feat: implement AI bouquet generation flow with Gemini/OpenAI

* feat: scaffold message, generating, and map pages and align header steps

* feat: implement AI bouquet generation flow with Gemini/OpenAI

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

src/lib/flowerFlow/api.js

@@ -0,0 +1,115 @@
+/**
+ * Error thrown for a non-OK API response, carrying the structured fields the
+ * server attaches (code/retryable/permanent/retryAfterMs) so callers can decide
+ * how to react instead of regex-matching the message.
+ */
+export class GenerationError extends Error {
+	/**
+	 * @param {string} message
+	 * @param {{ status?: number, code?: string, retryable?: boolean, permanent?: boolean, retryAfterMs?: number }} [info]
+	 */
+	constructor(message, info = {}) {
+		super(message);
+		this.name = 'GenerationError';
+		this.status = info.status ?? 0;
+		this.code = info.code ?? 'unknown';
+		this.retryable = Boolean(info.retryable);
+		this.permanent = Boolean(info.permanent);
+		this.retryAfterMs = typeof info.retryAfterMs === 'number' ? info.retryAfterMs : 0;
+	}
+}
+
+/**
+ * @param {Response} response
+ */
+async function parseResponse(response) {
+	const data = await response.json().catch(() => ({}));
+	if (!response.ok) {
+		throw new GenerationError(
+			typeof data.error === 'string' ? data.error : `Request failed (${response.status})`,
+			{
+				status: response.status,
+				code: typeof data.code === 'string' ? data.code : undefined,
+				retryable: data.retryable,
+				permanent: data.permanent,
+				retryAfterMs: data.retryAfterMs
+			}
+		);
+	}
+	return data;
+}
+
+/**
+ * @param {File} image
+ * @param {Record<string, unknown>} userInput
+ */
+export async function analyzeMood(image, userInput) {
+	const formData = new FormData();
+	formData.append('image', image);
+
+	for (const [key, value] of Object.entries(userInput)) {
+		if (value !== undefined && value !== null && value !== '') {
+			formData.append(key, String(value));
+		}
+	}
+
+	const response = await fetch('/api/flower-flow/mood-analysis', {
+		method: 'POST',
+		body: formData
+	});
+
+	return parseResponse(response);
+}
+
+/**
+ * @param {string} jobId
+ * @param {Record<string, unknown>} [userInput]
+ */
+export async function buildRecipe(jobId, userInput) {
+	const response = await fetch('/api/flower-flow/recipe', {
+		method: 'POST',
+		headers: { 'Content-Type': 'application/json' },
+		body: JSON.stringify({ jobId, userInput })
+	});
+
+	return parseResponse(response);
+}
+
+/** @param {string} jobId */
+export async function generateImages(jobId) {
+	const response = await fetch('/api/flower-flow/generate-images', {
+		method: 'POST',
+		headers: { 'Content-Type': 'application/json' },
+		body: JSON.stringify({ jobId })
+	});
+
+	return parseResponse(response);
+}
+
+/**
+ * @param {string} jobId
+ * @param {'S' | 'M' | 'L'} size
+ */
+export async function selectOption(jobId, size) {
+	const response = await fetch('/api/flower-flow/select-option', {
+		method: 'POST',
+		headers: { 'Content-Type': 'application/json' },
+		body: JSON.stringify({ jobId, size })
+	});
+
+	return parseResponse(response);
+}
+
+/** @param {string} jobId */
+export async function fetchJob(jobId) {
+	const response = await fetch(`/api/flower-flow/job?jobId=${encodeURIComponent(jobId)}`);
+	return parseResponse(response);
+}
+
+/**
+ * @param {{ mimeType?: string, base64?: string } | null | undefined} image
+ */
+export function toDataUrl(image) {
+	if (!image?.base64) return '';
+	return `data:${image.mimeType || 'image/png'};base64,${image.base64}`;
+}