Feat: Add Configurable Timeout, internal Project ID fields, and patch model testing payload

TODO.md

@@ -1 +1,19 @@
-# TODO
+# MoodleGPT TODO List
+
+Based on the open GitHub issues, the following features and improvements are planned:
+
+## Configuration & Settings
+
+- [x] **Configurable Request Timeout** (#70): Add a setting to manually configure the API request timeout, preventing hangups on deeply nested or slow LLM requests.
+- [x] **OpenAI Project ID Support** (#55): Add a field under Advanced Settings to provide an OpenAI `project_id`, enabling compatibility with organizational level API keys.
+- [ ] **Toggleable Visual Indications** (#71): Add an option to completely disable the cursor hijack and loading animations to prevent unwanted visual behaviors during strictly proctored quizzes.
+- [ ] **Custom LMS Hostname Whitelisting** (#71): Implement internal configuration logic that seamlessly whitelists custom university domains (e.g., `*.oes.kz`) to ensure extension logic runs on rebranded Moodle portals.
+
+## Context & Advanced AI Integration
+
+- [ ] **Document Context Injection** (#39, #42): Allow users to upload or attach PDFs/lecture notes to be embedded in the AI's prompt resolution context to strictly enforce the correct syllabus answers.
+- [ ] **ChatGPT Assistant Integrations** (#37): Implement support for routing requests to customized pre-trained ChatGPT Assistants using specific Assistant IDs instead of relying entirely on standard ChatCompletion routes.
+
+## Cross-Platform
+
+- [ ] **Firefox Port** (#50): Migrate and package the Chrome extension codebase for native Firefox compatibility, adapting to standard `browser.*` APIs.

extension/popup/index.html

@@ -60,10 +60,18 @@
           <label for="baseURL" class="textLabel">Base URL:</label>
           <input id="baseURL" type="text" />
         </div>
+        <div class="line center">
+          <label for="projectId" class="textLabel">Project ID:</label>
+          <input id="projectId" type="text" placeholder="proj_..." />
+        </div>
         <div class="line center">
           <label for="maxTokens" class="textLabel">Max Tokens:</label>
           <input id="maxTokens" type="number" />
         </div>
+        <div class="line center">
+          <label for="timeoutValue" class="textLabel">Timeout (s):</label>
+          <input id="timeoutValue" type="number" placeholder="20" />
+        </div>
       </div>
 
       <!-- SWITCH SETTINGS MODE -->

issues/issues.md

@@ -0,0 +1,78 @@
+[HELP] Add host and desactivate the cursor indication #71
+Help Request
+
+Issue Summary
+Host name change due to private proctoring system integration and unwanted visual behavior in tests.
+
+Describe the problem
+Hello!
+Our university is integrating a private proctoring system, and because of this, the host name is changing — it now ends with .oes.kz.
+How can we properly add or register this new host in the LMS so that everything continues to work correctly?
+
+Additionally, we’d like to remove the visual behavior when hovering over a test question. Currently, when you hover, the cursor changes, and when you click, a loading animation appears. Is there a way to disable this effect?
+
+Environment
+• Operating System: Windows 10
+• Browser: Google Chrome
+• Version: [e.g. 130.0.6723.70]
+• Platform: LMS
+
+[FEATURE] Add "Request timeout" configuration option #70
+Help Request
+
+Issue Summary
+
+[Request was abort.]
+
+Describe the problem
+
+[I have no idea what happen maybe request is take too long time.it happen after using gpt 5 mode which not heppen on gpt4 before. can we have timeout setting feature so we can manually set it]
+
+Environment
+
+    Operating System: [W11]
+    Browser: [Chrome]
+    Version: [1.1.5]
+    Any other relevant information
+
+Additional Information
+
+[Any additional information that may be helpful in diagnosing the issue.]
+
+[FEATURE] Add Project ID under Advanced Settings #55
+Add Project ID under Advanced Settings
+
+To use organization API keys it is required to provide the project id in the request. I never used openai APIs before and I was getting 401 - Incorrect API key provided.
+
+I fixed changing this line in MoodleGPT.js :
+
+project: s = Be("OPENAI_PROJECT_ID") ?? <my-project-id>
+
+and then loading the extension manually.
+
+It would be very nice to add an optional project id under the advanced settings section.
+
+[FEATURE] Firefox port? #50
+Feature Request
+
+Description of the Feature
+Would be really good if the extension was available for Firefox too.
+
+Additional Information
+It wouldn't require too much code modification, since Firefox supports chrome. APIs natively too.
+
+[FEATURE] Preprompt for test with existing documents #42
+Hello, could there be an option to maybe choose a chatgpt chat you have made already or do I need to make a separate API and enter the files in there maybe?
+
+TLDR: I have specific exams for what correct answers only come from there. For this I would usually make a chatgpt chat that had all the documents pre uploaded in there. But if I want to use the plugin, it takes new chats. Any way to work this in or do I need to generate a specific API for this?
+
+[FEATURE] Add a document with the extension #39
+Hello, quick question: is it possible to provide my PDF of lecture notes in advance, to prepare and optimise ChatGPT’s answers to the multiple-choice questions?
+For example, for my ‘Strategic Communication’ exam, I provide my PDF of about 30 pages of notes, so that the answers are faithful to what we covered in class.
+Thanks and have a good day
+Leo
+
+[FEATURE] ChatGPT Assistants #37
+Feature Request
+
+It would be really helpful if ChatGPT's assistants would also be supported in the extension. It should be only an option to select and use assistants which you have made for a specific topic to give more exact results.

