model."
-			);
-			throw new Error("Invalid code model name");
 		}
+
+		this.handleError(
+			"Invalid code model name, currently code supports the CodeLlama model."
+		);
 	}
 
-	private getChatModel(chatModel: string): HuggingFaceAIModel {
+	private getChatModel(chatModel: string): HuggingFaceAIModel | undefined {
 		if (chatModel.includes("codellama")) {
 			return new CodeLlama();
 		} else if (chatModel.includes("mistral")) {
 			return new Mistral();
-		} else {
-			vscode.window.showErrorMessage(
-				"Invalid chat model name, currently chat supports the Mistral model."
-			);
-			throw new Error("Invalid chat model name");
 		}
+
+		this.handleError(
+			"Invalid chat model name, currently chat supports the Mistral model."
+		);
 	}
 
 	private getSafeUrl() {
@@ -94,7 +106,7 @@ export class HuggingFace implements AIProvider {
 		signal: AbortSignal
 	) {
 		if (signal.aborted) {
-			return null;
+			return undefined;
 		}
 		return fetch(new URL(`${this.getSafeUrl()}${modelName}`), {
 			method: "POST",
@@ -112,10 +124,26 @@ export class HuggingFace implements AIProvider {
 		modelName: string,
 		signal: AbortSignal
 	) {
-		const response = await this.fetchModelResponse(
-			payload,
-			modelName,
-			signal
+		const startTime = new Date().getTime();
+		let response: Response | undefined;
+
+		try {
+			response = await this.fetchModelResponse(
+				payload,
+				modelName,
+				signal
+			);
+		} catch (error) {
+			loggingProvider.logError(
+				`HuggingFace - chat request with model: ${modelName} failed with the following error: ${error}`
+			);
+		}
+
+		const endTime = new Date().getTime();
+		const executionTime = (endTime - startTime) / 1000;
+
+		loggingProvider.logInfo(
+			`HuggingFace - chat execution time: ${executionTime} seconds`
 		);
 
 		if (!response?.body) {
@@ -141,6 +169,8 @@ export class HuggingFace implements AIProvider {
 		ending: string,
 		signal: AbortSignal
 	): Promise<string> {
+		const startTime = new Date().getTime();
+
 		const codeRequestOptions: HuggingFaceRequest = {
 			inputs: this.codeModel!.CodeCompletionPrompt.replace(
 				"{beginning}",
@@ -159,10 +189,25 @@ export class HuggingFace implements AIProvider {
 			},
 		};
 
-		const response = await this.fetchModelResponse(
-			codeRequestOptions,
-			this.settings?.codeModel!,
-			signal
+		let response: Response | undefined;
+
+		try {
+			response = await this.fetchModelResponse(
+				codeRequestOptions,
+				this.settings?.codeModel!,
+				signal
+			);
+		} catch (error) {
+			loggingProvider.logError(
+				`HuggingFace - code completion request with model ${this.settings?.codeModel} failed with the following error: ${error}`
+			);
+		}
+
+		const endTime = new Date().getTime();
+		const executionTime = (endTime - startTime) / 1000;
+
+		loggingProvider.logInfo(
+			`HuggingFace - Code Completion execution time: ${executionTime} seconds`
 		);
 
 		if (!response?.body) {
@@ -205,6 +250,7 @@ export class HuggingFace implements AIProvider {
 
 		this.clearChatHistory();
 
+		//left incase HF implements streaming.
 		for await (const chunk of this.generate(
 			chatPayload,
 			this.settings?.chatModel!,