@apmantza/greedysearch-pi 1.6.2 → 1.6.3

package/extractors/google-ai.mjs

@@ -1,156 +0,0 @@
-#!/usr/bin/env node
-
-// extractors/google-ai.mjs
-// Navigate Google AI Mode (udm=50), wait for answer, return clean answer + sources.
-//
-// Usage:
-//   node extractors/google-ai.mjs "<query>" [--tab <prefix>]
-//
-// Output (stdout): JSON { answer, sources, query, url }
-// Errors go to stderr only — stdout is always clean JSON for piping.
-
-import {
-	cdp,
-	formatAnswer,
-	getOrOpenTab,
-	handleError,
-	outputJson,
-	parseArgs,
-	validateQuery,
-} from "./common.mjs";
-import { dismissConsent, handleVerification } from "./consent.mjs";
-import { SELECTORS } from "./selectors.mjs";
-
-const S = SELECTORS.google;
-
-const STREAM_POLL_INTERVAL = 600;
-const STREAM_STABLE_ROUNDS = 3;
-const STREAM_TIMEOUT = 45000;
-const MIN_ANSWER_LENGTH = 50;
-
-// ============================================================================
-// Google AI-specific helpers
-// ============================================================================
-
-async function waitForGoogleStreamComplete(tab) {
-	const deadline = Date.now() + STREAM_TIMEOUT;
-	let stableCount = 0;
-	let lastLen = -1;
-
-	while (Date.now() < deadline) {
-		await new Promise((r) => setTimeout(r, STREAM_POLL_INTERVAL));
-
-		const lenStr = await cdp([
-			"eval",
-			tab,
-			`(document.querySelector('${S.answerContainer}')?.innerText?.length || 0) + ''`,
-		]).catch(() => "0");
-
-		const len = parseInt(lenStr, 10) || 0;
-
-		if (len >= MIN_ANSWER_LENGTH && len === lastLen) {
-			stableCount++;
-			if (stableCount >= STREAM_STABLE_ROUNDS) return len;
-		} else {
-			stableCount = 0;
-			lastLen = len;
-		}
-	}
-
-	if (lastLen >= MIN_ANSWER_LENGTH) return lastLen;
-	throw new Error(
-		`Google AI answer did not stabilise within ${STREAM_TIMEOUT}ms`,
-	);
-}
-
-async function extractAnswer(tab) {
-	const excludeFilter = S.sourceExclude
-		.map((e) => `!a.href.includes('${e}')`)
-		.join(" && ");
-	const raw = await cdp([
-		"eval",
-		tab,
-		`
-    (function() {
-      var el = document.querySelector('${S.answerContainer}');
-      if (!el) return JSON.stringify({ answer: '', sources: [] });
-      var answer = el.innerText.trim();
-      var sources = Array.from(document.querySelectorAll('${S.sourceLink}'))
-        .filter(a => ${excludeFilter})
-        .map(a => ({ url: a.href.split('#')[0], title: (a.closest('${S.sourceHeadingParent}')?.querySelector('h3, [role=heading]')?.innerText || a.innerText?.trim().split('\\n')[0] || '').slice(0, 100) }))
-        .filter(s => s.url && s.url.length > 10)
-        .filter((v, i, arr) => arr.findIndex(x => x.url === v.url) === i)
-        .slice(0, 10);
-      return JSON.stringify({ answer, sources });
-    })()
-  `,
-	]);
-	return JSON.parse(raw);
-}
-
-// ============================================================================
-// Main
-// ============================================================================
-
-const USAGE =
-	'Usage: node extractors/google-ai.mjs "<query>" [--tab <prefix>]\n';
-
-async function main() {
-	const args = process.argv.slice(2);
-	validateQuery(args, USAGE);
-
-	const { query, tabPrefix, short } = parseArgs(args);
-
-	try {
-		await cdp(["list"]);
-		const tab = await getOrOpenTab(tabPrefix);
-
-		const url = `https://www.google.com/search?q=${encodeURIComponent(query)}&udm=50`;
-		await cdp(["nav", tab, url], 35000);
-		await new Promise((r) => setTimeout(r, 1500));
-		await dismissConsent(tab, cdp);
-
-		// If consent redirected us away, navigate back
-		const currentUrl = await cdp(["eval", tab, "document.location.href"]).catch(
-			() => "",
-		);
-		if (!currentUrl.includes("google.com/search")) {
-			await cdp(["nav", tab, url], 35000);
-			await new Promise((r) => setTimeout(r, 1500));
-		}
-
-		// Handle "verify you're human" — auto-click simple buttons, wait for user on hard CAPTCHA
-		const verifyResult = await handleVerification(tab, cdp, 60000);
-		if (verifyResult === "needs-human")
-			throw new Error(
-				"Google verification required — could not be completed automatically",
-			);
-		if (verifyResult === "clicked" || verifyResult === "cleared-by-user") {
-			// Re-navigate to the search URL after verification
-			await cdp(["nav", tab, url], 35000);
