supermemory/apps/web/lib/url-helpers.ts

const PROXY_LOCAL_HOSTS = new Set(["localhost", "127.0.0.1", "::1"])
const DEV_APP_ORIGIN = "https://app.dev.supermemory.ai"
const PROD_APP_ORIGIN = "https://app.supermemory.ai"

export function getAppOriginForCurrentEnvironment(hostname?: string): string {
	const currentHostname =
		hostname ?? (typeof window !== "undefined" ? window.location.hostname : "")
	const normalized = currentHostname.toLowerCase()
	const isLocalOrDev =
		process.env.NODE_ENV !== "production" ||
		PROXY_LOCAL_HOSTS.has(normalized) ||
		normalized.includes("app.dev.supermemory")

	return isLocalOrDev ? DEV_APP_ORIGIN : PROD_APP_ORIGIN
}

export function getBillingSettingsUrl(hostname?: string): string {
	return `${getAppOriginForCurrentEnvironment(hostname)}/settings#billing`
}

/** Reconstruct the browser-facing URL when running behind portless (or similar). */
export function getPublicRequestUrl(request: Request): URL {
	const internal = new URL(request.url)
	const forwardedHost = request.headers
		.get("x-forwarded-host")
		?.split(",")[0]
		?.trim()
	if (forwardedHost) {
		const proto = request.headers.get("x-forwarded-proto") || "https"
		return new URL(
			`${proto}://${forwardedHost}${internal.pathname}${internal.search}`,
		)
	}
	const portlessUrl = process.env.PORTLESS_URL
	if (portlessUrl) {
		try {
			const base = new URL(portlessUrl)
			return new URL(`${base.origin}${internal.pathname}${internal.search}`)
		} catch {}
	}
	return internal
}

/** Map portless proxy localhost redirects back to the current public origin. */
export function resolveAuthRedirectUrl(
	redirectUrl: string | null,
	origin: string,
): URL {
	const fallback = new URL(origin)
	if (!redirectUrl) return fallback
	try {
		const target = new URL(redirectUrl)
		if (PROXY_LOCAL_HOSTS.has(target.hostname)) {
			return new URL(`${target.pathname}${target.search}`, origin)
		}
		if (target.origin === origin) return target
		return fallback
	} catch {
		return fallback
	}
}

/**
 * Validates if a string is a valid URL.
 */
export const isValidUrl = (url: string): boolean => {
	try {
		new URL(url)
		return true
	} catch {
		return false
	}
}

/**
 * Normalizes a URL by adding https:// prefix if missing.
 */
export const normalizeUrl = (url: string): string => {
	if (!url.trim()) return ""
	if (url.startsWith("http://") || url.startsWith("https://")) {
		return url
	}
	return `https://${url}`
}

/**
 * Checks if a URL is a Twitter/X URL.
 */
export const isTwitterUrl = (url: string): boolean => {
	const normalizedUrl = url.toLowerCase()
	return (
		normalizedUrl.includes("twitter.com") || normalizedUrl.includes("x.com")
	)
}

/**
 * Checks if a URL is a LinkedIn profile URL (not a company page).
 */
export const isLinkedInProfileUrl = (url: string): boolean => {
	const normalizedUrl = url.toLowerCase()
	return (
		normalizedUrl.includes("linkedin.com/in/") &&
		!normalizedUrl.includes("linkedin.com/company/")
	)
}

/**
 * Collects and validates URLs from LinkedIn profile and other links, excluding Twitter.
 */
export const collectValidUrls = (
	linkedinProfile: string,
	otherLinks: string[],
): string[] => {
	const urls: string[] = []

	if (linkedinProfile.trim()) {
		const normalizedLinkedIn = normalizeUrl(linkedinProfile.trim())
		if (
			isValidUrl(normalizedLinkedIn) &&
			isLinkedInProfileUrl(normalizedLinkedIn)
		) {
			urls.push(normalizedLinkedIn)
		}
	}

	otherLinks
		.filter((link) => link.trim())
		.forEach((link) => {
			const normalizedLink = normalizeUrl(link.trim())
			if (isValidUrl(normalizedLink) && !isTwitterUrl(normalizedLink)) {
				urls.push(normalizedLink)
			}
		})

	return urls
}

/**
 * Extracts X/Twitter handle from various input formats (URLs, handles with @, etc.).
 */
export function parseXHandle(input: string): string {
	if (!input.trim()) return ""

	let value = input.trim()

	if (value.startsWith("@")) {
		value = value.slice(1)
	}

	const lowerValue = value.toLowerCase()
	if (lowerValue.includes("x.com") || lowerValue.includes("twitter.com")) {
		try {
			let url: URL
			if (value.startsWith("http://") || value.startsWith("https://")) {
				url = new URL(value)
			} else {
				url = new URL(`https://${value}`)
			}

