browser-extension/background.js

// WidgeTDC Harvester Background Script

let API_BASE = 'http://localhost:3001';

// Load config on startup
chrome.storage.sync.get({ widgetdc_api_url: 'http://localhost:3001' }, (items) => {
    API_BASE = items.widgetdc_api_url;
});

// Listen for config changes
chrome.storage.onChanged.addListener((changes, namespace) => {
    if (namespace === 'sync' && changes.widgetdc_api_url) {
        API_BASE = changes.widgetdc_api_url.newValue;
    }
});

// Listen for messages from popup or content script
chrome.runtime.onMessage.addListener((request, sender, sendResponse) => {
    if (request.action === 'ingest_page') {
        handleIngestion(request.data).then(sendResponse);
        return true; // Keep channel open for async response
    }
    if (request.action === 'crawl_site') {
        handleCrawl(request.data).then(sendResponse);
        return true;
    }
    if (request.action === 'check_connection') {
        checkBackendHealth().then(sendResponse);
        return true;
    }
});

async function checkBackendHealth() {
    try {
        const response = await fetch(`${API_BASE}/health`);
        return { success: response.ok };
    } catch (error) {
        return { success: false, error: error.message };
    }
}

async function handleIngestion(data) {
    const { url, title, html } = data;
    console.log('Ingesting page:', url);

    try {
        // Option 1: Send raw HTML (heavy)
        // Option 2: Send URL and let backend crawl (cleaner) -> We use this via ingestion.crawl

        const response = await fetch(`${API_BASE}/api/mcp/route`, {
            method: 'POST',
            headers: { 'Content-Type': 'application/json' },
            body: JSON.stringify({
                tool: 'ingestion.crawl',
                params: {
                    url: url,
                    depth: 0 // Single page
                }
            })
        });

        const result = await response.json();
        return { success: true, data: result };

    } catch (error) {
        console.error('Ingestion failed:', error);
        return { success: false, error: error.message };
    }
}

async function handleCrawl(data) {
    const { url, depth } = data;
    console.log(`Crawling site: ${url} (Depth: ${depth})`);

    try {
        const response = await fetch(`${API_BASE}/api/mcp/route`, {
            method: 'POST',
            headers: { 'Content-Type': 'application/json' },
            body: JSON.stringify({
                tool: 'ingestion.crawl',
                params: {
                    url: url,
                    depth: depth || 1,
                    mode: data.mode || 'single',
                    keywords: data.keywords || [],
                    user: data.user
                }
            })
        });

        const result = await response.json();
        return { success: true, data: result };

    } catch (error) {
        console.error('Crawl failed:', error);
        return { success: false, error: error.message };
    }
}

// Context menu for quick actions
chrome.runtime.onInstalled.addListener(() => {
    chrome.contextMenus.create({
        id: "widgetdc-ingest",
        title: "Ingest to WidgeTDC",
        contexts: ["page", "selection", "link"]
    });
});

chrome.contextMenus.onClicked.addListener((info, tab) => {
    if (info.menuItemId === "widgetdc-ingest") {
        const targetUrl = info.linkUrl || info.pageUrl;
        handleIngestion({ url: targetUrl });
    }
});