-			await new Promise((r) => setTimeout(r, 1500));
-		}
-
-		await waitForGoogleStreamComplete(tab);
-
-		const { answer, sources } = await extractAnswer(tab);
-		if (!answer)
-			throw new Error(
-				"No answer extracted — Google AI Mode may not have responded",
-			);
-
-		const finalUrl = await cdp(["eval", tab, "document.location.href"]).catch(
-			() => url,
-		);
-		outputJson({
-			query,
-			url: finalUrl,
-			answer: formatAnswer(answer, short),
-			sources,
-		});
-	} catch (e) {
-		handleError(e);
-	}
-}
-
-main();

package/extractors/perplexity.mjs

@@ -1,128 +0,0 @@
-#!/usr/bin/env node
-
-// extractors/perplexity.mjs
-// Navigate Perplexity, wait for streaming to complete, return clean answer + sources.
-//
-// Usage:
-//   node extractors/perplexity.mjs "<query>" [--tab <prefix>]
-//
-// Output (stdout): JSON { answer, sources, query, url }
-// Errors go to stderr only — stdout is always clean JSON for piping.
-//
-// TODO: Refactor - this file has 42 lines duplicated with google-ai.mjs (line 28)
-
-import {
-	cdp,
-	formatAnswer,
-	getOrOpenTab,
-	handleError,
-	injectClipboardInterceptor,
-	outputJson,
-	parseArgs,
-	parseSourcesFromMarkdown,
-	validateQuery,
-	waitForStreamComplete,
-} from "./common.mjs";
-import { dismissConsent } from "./consent.mjs";
-import { SELECTORS } from "./selectors.mjs";
-
-const S = SELECTORS.perplexity;
-const GLOBAL_VAR = "__pplxClipboard";
-
-// ============================================================================
-// Extraction
-// ============================================================================
-
-async function extractAnswer(tab) {
-	await cdp([
-		"eval",
-		tab,
-		`document.querySelector('${S.copyButton}')?.click()`,
-	]);
-	await new Promise((r) => setTimeout(r, 400));
-
-	const answer = await cdp(["eval", tab, `window.${GLOBAL_VAR} || ''`]);
-	if (!answer) throw new Error("Clipboard interceptor returned empty text");
-
-	const sources = parseSourcesFromMarkdown(answer);
-	return { answer: answer.trim(), sources };
-}
-
-// ============================================================================
-// Main
-// ============================================================================
-
-const USAGE =
-	'Usage: node extractors/perplexity.mjs "<query>" [--tab <prefix>]\n';
-
-async function main() {
-	const args = process.argv.slice(2);
-	validateQuery(args, USAGE);
-
-	const { query, tabPrefix, short } = parseArgs(args);
-
-	try {
-		// Refresh page list so cache is current
-		await cdp(["list"]);
-
-		const tab = await getOrOpenTab(tabPrefix);
-
-		// Navigate to homepage and use the search box (direct ?q= URLs trigger bot redirect)
-		await cdp(["nav", tab, "https://www.perplexity.ai/"], 35000);
-		await dismissConsent(tab, cdp);
-
-		// Wait for React app to mount input (up to 8s)
-		const deadline = Date.now() + 8000;
-		while (Date.now() < deadline) {
-			const found = await cdp([
-				"eval",
-				tab,
-				`!!document.querySelector('${S.input}')`,
-			]).catch(() => "false");
-			if (found === "true") break;
-			await new Promise((r) => setTimeout(r, 400));
-		}
-		await new Promise((r) => setTimeout(r, 300));
-
-		await injectClipboardInterceptor(tab, GLOBAL_VAR);
-		await cdp(["click", tab, S.input]);
-		await new Promise((r) => setTimeout(r, 400));
-		await cdp(["type", tab, query]);
-		await new Promise((r) => setTimeout(r, 400));
-
-		// Submit with Enter (most reliable across Chrome instances)
-		await cdp([
-			"eval",
-			tab,
-			`document.querySelector('${S.input}')?.dispatchEvent(new KeyboardEvent('keydown',{key:'Enter',bubbles:true,keyCode:13})), 'ok'`,
-		]);
-
-		await waitForStreamComplete(tab, {
-			timeout: 30000,
-			interval: 600,
-			stableRounds: 3,
-			selector: "document.body",
-		});
-
-		const { answer, sources } = await extractAnswer(tab);
-
-		if (!answer)
-			throw new Error(
-				"No answer extracted — Perplexity may not have responded",
-			);
-
-		const finalUrl = await cdp(["eval", tab, "document.location.href"]).catch(
-			() => "",
-		);
-		outputJson({
-			query,
-			url: finalUrl,
-			answer: formatAnswer(answer, short),
-			sources,
-		});
-	} catch (e) {
-		handleError(e);
-	}
-}
-
-main();

package/extractors/selectors.mjs

@@ -1,52 +0,0 @@
-// extractors/selectors.mjs
-// Centralized CSS selectors for all engines.
