@asterxsk/kiln 0.2.0 → 0.3.0

package/agent.bak.20260905-190020/extensions/pi-web-access/gemini-web-config.ts

@@ -0,0 +1,101 @@
+import { existsSync, readFileSync } from "node:fs";
+import { getWebSearchConfigPath } from "./utils.ts";
+
+const CONFIG_PATH = getWebSearchConfigPath();
+
+interface GeminiWebConfig {
+	browserCookies?: BrowserCookieSelection;
+	allowBrowserCookies?: boolean;
+}
+
+const BROWSER_COOKIE_PRESETS = ["helium", "chrome", "brave", "arc", "chromium", "edge"] as const;
+export type BrowserCookiePreset = typeof BROWSER_COOKIE_PRESETS[number];
+
+export interface BrowserCookieSelection {
+	browser?: BrowserCookiePreset;
+	profile?: string;
+}
+
+let cachedConfig: GeminiWebConfig | null = null;
+
+export function normalizeChromeProfile(value: unknown): string | undefined {
+	if (typeof value !== "string") return undefined;
+	const normalized = value.trim();
+	return normalized.length > 0 ? normalized : undefined;
+}
+
+function parseBrowserCookies(value: unknown): BrowserCookieSelection | undefined {
+	if (value === undefined) return undefined;
+	if (!value || typeof value !== "object" || Array.isArray(value)) {
+		throw new Error(`browserCookies in ${CONFIG_PATH} must be an object`);
+	}
+	const raw = value as Record<string, unknown>;
+	if ("profilePath" in raw) {
+		throw new Error(`browserCookies.profilePath is not supported; use a browser preset and profile directory name in ${CONFIG_PATH}`);
+	}
+	let browser: BrowserCookiePreset | undefined;
+	if (raw.browser !== undefined) {
+		if (typeof raw.browser !== "string") throw new Error(`browserCookies.browser in ${CONFIG_PATH} must be a supported browser name`);
+		const normalized = raw.browser.trim().toLowerCase();
+		const preset = BROWSER_COOKIE_PRESETS.find((candidate) => candidate === normalized);
+		if (!preset) {
+			throw new Error(`Unsupported browserCookies.browser ${JSON.stringify(raw.browser)} in ${CONFIG_PATH}`);
+		}
+		browser = preset;
+	}
+	if (raw.profile !== undefined && typeof raw.profile !== "string") {
+		throw new Error(`browserCookies.profile in ${CONFIG_PATH} must be a profile directory name`);
+	}
+	const profile = normalizeChromeProfile(raw.profile);
+	if (profile && (profile === "." || profile === ".." || profile.includes("/") || profile.includes("\\"))) {
+		throw new Error(`browserCookies.profile in ${CONFIG_PATH} must be a profile directory name, not a path`);
+	}
+	return { ...(browser ? { browser } : {}), ...(profile ? { profile } : {}) };
+}
+
+function loadConfig(): GeminiWebConfig {
+	if (cachedConfig) return cachedConfig;
+	if (!existsSync(CONFIG_PATH)) {
+		cachedConfig = {};
+		return cachedConfig;
+	}
+
+	const rawText = readFileSync(CONFIG_PATH, "utf-8");
+	let raw: { chromeProfile?: unknown; browserCookies?: unknown; allowBrowserCookies?: unknown };
+	try {
+		raw = JSON.parse(rawText) as { chromeProfile?: unknown; browserCookies?: unknown; allowBrowserCookies?: unknown };
+	} catch (err) {
+		const message = err instanceof Error ? err.message : String(err);
+		throw new Error(`Failed to parse ${CONFIG_PATH}: ${message}`);
+	}
+
+	if (raw.chromeProfile !== undefined) {
+		throw new Error(`chromeProfile in ${CONFIG_PATH} is no longer supported; use browserCookies.profile`);
+	}
+	cachedConfig = {
+		browserCookies: parseBrowserCookies(raw.browserCookies),
