Skip to content

Add OpenAI-Compatible API Support #41

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
ninjaeon wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Add OpenAI-Compatible API Support #41

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
47 changes: 42 additions & 5 deletions preferences.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,6 +35,10 @@
{
"value": 3,
"label": "Mistral"
},
{
"value": 4,
"label": "OpenAI-Compatible"
}
],
"conditions": [
Expand Down Expand Up @@ -130,13 +134,46 @@
"operator": "AND"
},
{
"property": "extensions.tabgroups.mistral_model",
"label": "Mistral Model to Use For Renaming",
"default": "mistral-large-latest",
"type": "separator",
"label": "OpenAI-Compatible",
"id": "openai-separator",
"conditions": [
{ "if": { "property": "extensions.tabgroups.ai_model", "value": 4 } },
{ "if": { "property": "extensions.tabgroups.enable_sort", "value":true}}
],
"operator": "AND"
},
{
"property": "extensions.tabgroups.openai_endpoint",
"label": "OpenAI-Compatible Endpoint URL",
"default": "https://api.openai.com/v1/chat/completions",
"type": "string",
"placeholder": "Enter your Mistral Model here",
"placeholder": "Enter your OpenAI-Compatible endpoint URL here",
"conditions": [
{ "if": { "property": "extensions.tabgroups.ai_model", "value": 3 } },
{ "if": { "property": "extensions.tabgroups.ai_model", "value": 4 } },
{ "if": { "property": "extensions.tabgroups.enable_sort", "value":true}}
],
"operator": "AND"
},
{
"property": "extensions.tabgroups.openai_api_key",
"label": "OpenAI-Compatible API Key",
"type": "string",
"placeholder": "Enter your OpenAI-Compatible API key here",
"conditions": [
{ "if": { "property": "extensions.tabgroups.ai_model", "value": 4 } },
{ "if": { "property": "extensions.tabgroups.enable_sort", "value":true}}
],
"operator": "AND"
},
{
"property": "extensions.tabgroups.openai_model",
"label": "OpenAI-Compatible Model to Use For Renaming",
"default": "gpt-4o-mini",
"type": "string",
"placeholder": "Enter your OpenAI-Compatible model here",
"conditions": [
{ "if": { "property": "extensions.tabgroups.ai_model", "value": 4 } },
{ "if": { "property": "extensions.tabgroups.enable_sort", "value":true}}
],
"operator": "AND"
Expand Down
131 changes: 127 additions & 4 deletions tab_sort_clear.uc.js
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,19 +1,22 @@
// VERSION 4.11.0 (Added Mistral API support)
// VERSION 4.12.0 (Added OpenAI-Compatible API support)
(() => {
// --- Configuration ---

// Feature toggle preference keys
const ENABLE_SORT_PREF = "extensions.tabgroups.enable_sort";
const ENABLE_CLEAR_PREF = "extensions.tabgroups.enable_clear";
// Preference Key for AI Model Selection
const AI_MODEL_PREF = "extensions.tabgroups.ai_model"; // '1' for Gemini, '2' for Ollama, '3' for Mistral
const AI_MODEL_PREF = "extensions.tabgroups.ai_model"; // '1' for Gemini, '2' for Ollama, '3' for Mistral, '4' for OpenAI-Compatible
// Preference Keys for AI Config
const OLLAMA_ENDPOINT_PREF = "extensions.tabgroups.ollama_endpoint";
const OLLAMA_MODEL_PREF = "extensions.tabgroups.ollama_model";
const GEMINI_API_KEY_PREF = "extensions.tabgroups.gemini_api_key";
const GEMINI_MODEL_PREF = "extensions.tabgroups.gemini_model";
const MISTRAL_API_KEY_PREF = "extensions.tabgroups.mistral_api_key";
const MISTRAL_MODEL_PREF = "extensions.tabgroups.mistral_model";
const OPENAI_ENDPOINT_PREF = "extensions.tabgroups.openai_endpoint";
const OPENAI_API_KEY_PREF = "extensions.tabgroups.openai_api_key";
const OPENAI_MODEL_PREF = "extensions.tabgroups.openai_model";