-// Update selectors here when a site changes its UI.
-
-export const SELECTORS = {
-	// ──────────────────────────────────────────────
-	// Perplexity (perplexity.ai)
-	// ──────────────────────────────────────────────
-	perplexity: {
-		input: "#ask-input",
-		copyButton: 'button[aria-label="Copy"]',
-		sourceItem: "[data-pplx-citation-url]",
-		sourceLink: "a",
-		consent: "#onetrust-accept-btn-handler",
-	},
-
-	// ──────────────────────────────────────────────
-	// Bing Copilot (copilot.microsoft.com)
-	// ──────────────────────────────────────────────
-	bing: {
-		input: "#userInput",
-		copyButton: 'button[data-testid="copy-ai-message-button"]',
-		sourceLink: 'a[href^="http"][target="_blank"]',
-		sourceExclude: "copilot.microsoft.com",
-		consent: "#onetrust-accept-btn-handler",
-	},
-
-	// ──────────────────────────────────────────────
-	// Google AI Mode (google.com/search?udm=50)
-	// ──────────────────────────────────────────────
-	google: {
-		answerContainer: ".pWvJNd",
-		sourceLink: 'a[href^="http"]',
-		sourceExclude: ["google.", "gstatic", "googleapis"],
-		sourceHeadingParent: "[data-snhf]",
-		consent: '#L2AGLb, button[jsname="b3VHJd"], .tHlp8d',
-	},
-
-	// ──────────────────────────────────────────────
-	// Gemini (gemini.google.com/app)
-	// ──────────────────────────────────────────────
-	gemini: {
-		input: "rich-textarea .ql-editor",
-		copyButton: 'button[aria-label="Copy"]',
-		sendButton: 'button[aria-label*="Send"]',
-		sourcesSidebarButton: "button.legacy-sources-sidebar-button",
-		sourcesExclude: ["gemini.google", "gstatic", "google.com/search"],
-		citationButtonPattern: 'button[aria-label*="citation from"]',
-		// For parsing citation aria-labels: "View source details for citation from {name}. Opens side panel."
-		citationNameRegex: /from\s+(.+?)\.\s/,
-	},
-};

package/launch.mjs

@@ -1,288 +0,0 @@
-#!/usr/bin/env node
-// launch.mjs — start a dedicated Chrome instance for GreedySearch
-//
-// This Chrome instance uses --disable-features=DevToolsPrivacyUI which suppresses
-// the "Allow remote debugging?" dialog entirely. It runs on port 9222 so it doesn't
-// conflict with your main Chrome session (which may use port 9223).
-//
-// search.mjs passes CDP_PROFILE_DIR so cdp.mjs targets this dedicated Chrome
-// without ever touching the user's main Chrome DevToolsActivePort file.