			const pathSegments = url.pathname.split("/").filter(Boolean)
			if (pathSegments.length > 0) {
				const firstSegment = pathSegments[0]
				if (firstSegment && firstSegment !== "status" && firstSegment !== "i") {
					return firstSegment
				}
			}
		} catch {
			const match = value.match(/(?:x\.com|twitter\.com)\/([^/\s?#]+)/i)
			const handle = match?.[1]
			if (handle && handle !== "status") {
				return handle
			}
		}
	}

	if (
		value.includes("/") &&
		!lowerValue.includes("x.com") &&
		!lowerValue.includes("twitter.com")
	) {
		const parts = value.split("/").filter(Boolean)
		const firstPart = parts[0]
		if (firstPart) {
			return firstPart
		}
	}

	return value
}

/**
 * Extracts LinkedIn handle from various input formats (URLs, handles with @, etc.).
 */
export function parseLinkedInHandle(input: string): string {
	if (!input.trim()) return ""

	let value = input.trim()

	if (value.startsWith("@")) {
		value = value.slice(1)
	}

	const lowerValue = value.toLowerCase()
	if (lowerValue.includes("linkedin.com")) {
		try {
			let url: URL
			if (value.startsWith("http://") || value.startsWith("https://")) {
				url = new URL(value)
			} else {
				url = new URL(`https://${value}`)
			}

			const pathMatch = url.pathname.match(/\/(in|pub)\/([^/\s?#]+)/i)
			const handle = pathMatch?.[2]
			if (handle) {
				return handle
			}
		} catch {
			const match = value.match(/linkedin\.com\/(?:in|pub)\/([^/\s?#]+)/i)
			const handle = match?.[1]
			if (handle) {
				return handle
			}
		}
	}

	if (value.includes("/in/") || value.includes("/pub/")) {
		const match = value.match(/\/(?:in|pub)\/([^/\s?#]+)/i)
		const handle = match?.[1]
		if (handle) {
			return handle
		}
	}

	return value
}

/**
 * Converts X/Twitter handle to full profile URL.
 */
export function toXProfileUrl(handle: string): string {
	if (!handle.trim()) return ""
	return `https://x.com/${handle.trim()}`
}

/**
 * Converts LinkedIn handle to full profile URL.
 */
export function toLinkedInProfileUrl(handle: string): string {
	if (!handle.trim()) return ""
	return `https://linkedin.com/in/${handle.trim()}`
}

/**
 * Checks if a URL points to a supermemory-hosted file.
 * Matches the public bucket domain (files.supermemory.ai) and
 * presigned R2 URLs whose hostname ends with `.r2.cloudflarestorage.com`.
 *
 * Note: The R2 check is intentionally broad — it matches any Cloudflare R2
 * presigned URL, not only supermemory's account.  This is acceptable because
 * the function is only called on `document.url` values returned by our own
 * backend, where all R2 URLs originate from the supermemory bucket.
 * If user-supplied external R2 URLs ever appear in this field, tighten the
 * check by also validating the account-id subdomain or the bucket path prefix.
 */
export const isSupermemoryFileUrl = (url: string): boolean => {
	try {
		const parsed = new URL(url)
		if (parsed.hostname === "files.supermemory.ai") return true
		if (parsed.hostname.endsWith(".r2.cloudflarestorage.com")) return true
		return false
	} catch {
		return false
	}
}

/**
 * Gets the favicon URL for a given URL.
 */
export function getFaviconUrl(url: string | null | undefined): string | null {
	if (!url) return null
	try {
		const urlObj = new URL(url)
		return `https://www.google.com/s2/favicons?domain=${urlObj.hostname}&sz=16`
	} catch {
		return null
	}
}

/**
 * Extracts the document ID from a Google Docs/Sheets/Slides URL.
 * Works with various URL formats:
 * - https://docs.google.com/document/d/{id}/edit
 * - https://docs.google.com/spreadsheets/d/{id}/edit#gid=0
 * - https://docs.google.com/presentation/d/{id}/edit
 */
export function extractGoogleDocId(url: string): string | null {
	try {
		const match = url.match(/\/d\/([a-zA-Z0-9_-]+)/)
		return match?.[1] ?? null
	} catch {
		return null
	}
}

/**
 * Generates the embed URL for a Google document based on its type.
 */
export function getGoogleEmbedUrl(
	docId: string,
	type: "google_doc" | "google_sheet" | "google_slide",
): string {
	switch (type) {
		case "google_doc":
			return `https://docs.google.com/document/d/${docId}/preview`
		case "google_sheet":
			return `https://docs.google.com/spreadsheets/d/${docId}/preview`
		case "google_slide":
			return `https://docs.google.com/presentation/d/${docId}/embed?start=false&loop=false&delayms=3000`
	}
}