+		allowBrowserCookies: raw.allowBrowserCookies === true,
+	};
+	return cachedConfig;
+}
+
+export function getBrowserCookieSelectionFromConfig(): BrowserCookieSelection {
+	return { ...loadConfig().browserCookies };
+}
+
+export function isBrowserCookieAccessAllowed(): boolean {
+	if (process.env.PI_ALLOW_BROWSER_COOKIES === "1" || process.env.FEYNMAN_ALLOW_BROWSER_COOKIES === "1") {
+		return true;
+	}
+	if (!existsSync(CONFIG_PATH)) return false;
+
+	const rawText = readFileSync(CONFIG_PATH, "utf-8");
+	let raw: { allowBrowserCookies?: unknown };
+	try {
+		raw = JSON.parse(rawText) as { allowBrowserCookies?: unknown };
+	} catch (err) {
+		const message = err instanceof Error ? err.message : String(err);
+		throw new Error(`Failed to parse ${CONFIG_PATH}: ${message}`);
+	}
+	return raw.allowBrowserCookies === true;
+}

package/agent.bak.20260905-190020/extensions/pi-web-access/gemini-web.ts

@@ -0,0 +1,487 @@
+import { readFileSync } from "node:fs";
+import { basename } from "node:path";
+import { getLastGoogleCookieDiagnostic, getLastGoogleCookieDiagnosticDetails, type BrowserCookieDiagnosticDetails, type CookieMap, getGoogleCookies } from "./chrome-cookies.ts";
+import { getBrowserCookieSelectionFromConfig, isBrowserCookieAccessAllowed, normalizeChromeProfile } from "./gemini-web-config.ts";
+
+const GEMINI_APP_URL = "https://gemini.google.com/app";
+const GEMINI_STREAM_GENERATE_URL =
+	"https://gemini.google.com/_/BardChatUi/data/assistant.lamda.BardFrontendService/StreamGenerate";
+const GEMINI_UPLOAD_URL = "https://content-push.googleapis.com/upload";
+const GEMINI_UPLOAD_PUSH_ID = "feeds/mcudyrk2a4khkz";
+const GOOGLE_LIST_ACCOUNTS_URL =
+	"https://accounts.google.com/ListAccounts?gpsia=1&source=ChromiumBrowser&laf=b64bin&json=standard";
+
+const USER_AGENT =
+	"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36";
+
+const MODEL_HEADER_NAME = "x-goog-ext-525001261-jspb";
+export const DEFAULT_GEMINI_WEB_MODEL = "gemini-3.1-pro";
+const MODEL_HEADERS: Record<string, string> = {
+	[DEFAULT_GEMINI_WEB_MODEL]: '[1,null,null,null,"9d8ca3786ebdfbea",null,null,0,[4]]',
+	"gemini-2.5-pro": '[1,null,null,null,"4af6c7f5da75d65d",null,null,0,[4]]',
+	"gemini-2.5-flash": '[1,null,null,null,"9ec249fc9ad08861",null,null,0,[4]]',
+};
+
+const REQUIRED_COOKIES = ["__Secure-1PSID", "__Secure-1PSIDTS"];
+
+// Gemini Web talks to Google's frontend inside the host agent (pi), whose
+// global undici dispatcher runs HTTP/1.1 with undici's default 16 KiB
+// maxHeaderSize. Google's gemini.google.com /app response headers routinely
+// exceed that budget over HTTP/1.1, so cookie-bearing requests fail with
+// UND_ERR_HEADERS_OVERFLOW surfaced as "fetch failed". Route these requests
+// through a dedicated undici Agent with a larger header budget instead; fall
+// back to the ambient global fetch when undici is not resolvable (e.g.
+// standalone test environments).
+let geminiFetchImpl: typeof fetch | null = null;
+
+// Test-only hook: lets tests inject a simulated transport instead of patching
+// globalThis.fetch. Gemini Web deliberately bypasses the ambient global fetch
+// (inside the pi host agent it is constrained to undici's 16 KiB
+// maxHeaderSize and cannot serve Google's /app page over HTTP/1.1), so
+// globalThis.fetch patching is no longer a reliable way to intercept these
+// requests.