-//
-// Usage:
-//   node launch.mjs          — launch (or report if already running)
-//   node launch.mjs --kill   — stop and restore original DevToolsActivePort
-//   node launch.mjs --status — check if running
-
-import { execSync, spawn } from "node:child_process";
-import {
-	existsSync,
-	mkdirSync,
-	readFileSync,
-	unlinkSync,
-	writeFileSync,
-} from "node:fs";
-import http from "node:http";
-import { platform, tmpdir } from "node:os";
-import { join } from "node:path";
-
-const PORT = 9222;
-const PROFILE_DIR = join(tmpdir(), "greedysearch-chrome-profile");
-const ACTIVE_PORT = join(PROFILE_DIR, "DevToolsActivePort");
-const PID_FILE = join(tmpdir(), "greedysearch-chrome.pid");
-
-function findChrome() {
-	const os = platform();
-	const candidates =
-		os === "win32"
-			? [
-					"C:\\Program Files\\Google\\Chrome\\Application\\chrome.exe",
-					"C:\\Program Files (x86)\\Google\\Chrome\\Application\\chrome.exe",
-				]
-			: os === "darwin"
-				? [
-						"/Applications/Google Chrome.app/Contents/MacOS/Google Chrome",
-						"/Applications/Chromium.app/Contents/MacOS/Chromium",
-					]
-				: [
-						"/usr/bin/google-chrome",
-						"/usr/bin/google-chrome-stable",
-						"/usr/bin/chromium-browser",
-						"/usr/bin/chromium",
-						"/snap/bin/chromium",
-					];
-	return candidates.find(existsSync) || null;
-}
-
-const CHROME_FLAGS = [
-	`--remote-debugging-port=${PORT}`,
-	"--disable-features=DevToolsPrivacyUI", // suppresses "Allow remote debugging?" dialog
-	"--no-first-run",
-	"--no-default-browser-check",
-	"--disable-default-apps",
-	`--user-data-dir=${PROFILE_DIR}`,
-	"--profile-directory=Default",
-	"about:blank",
-];
-
-// ---------------------------------------------------------------------------
-
-function isRunning() {
-	if (!existsSync(PID_FILE)) return false;
-	const pid = parseInt(readFileSync(PID_FILE, "utf8").trim(), 10);
-	if (!pid) return false;
-	try {
-		process.kill(pid, 0);
-		return pid;
-	} catch {
-		return false;
-	}
-}
-
-// Get the PID of the process listening on a port (Windows + Unix)
-function getPortPid(port) {
-	try {
-		const os = platform();
-		if (os === "win32") {
-			// Windows: netstat -ano returns PID in last column
-			const out = execSync(`netstat -ano -p TCP 2>nul`, { encoding: "utf8" });
-			// Match lines like: TCP    127.0.0.1:9222    0.0.0.0:0    LISTENING    12345
-			const regex = new RegExp(
-				`TCP\\s+[^\\s]*:${port}\\s+[^\\s]*:0\\s+LISTENING\\s+(\\d+)`,
-				"i",
-			);
-			const match = out.match(regex);
-			return match ? parseInt(match[1], 10) : null;
-		} else {
-			// Unix: use lsof or ss
-			try {
-				const out = execSync(`lsof -i :${port} -t 2>/dev/null`, {
-					encoding: "utf8",
-				}).trim();
-				return out ? parseInt(out.split("\n")[0], 10) : null;
-			} catch {
-				const out = execSync(`ss -tlnp 2>/dev/null | grep :${port}`, {
-					encoding: "utf8",
-				});
-				const match = out.match(/pid=(\d+)/);
-				return match ? parseInt(match[1], 10) : null;
-			}
-		}
-	} catch {
-		return null;
-	}
-}
-
-// Kill a process by PID (with Windows/Unix compatibility)
-function killProcess(pid) {
-	try {
-		if (platform() === "win32") {
-			execSync(`taskkill //F //PID ${pid}`, { stdio: "ignore" });
-		} else {
-			process.kill(pid, "SIGTERM");
-		}
-		return true;
-	} catch {
-		return false;
-	}
-}
-
-// Clean up ghost Chrome on port 9222 that isn't tracked by our PID file
-function cleanupGhostChrome() {
-	const portPid = getPortPid(PORT);
-	if (!portPid) return; // Nothing on port 9222, all good
-
-	const trackedPid = isRunning();
-
-	if (trackedPid && portPid === trackedPid) {
-		return; // Port 9222 is our Chrome, all good
-	}
-
-	// Ghost Chrome detected — something on 9222 that isn't ours
-	if (trackedPid && portPid !== trackedPid) {
-		console.log(
-			`Ghost Chrome detected: port ${PORT} has pid ${portPid}, but our PID file says ${trackedPid}.`,
-		);
-	} else if (!trackedPid) {
-		console.log(
-			`Ghost Chrome detected: unknown process ${portPid} on port ${PORT} (no PID file).`,
-		);
-	}
-
-	console.log(`Killing ghost Chrome (pid ${portPid})...`);
-	killProcess(portPid);
-
-	// Clean up stale files
-	try {
-		unlinkSync(PID_FILE);
-	} catch {}
-	try {
-		unlinkSync(ACTIVE_PORT);
-	} catch {}
-	console.log("Cleaned up stale Chrome files.");
-}
-
-function httpGet(url, timeoutMs = 1000) {
-	return new Promise((resolve) => {
-		const req = http.get(url, (res) => {
-			let body = "";
-			res.on("data", (d) => (body += d));
-			res.on("end", () => resolve({ ok: res.statusCode === 200, body }));
-		});
-		req.on("error", () => resolve({ ok: false }));
-		req.setTimeout(timeoutMs, () => {
-			req.destroy();
-			resolve({ ok: false });
-		});
-	});
-}
-
-async function writePortFile(timeoutMs = 15000) {
-	// Chrome on Windows doesn't write DevToolsActivePort — we build it from the HTTP API.