// Helper function to read preferences with fallbacks
const getPref = (prefName, defaultValue = "") => {
Expand Down Expand Up @@ -45,6 +48,9 @@
const GEMINI_MODEL_VALUE = getPref(GEMINI_MODEL_PREF, "gemini-2.0-flash");
const MISTRAL_API_KEY_VALUE = getPref(MISTRAL_API_KEY_PREF, "");
const MISTRAL_MODEL_VALUE = getPref(MISTRAL_MODEL_PREF, "mistral-large-latest");
const OPENAI_ENDPOINT_VALUE = getPref(OPENAI_ENDPOINT_PREF, "https://api.openai.com/v1/chat/completions");
const OPENAI_API_KEY_VALUE = getPref(OPENAI_API_KEY_PREF, "");
const OPENAI_MODEL_VALUE = getPref(OPENAI_MODEL_PREF, "gpt-4o-mini");

const CONFIG = {
featureConfig: {
Expand Down Expand Up @@ -167,6 +173,47 @@
presence_penalty: 0.0
}
},
openai: {
enabled: AI_MODEL_VALUE == "4",
endpoint: OPENAI_ENDPOINT_VALUE,
apiKey: OPENAI_API_KEY_VALUE,
model: OPENAI_MODEL_VALUE,
promptTemplateBatch: `Analyze the following numbered list of tab data (Title, URL, Description) and assign a concise category (1-2 words, Title Case) for EACH tab.

Existing Categories (Use these EXACT names if a tab fits):
{EXISTING_CATEGORIES_LIST}

---
Instructions for Assignment:
1. **Prioritize Existing:** For each tab below, determine if it clearly belongs to one of the 'Existing Categories'. Base this primarily on the URL/Domain, then Title/Description. If it fits, you MUST use the EXACT category name provided in the 'Existing Categories' list. DO NOT create a minor variation (e.g., if 'Project Docs' exists, use that, don't create 'Project Documentation').
2. **Assign New Category (If Necessary):** Only if a tab DOES NOT fit an existing category, assign the best NEW concise category (1-2 words, Title Case).
* PRIORITIZE the URL/Domain (e.g., 'GitHub', 'YouTube', 'StackOverflow').
* Use Title/Description for specifics or generic domains.
3. **Consistency is CRITICAL:** Use the EXACT SAME category name for all tabs belonging to the same logical group (whether assigned an existing or a new category). If multiple tabs point to 'google.com/search?q=recipes', categorize them consistently (e.g., 'Google Search' or 'Recipes', but use the same one for all).
4. **Format:** 1-2 words, Title Case.

---
Input Tab Data:
{TAB_DATA_LIST}

---
Instructions for Output:
1. Output ONLY the category names.
2. Provide EXACTLY ONE category name per line.
3. The number of lines in your output MUST EXACTLY MATCH the number of tabs in the Input Tab Data list above.
4. DO NOT include numbering, explanations, apologies, markdown formatting, or any surrounding text like "Output:" or backticks.
5. Just the list of categories, separated by newlines.
---

Output:`,
generationConfig: {
temperature: 0.1, // Low temp for consistency
max_tokens: 512, // Default, will be calculated dynamically
top_p: 0.9,
frequency_penalty: 0.0,
presence_penalty: 0.0
}
},
customApi: {
enabled: false,
// ... (custom API config if needed)
Expand Down Expand Up @@ -678,7 +725,7 @@
return [];
}

const { gemini, ollama, mistral } = CONFIG.apiConfig;
const { gemini, ollama, mistral, openai } = CONFIG.apiConfig;
let result = [];
let apiChoice = "None";