src/background/core/get-response.ts

@@ -18,7 +18,7 @@ async function getChatGPTResponse(
   question: string
 ): Promise<GPTAnswer> {
   const controller = new AbortController();
-  const timeoutControler = setTimeout(() => controller.abort(), 20 * 1000);
+  const timeoutControler = setTimeout(() => controller.abort(), (config.timeoutValue || 20) * 1000);
 
   // Get the content to send to chatgpt
   // Including the instructions to the AI, the images as base64 if needed, the question and the past conversation if history is set to true
@@ -27,6 +27,7 @@ async function getChatGPTResponse(
   const client = new OpenAI({
     apiKey: config.apiKey,
     baseURL: config.baseURL,
+    project: config.projectId,
     dangerouslyAllowBrowser: true
   });
 

src/background/types/config.ts

@@ -13,7 +13,9 @@ type Config = {
   includeImages?: boolean;
   mode?: 'autocomplete' | 'question-to-answer' | 'clipboard';
   baseURL?: string;
+  projectId?: string;
   maxTokens?: number;
+  timeoutValue?: number;
 };
 
 export default Config;

src/popup/gpt-version.ts

@@ -6,6 +6,8 @@ const inputModel: HTMLInputElement = document.querySelector('#model')!;
 const modelsList: HTMLElement = document.querySelector('#models')!;
 const imagesIntegrationLine: HTMLInputElement = document.querySelector('#includeImages-line')!;
 const baseURLSelector: HTMLInputElement = document.querySelector('#baseURL')!;
+const projectIdSelector: HTMLInputElement = document.querySelector('#projectId')!;
+const maxTokensSelector: HTMLInputElement = document.querySelector('#maxTokens')!;
 /**
  * Check if the gpt version is at least 4 to show the option 'Include images'
  */
@@ -24,6 +26,7 @@ inputModel.addEventListener('input', checkCanIncludeImages);
 export async function populateDatalistWithGptVersions() {
   const apiKey = apiKeySelector.value?.trim();
   const baseURL = baseURLSelector.value?.trim();
+  const projectId = projectIdSelector.value?.trim();
 
   if (!apiKey) return;
 
@@ -33,6 +36,7 @@ export async function populateDatalistWithGptVersions() {
     const client = new OpenAI({
       apiKey,
       baseURL,
+      project: projectId,
       dangerouslyAllowBrowser: true
     });
 
@@ -66,13 +70,21 @@ export async function checkModel() {
   const model = inputModel.value?.trim();
   const apiKey = apiKeySelector.value?.trim();
   const baseURL = baseURLSelector.value?.trim();
+  const projectId = projectIdSelector.value?.trim();
+  const maxTokens = maxTokensSelector.value ? parseInt(maxTokensSelector.value) : undefined;
 
   try {
     showMessage({ msg: 'Checking GPT version...', isInfinite: true, isError: false });
-    const client = new OpenAI({ apiKey, baseURL, dangerouslyAllowBrowser: true });
+    const client = new OpenAI({
+      apiKey,
+      baseURL,
+      project: projectId,
+      dangerouslyAllowBrowser: true
+    });
     await client.chat.completions.create({
       model,
-      messages: [{ role: 'user', content: 'reply just pong' }]
+      messages: [{ role: 'user', content: 'reply just pong' }],
+      max_completion_tokens: maxTokens || 2000
     });
     showMessage({ msg: 'The model is valid!' });
   } catch (err: any) {

src/popup/index.ts

@@ -9,12 +9,12 @@ import { showMessage } from './utils';
 const saveBtn = document.querySelector('.save')!;
 
 // inputs id
-const inputsText = ['apiKey', 'code', 'model', 'baseURL', 'maxTokens'];
+const inputsText = ['apiKey', 'code', 'model', 'baseURL', 'maxTokens', 'projectId', 'timeoutValue'];
 
 // Save the configuration
 saveBtn.addEventListener('click', function () {
-  const [apiKey, code, model, baseURL, maxTokens] = inputsText.map(selector =>
+  const [apiKey, code, model, baseURL, maxTokens, projectId, timeoutValue] = inputsText.map(
-    (document.querySelector('#' + selector) as HTMLInputElement).value.trim()
+    selector => (document.querySelector('#' + selector) as HTMLInputElement).value.trim()
   );
   const [logs, title, cursor, typing, mouseover, infinite, timeout, history, includeImages] =
     inputsCheckbox.map(selector => {
@@ -42,6 +42,8 @@ saveBtn.addEventListener('click', function () {
       model,
       baseURL,
       maxTokens: maxTokens ? parseInt(maxTokens) : undefined,
+      projectId,
+      timeoutValue: timeoutValue ? parseInt(timeoutValue) : undefined,
       logs,
       title,
       cursor,