Improve ai error handling (#4180)

* Introduce a result type

* Update AI error handling to use the result type

* Handle ollama json parse error

* Migrate using Error as the type that represents errors

* Remove now useless condition

* asdfasdf

* Use andThen

* Correct unit tests

Author: Caleb-T-Owens

app/src/lib/ai/anthropicClient.ts

@@ -1,8 +1,12 @@
 import { SHORT_DEFAULT_COMMIT_TEMPLATE, SHORT_DEFAULT_BRANCH_TEMPLATE } from '$lib/ai/prompts';
+import { type AIClient, type AnthropicModelName, type Prompt } from '$lib/ai/types';
+import { buildFailureFromAny, ok, type Result } from '$lib/result';
 import { fetch, Body } from '@tauri-apps/api/http';
-import type { AIClient, AnthropicModelName, Prompt } from '$lib/ai/types';
 
-type AnthropicAPIResponse = { content: { text: string }[] };
+type AnthropicAPIResponse = {
+	content: { text: string }[];
+	error: { type: string; message: string };
+};
 
 export class AnthropicAIClient implements AIClient {
 	defaultCommitTemplate = SHORT_DEFAULT_COMMIT_TEMPLATE;
@@ -13,7 +17,7 @@ export class AnthropicAIClient implements AIClient {
 		private modelName: AnthropicModelName
 	) {}
 
-	async evaluate(prompt: Prompt) {
+	async evaluate(prompt: Prompt): Promise<Result<string, Error>> {
 		const body = Body.json({
 			messages: prompt,
 			max_tokens: 1024,
@@ -30,6 +34,12 @@ export class AnthropicAIClient implements AIClient {
 			body
 		});
 
-		return response.data.content[0].text;
+		if (response.ok && response.data?.content?.[0]?.text) {
+			return ok(response.data.content[0].text);
+		} else {
+			return buildFailureFromAny(
+				`Anthropic returned error code ${response.status} ${response.data?.error?.message}`
+			);
+		}
 	}
 }

app/src/lib/ai/butlerClient.ts

@@ -1,4 +1,5 @@
 import { SHORT_DEFAULT_BRANCH_TEMPLATE, SHORT_DEFAULT_COMMIT_TEMPLATE } from '$lib/ai/prompts';
+import { map, type Result } from '$lib/result';
 import type { AIClient, ModelKind, Prompt } from '$lib/ai/types';
 import type { HttpClient } from '$lib/backend/httpClient';
 
@@ -12,16 +13,19 @@ export class ButlerAIClient implements AIClient {
 		private modelKind: ModelKind
 	) {}
 
-	async evaluate(prompt: Prompt) {
-		const response = await this.cloud.post<{ message: string }>('evaluate_prompt/predict.json', {
-			body: {
-				messages: prompt,
-				max_tokens: 400,
-				model_kind: this.modelKind
-			},
-			token: this.userToken
-		});
+	async evaluate(prompt: Prompt): Promise<Result<string, Error>> {
+		const response = await this.cloud.postSafe<{ message: string }>(
+			'evaluate_prompt/predict.json',
+			{
+				body: {
+					messages: prompt,
+					max_tokens: 400,
+					model_kind: this.modelKind
+				},
+				token: this.userToken
+			}
+		);
 
-		return response.message;
+		return map(response, ({ message }) => message);
 	}
 }

app/src/lib/ai/ollamaClient.ts

@@ -1,5 +1,6 @@
 import { LONG_DEFAULT_BRANCH_TEMPLATE, LONG_DEFAULT_COMMIT_TEMPLATE } from '$lib/ai/prompts';
 import { MessageRole, type PromptMessage, type AIClient, type Prompt } from '$lib/ai/types';
+import { andThen, buildFailureFromAny, ok, wrap, wrapAsync, type Result } from '$lib/result';
 import { isNonEmptyObject } from '$lib/utils/typeguards';
 import { fetch, Body, Response } from '@tauri-apps/api/http';
 
@@ -81,15 +82,22 @@ export class OllamaClient implements AIClient {
 		private modelName: string
 	) {}
 
-	async evaluate(prompt: Prompt) {
+	async evaluate(prompt: Prompt): Promise<Result<string, Error>> {
 		const messages = this.formatPrompt(prompt);
-		const response = await this.chat(messages);
-		const rawResponse = JSON.parse(response.message.content);
-		if (!isOllamaChatMessageFormat(rawResponse)) {
-			throw new Error('Invalid response: ' + response.message.content);
-		}
 