+export function setGeminiFetchOverrideForTests(fetchImpl: typeof fetch | null): void {
+	geminiFetchOverride = fetchImpl;
+}
+
+let geminiFetchOverride: typeof fetch | null = null;
+
+// The header budget Gemini Web needs for Google frontend responses.
+// Google's /app page exceeds undici's default 16 KiB maxHeaderSize when
+// served over HTTP/1.1, which is how pi's global dispatcher operates.
+export const GEMINI_MAX_HEADER_SIZE = 4 * 1024 * 1024;
+
+// Build a fetch that routes through a dedicated undici Agent with a larger
+// header budget than the host agent's global dispatcher allows. Exported for
+// tests so the agent configuration can be asserted without network access.
+export function createGeminiFetch(undiciImpl: typeof import("undici")): typeof fetch {
+	const agent = new undiciImpl.EnvHttpProxyAgent({
+		allowH2: false,
+		connectTimeout: 30_000,
+		maxHeaderSize: GEMINI_MAX_HEADER_SIZE,
+		pipelining: 1,
+	});
+	// undici 8 types its fetch request/response structurally differently from
+	// the DOM lib types (duplex/textStream on Request, Symbol.dispose on
+	// Headers); the shape this module uses (status, ok, headers.get, text) is
+	// identical, so bridge the two.
+	type UndiciFetch = typeof undiciImpl.fetch;
+	return (input, init) =>
+		undiciImpl.fetch(
+			input as unknown as Parameters<UndiciFetch>[0],
+			{ ...init, dispatcher: agent } as unknown as Parameters<UndiciFetch>[1],
+		) as unknown as Promise<Response>;
+}
+
+export async function resolveGeminiFetch(): Promise<typeof fetch> {
+	if (geminiFetchOverride) return geminiFetchOverride;
+	if (geminiFetchImpl) return geminiFetchImpl;
+
+	let undici: typeof import("undici");
+	try {
+		undici = await import("undici");
+	} catch {
+		geminiFetchImpl = fetch;
+		return geminiFetchImpl;
+	}
+
+	geminiFetchImpl = createGeminiFetch(undici);
+	return geminiFetchImpl;
+}
+
+export interface GeminiWebOptions {
+	youtubeUrl?: string;
+	model?: string;
+	files?: string[];
+	signal?: AbortSignal;
+	timeoutMs?: number;
+}
+
+export async function isGeminiWebAvailable(chromeProfile?: string): Promise<CookieMap | null> {
+	if (!isBrowserCookieAccessAllowed()) return null;
+
+	const configured = getBrowserCookieSelectionFromConfig();
+	const result = await getGoogleCookies({
+		browser: configured.browser,
+		profile: normalizeChromeProfile(chromeProfile) ?? configured.profile,
+		requiredCookies: REQUIRED_COOKIES,
+	});
+	if (!result) return null;
+	return result.cookies;
+}
+
+export function getGeminiWebAvailabilityDiagnostic(): string | null {
+	return isBrowserCookieAccessAllowed() ? getLastGoogleCookieDiagnostic() : null;
+}
+
+export function getGeminiWebAvailabilityDiagnosticDetails(): BrowserCookieDiagnosticDetails | null {
+	return isBrowserCookieAccessAllowed() ? getLastGoogleCookieDiagnosticDetails() : null;
+}
+
+export async function getActiveGoogleEmail(cookies: CookieMap): Promise<string | null> {
+	const cookieHeader = buildCookieHeader(cookies);
+	if (!cookieHeader) return null;
+
+	try {
+		const html = await fetchWithCookieRedirects(
+			GEMINI_APP_URL,
+			cookieHeader,
+			10,
+			AbortSignal.timeout(10000),
+		);
+		const email = extractEmailFromGeminiHtml(html);
+		if (email) return email;
+	} catch {
+	}
+
+	try {
+		const response = await fetchWithCookieRedirects(
+			GOOGLE_LIST_ACCOUNTS_URL,
+			cookieHeader,
+			10,
+			AbortSignal.timeout(10000),
+		);
+		return extractEmailFromListAccounts(response);
+	} catch {
+		return null;
+	}
+}
+
+export async function queryWithCookies(
+	prompt: string,
+	cookieMap: CookieMap,
+	options: GeminiWebOptions = {},
+): Promise<string> {
+	const model = options.model ?? DEFAULT_GEMINI_WEB_MODEL;
+	if (!MODEL_HEADERS[model]) {
+		throw new Error(`Gemini Web does not support model ${model}; configure Gemini API or choose a supported Gemini Web model.`);
+	}
+	const timeoutMs = options.timeoutMs ?? 120000;
+