-	const deadline = Date.now() + timeoutMs;
-	while (Date.now() < deadline) {
-		const { ok, body } = await httpGet(
-			`http://localhost:${PORT}/json/version`,
-			1500,
-		);
-		if (ok) {
-			try {
-				const { webSocketDebuggerUrl } = JSON.parse(body);
-				// webSocketDebuggerUrl = "ws://localhost:9223/devtools/browser/..."
-				const wsPath = new URL(webSocketDebuggerUrl).pathname;
-				// Write in DevToolsActivePort format: port on line 1, path on line 2
-				const content = `${PORT}\n${wsPath}`;
-				writeFileSync(ACTIVE_PORT, content, "utf8");
-				return true;
-			} catch {
-				/* malformed response, retry */
-			}
-		}
-		await new Promise((r) => setTimeout(r, 400));
-	}
-	return false;
-}
-
-// ---------------------------------------------------------------------------
-
-async function main() {
-	const arg = process.argv[2];
-
-	// Clean up any ghost Chrome on port 9222 before doing anything else
-	cleanupGhostChrome();
-
-	if (arg === "--kill") {
-		const pid = isRunning();
-		if (pid) {
-			const ok = killProcess(pid);
-			if (ok) console.log(`Stopped Chrome (pid ${pid}).`);
-			else console.error(`Failed to stop pid ${pid}.`);
-		} else {
-			console.log("GreedySearch Chrome is not running.");
-		}
-		return;
-	}
-
-	if (arg === "--status") {
-		const pid = isRunning();
-		if (pid)
-			console.log(
-				`Running — pid ${pid}, port ${PORT}, DevToolsActivePort redirected.`,
-			);
-		else console.log("Not running.");
-		return;
-	}
-
-	// Already running?
-	const existing = isRunning();
-	if (existing) {
-		const ready = await writePortFile(5000);
-		if (ready) {
-			console.log(
-				`GreedySearch Chrome already running (pid ${existing}, port ${PORT}).`,
-			);
-			console.log("Dedicated GreedySearch DevToolsActivePort is ready.");
-			return;
-		}
-		// Stale PID — process alive but not Chrome on port 9223. Fall through to fresh launch.
-		console.log(
-			`Stale PID ${existing} detected (not Chrome on port ${PORT}) — launching fresh.`,
-		);
-		try {
-			unlinkSync(PID_FILE);
-		} catch {}
-	}
-
-	const CHROME_EXE = process.env.CHROME_PATH || findChrome();
-	if (!CHROME_EXE) {
-		console.error("Chrome not found. Tried standard paths for your OS.");
-		console.error(
-			"Set the CHROME_PATH environment variable to point to your Chrome binary.",
-		);
-		process.exit(1);
-	}
-
-	mkdirSync(PROFILE_DIR, { recursive: true });
-
-	console.log(`Launching GreedySearch Chrome on port ${PORT}...`);
-	const proc = spawn(CHROME_EXE, CHROME_FLAGS, {
-		detached: true,
-		stdio: "ignore",
-		windowsHide: false,
-	});
-	proc.unref();
-	writeFileSync(PID_FILE, String(proc.pid));
-
-	// Wait for Chrome HTTP endpoint and build the dedicated DevToolsActivePort file
-	const portFileReady = await writePortFile();
-	if (!portFileReady) {
-		console.error("Chrome did not become ready within 15s.");
-		process.exit(1);
-	}
-
-	console.log(`Ready. No more "Allow remote debugging?" dialogs.`);
-	console.log(
-		"GreedySearch now uses its own isolated DevToolsActivePort file.",
-	);
-}
-
-main();