Expand Down Expand Up @@ -928,8 +975,84 @@
console.log("Batch AI (Mistral): Processed Topics:", processedTopics);
result = validTabs.map((tab, index) => ({ tab: tab, topic: processedTopics[index] }));
}
} else if (openai.enabled) {
// --- OPENAI-COMPATIBLE LOGIC ---
apiChoice = "OpenAI-Compatible";
console.log(`Batch AI (OpenAI-Compatible): Requesting categories for ${validTabs.length} tabs, considering ${existingCategoryNames.length} existing categories...`);
let apiUrl = openai.endpoint;
let headers = {
'Content-Type': 'application/json',
'Authorization': `Bearer ${openai.apiKey}`
};

const tabDataArray = validTabs.map(getTabData);
const formattedTabDataList = tabDataArray.map((data, index) =>
`${index + 1}.\nTitle: "${data.title}"\nURL: "${data.url}"\nDescription: "${data.description}"`
).join('\n\n');
const formattedExistingCategories = existingCategoryNames.length > 0
? existingCategoryNames.map(name => `- ${name}`).join('\n')
: "None";

const prompt = openai.promptTemplateBatch
.replace("{EXISTING_CATEGORIES_LIST}", formattedExistingCategories)
.replace("{TAB_DATA_LIST}", formattedTabDataList);

const estimatedOutputTokens = Math.max(256, validTabs.length * 16); // Dynamic estimation
const requestBody = {
model: openai.model,
messages: [{ role: "user", content: prompt }],
max_tokens: estimatedOutputTokens,
temperature: openai.generationConfig.temperature,
top_p: openai.generationConfig.top_p,
frequency_penalty: openai.generationConfig.frequency_penalty,
presence_penalty: openai.generationConfig.presence_penalty
};

const response = await fetch(apiUrl, {
method: 'POST',
headers: headers,
body: JSON.stringify(requestBody)
});

if (!response.ok) {
const errorText = await response.text().catch(() => 'Unknown API error reason');
throw new Error(`OpenAI-Compatible API Error ${response.status}: ${errorText}`);
}

const data = await response.json();
let aiText = data.choices?.[0]?.message?.content?.trim();

if (!aiText) {
throw new Error("OpenAI-Compatible: Empty API response");
}

const lines = aiText.split('\n').map(line => line.trim()).filter(Boolean);

if (lines.length !== validTabs.length) {
console.warn(`Batch AI (OpenAI-Compatible): Mismatch! Expected ${validTabs.length} topics, received ${lines.length}. AI Response:\n${aiText}`);
if (validTabs.length === 1 && lines.length > 0) {
const firstLineTopic = processTopic(lines[0]);
console.warn(` -> Mismatch Correction (Single Tab): Using first line "${lines[0]}" -> Topic: "${firstLineTopic}"`);
result = [{ tab: validTabs[0], topic: firstLineTopic }];
} else if (lines.length > validTabs.length) {
console.warn(` -> Mismatch Correction (Too Many Lines): Truncating response to ${validTabs.length} lines.`);
const processedTopics = lines.slice(0, validTabs.length).map(processTopic);
result = validTabs.map((tab, index) => ({ tab: tab, topic: processedTopics[index] }));
} else {
console.warn(` -> Fallback (Too Few Lines): Assigning remaining tabs "Uncategorized".`);
const processedTopics = lines.map(processTopic);
result = validTabs.map((tab, index) => ({
tab: tab,
topic: index < processedTopics.length ? processedTopics[index] : "Uncategorized"
}));
}
} else {
const processedTopics = lines.map(processTopic);
console.log("Batch AI (OpenAI-Compatible): Processed Topics:", processedTopics);
result = validTabs.map((tab, index) => ({ tab: tab, topic: processedTopics[index] }));
}
} else {
throw new Error("No AI API is enabled in the configuration (Gemini, Ollama, or Mistral).");
throw new Error("No AI API is enabled in the configuration (Gemini, Ollama, Mistral, or OpenAI-Compatible).");
}
return result;
} catch (error) {
Expand Down