+	let fullPrompt = prompt;
+	if (options.youtubeUrl) {
+		fullPrompt = `${fullPrompt}\n\nYouTube video: ${options.youtubeUrl}`;
+	}
+
+	const result = await runGeminiWebOnce(fullPrompt, cookieMap, model, options.files, timeoutMs, options.signal);
+
+	if (result.errorMessage) throw new Error(result.errorMessage);
+	if (!result.text) throw new Error("Gemini Web returned empty response");
+	return result.text;
+}
+
+interface GeminiWebResult {
+	text: string;
+	errorCode?: number;
+	errorMessage?: string;
+}
+
+async function runGeminiWebOnce(
+	prompt: string,
+	cookieMap: CookieMap,
+	model: string,
+	files: string[] | undefined,
+	timeoutMs: number,
+	signal?: AbortSignal,
+): Promise<GeminiWebResult> {
+	const effectiveSignal = withTimeout(signal, timeoutMs);
+	const cookieHeader = buildCookieHeader(cookieMap);
+	const accessToken = await fetchAccessToken(cookieHeader, effectiveSignal);
+
+	const uploaded: Array<{ id: string; name: string }> = [];
+	if (files) {
+		for (const filePath of files) {
+			uploaded.push(await uploadFile(filePath, cookieHeader, effectiveSignal));
+		}
+	}
+
+	const fReq = buildFReqPayload(prompt, uploaded);
+	const params = new URLSearchParams();
+	params.set("at", accessToken);
+	params.set("f.req", fReq);
+
+	const res = await (await resolveGeminiFetch())(GEMINI_STREAM_GENERATE_URL, {
+		method: "POST",
+		redirect: "error",
+		headers: {
+			"content-type": "application/x-www-form-urlencoded;charset=utf-8",
+			host: "gemini.google.com",
+			origin: "https://gemini.google.com",
+			referer: "https://gemini.google.com/",
+			"x-same-domain": "1",
+			"user-agent": USER_AGENT,
+			cookie: cookieHeader,
+			[MODEL_HEADER_NAME]: MODEL_HEADERS[model],
+		},
+		body: params.toString(),
+		signal: effectiveSignal,
+	});
+
+	const rawText = await res.text();
+
+	if (!res.ok) {
+		return { text: "", errorMessage: `Gemini request failed: ${res.status}` };
+	}
+
+	try {
+		return parseStreamGenerateResponse(rawText);
+	} catch (err) {
+		let errorCode: number | undefined;
+		try {
+			const json = JSON.parse(trimJsonEnvelope(rawText));
+			errorCode = extractErrorCode(json);
+		} catch {
+		}
+		return {
+			text: "",
+			errorCode,
+			errorMessage: err instanceof Error ? err.message : String(err),
+		};
+	}
+}
+
+async function fetchAccessToken(
+	cookieHeader: string,
+	signal: AbortSignal,
+): Promise<string> {
+	const html = await fetchWithCookieRedirects(GEMINI_APP_URL, cookieHeader, 10, signal);
+
+	for (const key of ["SNlM0e", "thykhd"]) {
+		const match = html.match(new RegExp(`"${key}":"(.*?)"`));
+		if (match?.[1]) return match[1];
+	}
+
+	throw new Error("Unable to authenticate with Gemini. Make sure you're signed into gemini.google.com in a supported Chromium-based browser.");
+}
+
+async function fetchWithCookieRedirects(
+	url: string,
+	cookieHeader: string,
+	maxRedirects: number,
+	signal: AbortSignal,
+): Promise<string> {
+	let current = url;
+	const allowedOrigin = new URL(url).origin;
+	for (let i = 0; i <= maxRedirects; i++) {
+		const res = await (await resolveGeminiFetch())(current, {
+			headers: { "user-agent": USER_AGENT, cookie: cookieHeader },
+			redirect: "manual",
+			signal,
+		});
+		if (res.status >= 300 && res.status < 400) {
+			const location = res.headers.get("location");
+			if (location) {
+				const next = new URL(location, current);
+				if (next.origin !== allowedOrigin) {
+					throw new Error(`Refusing to send Google cookies across origins: ${allowedOrigin} -> ${next.origin}`);
+				}
+				current = next.toString();
+				continue;
+			}
+		}
+		return await res.text();
+	}
+	throw new Error(`Too many redirects (>${maxRedirects})`);
+}
+
+function extractEmailFromGeminiHtml(html: string): string | null {
+	const patterns = [
+		// Gemini bootstraps the active account in oPEP7c. Prefer it over generic
+		// feature/config entries that may contain other signed-in Google accounts.