-		return rawResponse.result;
+		const responseResult = await this.chat(messages);
+
+		return andThen(responseResult, (response) => {
+			const rawResponseResult = wrap<unknown, Error>(() => JSON.parse(response.message.content));
+
+			return andThen(rawResponseResult, (rawResponse) => {
+				if (!isOllamaChatMessageFormat(rawResponse)) {
+					return buildFailureFromAny('Invalid response: ' + response.message.content);
+				}
+
+				return ok(rawResponse.result);
+			});
+		});
 	}
 
 	/**
@@ -124,31 +132,32 @@ ${JSON.stringify(OLLAMA_CHAT_MESSAGE_FORMAT_SCHEMA, null, 2)}`
 	 * @param request - The OllamaChatRequest object containing the request details.
 	 * @returns A Promise that resolves to the Response object.
 	 */
-	private async fetchChat(request: OllamaChatRequest): Promise<Response<any>> {
+	private async fetchChat(request: OllamaChatRequest): Promise<Result<Response<any>, Error>> {
 		const url = new URL(OllamaAPEndpoint.Chat, this.endpoint);
 		const body = Body.json(request);
-		const result = await fetch(url.toString(), {
-			method: 'POST',
-			headers: {
-				'Content-Type': 'application/json'
-			},
-			body
-		});
-		return result;
+		return await wrapAsync(
+			async () =>
+				await fetch(url.toString(), {
+					method: 'POST',
+					headers: {
+						'Content-Type': 'application/json'
+					},
+					body
+				})
+		);
 	}
 
 	/**
 	 * Sends a chat message to the LLM model and returns the response.
 	 *
 	 * @param messages - An array of LLMChatMessage objects representing the chat messages.
 	 * @param options - Optional LLMRequestOptions object for specifying additional options.
-	 * @throws Error if the response is invalid.
 	 * @returns A Promise that resolves to an LLMResponse object representing the response from the LLM model.
 	 */
 	private async chat(
 		messages: Prompt,
 		options?: OllamaRequestOptions
-	): Promise<OllamaChatResponse> {
+	): Promise<Result<OllamaChatResponse, Error>> {
 		const result = await this.fetchChat({
 			model: this.modelName,
 			stream: false,
@@ -157,10 +166,12 @@ ${JSON.stringify(OLLAMA_CHAT_MESSAGE_FORMAT_SCHEMA, null, 2)}`
 			format: 'json'
 		});
 
-		if (!isOllamaChatResponse(result.data)) {
-			throw new Error('Invalid response\n' + JSON.stringify(result.data));
-		}
+		return andThen(result, (result) => {
+			if (!isOllamaChatResponse(result.data)) {
+				return buildFailureFromAny('Invalid response\n' + JSON.stringify(result.data));
+			}
 
-		return result.data;
+			return ok(result.data);
+		});
 	}
 }

app/src/lib/ai/openAIClient.ts

@@ -1,6 +1,8 @@
 import { SHORT_DEFAULT_BRANCH_TEMPLATE, SHORT_DEFAULT_COMMIT_TEMPLATE } from '$lib/ai/prompts';
+import { andThen, buildFailureFromAny, ok, wrapAsync, type Result } from '$lib/result';
 import type { OpenAIModelName, Prompt, AIClient } from '$lib/ai/types';
 import type OpenAI from 'openai';
+import type { ChatCompletion } from 'openai/resources/index.mjs';
 
 export class OpenAIClient implements AIClient {
 	defaultCommitTemplate = SHORT_DEFAULT_COMMIT_TEMPLATE;
@@ -11,13 +13,21 @@ export class OpenAIClient implements AIClient {
 		private openAI: OpenAI
 	) {}
 
-	async evaluate(prompt: Prompt) {
-		const response = await this.openAI.chat.completions.create({
-			messages: prompt,
-			model: this.modelName,
-			max_tokens: 400
+	async evaluate(prompt: Prompt): Promise<Result<string, Error>> {
+		const responseResult = await wrapAsync<ChatCompletion, Error>(async () => {
+			return await this.openAI.chat.completions.create({
+				messages: prompt,
+				model: this.modelName,
+				max_tokens: 400
+			});
 		});
 
-		return response.choices[0].message.content || '';
+		return andThen(responseResult, (response) => {
+			if (response.choices[0]?.message.content) {
+				return ok(response.choices[0]?.message.content);
+			} else {
+				return buildFailureFromAny('Open AI generated an empty message');
+			}
+		});
 	}
 }