+		/"oPEP7c"\s*:\s*"([^"]+)"/,
+		// The Google account menu aria-label is rendered for the active account.
+		/aria-label="Google Account:[^"]*?\(([^)]+)\)"/,
+		/"displayEmail"\s*:\s*"([^"]+)"/,
+		/"defaultEmail"\s*:\s*"([^"]+)"/,
+		/"email"\s*:\s*"([^"]+)"/,
+		/"identifier"\s*:\s*"([^"]+)"/,
+		/"gaiaIdentifier"\s*:\s*"([^"]+)"/,
+	];
+
+	for (const pattern of patterns) {
+		const match = html.match(pattern);
+		const email = normalizeEmail(match?.[1]);
+		if (email) return email;
+	}
+
+	return findFirstEmail(html);
+}
+
+function extractEmailFromListAccounts(text: string): string | null {
+	const trimmed = text.replace(/^\)\]\}'\s*/, "");
+	try {
+		return findEmailInValue(JSON.parse(trimmed)) ?? findFirstEmail(trimmed);
+	} catch {
+		return findFirstEmail(trimmed);
+	}
+}
+
+function findEmailInValue(value: unknown): string | null {
+	if (typeof value === "string") return normalizeEmail(value);
+	if (Array.isArray(value)) {
+		for (const item of value) {
+			const email = findEmailInValue(item);
+			if (email) return email;
+		}
+		return null;
+	}
+	if (value && typeof value === "object") {
+		for (const item of Object.values(value as Record<string, unknown>)) {
+			const email = findEmailInValue(item);
+			if (email) return email;
+		}
+	}
+	return null;
+}
+
+function findFirstEmail(text: string): string | null {
+	const normalized = decodeEmailEscapes(text);
+	const match = normalized.match(/\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,}\b/i);
+	return match?.[0] ?? null;
+}
+
+function normalizeEmail(value: string | undefined): string | null {
+	if (!value) return null;
+	const normalized = decodeEmailEscapes(value.trim());
+	return /^[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,}$/i.test(normalized) ? normalized : null;
+}
+
+function decodeEmailEscapes(value: string): string {
+	return value
+		.replace(/\\u0040/gi, "@")
+		.replace(/\\x40/gi, "@")
+		.replace(/&#64;/gi, "@")
+		.replace(/&commat;/gi, "@")
+		.replace(/\\"/g, "\"")
+		.replace(/\\\\/g, "\\");
+}
+
+async function uploadFile(
+	filePath: string,
+	cookieHeader: string,
+	signal: AbortSignal,
+): Promise<{ id: string; name: string }> {
+	const data = readFileSync(filePath);
+	const fileName = basename(filePath);
+	const boundary = "----FormBoundary" + Math.random().toString(36).slice(2);
+	const header = `--${boundary}\r\nContent-Disposition: form-data; name="file"; filename="${fileName}"\r\nContent-Type: application/octet-stream\r\n\r\n`;
+	const footer = `\r\n--${boundary}--\r\n`;
+
+	const body = Buffer.concat([
+		Buffer.from(header, "utf-8"),
+		data,
+		Buffer.from(footer, "utf-8"),
+	]);
+
+	const res = await (await resolveGeminiFetch())(GEMINI_UPLOAD_URL, {
+		method: "POST",
+		redirect: "error",
+		headers: {
+			"content-type": `multipart/form-data; boundary=${boundary}`,
+			"push-id": GEMINI_UPLOAD_PUSH_ID,
+			"user-agent": USER_AGENT,
+			cookie: cookieHeader,
+		},
+		body,
+		signal,
+	});
+
+	if (!res.ok) {
+		const text = await res.text();
+		throw new Error(`File upload failed: ${res.status} (${text.slice(0, 200)})`);
+	}
+
+	return { id: await res.text(), name: fileName };
+}
+
+function buildFReqPayload(
+	prompt: string,
+	uploaded: Array<{ id: string; name: string }>,
+): string {
+	const promptPayload =
+		uploaded.length > 0
+			? [prompt, 0, null, uploaded.map((file) => [[file.id, 1]])]
+			: [prompt];
+	const innerList = [promptPayload, null, null];
+	return JSON.stringify([null, JSON.stringify(innerList)]);
+}
+
+function withTimeout(signal: AbortSignal | undefined, timeoutMs: number): AbortSignal {
+	const timeout = AbortSignal.timeout(timeoutMs);
+	return signal ? AbortSignal.any([signal, timeout]) : timeout;
+}
+
+function buildCookieHeader(cookieMap: CookieMap): string {
+	return Object.entries(cookieMap)
+		.filter(([, value]) => typeof value === "string" && value.length > 0)
+		.map(([name, value]) => `${name}=${value}`)
+		.join("; ");
+}
+
+function getNestedValue(value: unknown, pathParts: number[]): unknown {
+	let current: unknown = value;
+	for (const part of pathParts) {
+		if (current == null) return undefined;
+		if (!Array.isArray(current)) return undefined;
+		current = (current as unknown[])[part];
+	}
+	return current;
+}
+
+function trimJsonEnvelope(text: string): string {
+	const start = text.indexOf("[");
+	const end = text.lastIndexOf("]");
+	if (start === -1 || end === -1 || end <= start) {
+		throw new Error("Gemini response did not contain a JSON payload.");
+	}
+	return text.slice(start, end + 1);
+}
+
+function extractErrorCode(responseJson: unknown): number | undefined {
+	const code = getNestedValue(responseJson, [0, 5, 2, 0, 1, 0]);
+	return typeof code === "number" && code >= 0 ? code : undefined;
+}
+
+function extractCandidateText(candidate: unknown): string {
+	const textRaw = getNestedValue(candidate, [1, 0]);
+	let text = typeof textRaw === "string" ? textRaw : "";
+
+	if (/^http:\/\/googleusercontent\.com\/card_content\/\d+/.test(text)) {
+		const alt = getNestedValue(candidate, [22, 0]);
+		if (typeof alt === "string" && alt.length > 0) text = alt;
+	}
+
+	return text;
+}
+
+function parseStreamGenerateResponse(rawText: string): GeminiWebResult {
+	const responseJson = JSON.parse(trimJsonEnvelope(rawText));
+	const errorCode = extractErrorCode(responseJson);
+
+	const parts = Array.isArray(responseJson) ? responseJson : [];
+	let firstCandidateSeen: unknown = undefined;
+	let latestNonEmptyText = "";
+
+	for (let i = 0; i < parts.length; i++) {
+		const partBody = getNestedValue(parts[i], [2]);
+		if (!partBody || typeof partBody !== "string") continue;
+		try {
+			const parsed = JSON.parse(partBody);
+			const candidateList = getNestedValue(parsed, [4]);
+			if (!Array.isArray(candidateList) || candidateList.length === 0) continue;
+
+			const firstCandidate = (candidateList as unknown[])[0];
+			if (firstCandidateSeen === undefined) firstCandidateSeen = firstCandidate;
+
+			const text = extractCandidateText(firstCandidate);
+			if (text.length > 0) latestNonEmptyText = text;
+		} catch {
+		}
+	}
+
+	const text = latestNonEmptyText.length > 0
+		? latestNonEmptyText
+		: extractCandidateText(firstCandidateSeen);
+
+	return { text, errorCode };
+}