feat(platform): 开源发布 - 初始化 Open AI Canvas

web/src/lib/generation-error.ts

@@ -0,0 +1,47 @@
+export const CONTENT_MODERATION_ERROR_CODE = "sensitive_words_detected";
+
+export const CONTENT_MODERATION_MESSAGE = "内容审核未通过，本次平台积分未扣除或已退还。请修改提示词后重新生成。";
+
+export type GenerationFailureMetadata = {
+    errorDetails: string;
+    generationErrorCode?: string;
+    failedPromptFingerprint?: string;
+};
+
+export function generationFailureMetadata(error: unknown, prompt: string): GenerationFailureMetadata {
+    const raw = error instanceof Error ? error.message : String(error || "生成失败");
+    if (!isContentModerationError(raw)) return { errorDetails: raw || "生成失败" };
+    return {
+        errorDetails: CONTENT_MODERATION_MESSAGE,
+        generationErrorCode: CONTENT_MODERATION_ERROR_CODE,
+        failedPromptFingerprint: generationPromptFingerprint(prompt),
+    };
+}
+
+export function generationErrorMessage(error: unknown) {
+    const raw = error instanceof Error ? error.message : String(error || "生成失败");
+    return isContentModerationError(raw) ? CONTENT_MODERATION_MESSAGE : raw || "生成失败";
+}
+
+export function isContentModerationError(value: unknown) {
+    const text = value instanceof Error ? value.message : String(value || "");
+    return text.toLowerCase().includes(CONTENT_MODERATION_ERROR_CODE) || text.includes("内容审核未通过");
+}
+
+export function unchangedModeratedPrompt(metadata: { errorDetails?: string; generationErrorCode?: string; failedPromptFingerprint?: string } | undefined, prompt: string) {
+    const moderationFailure = metadata?.generationErrorCode === CONTENT_MODERATION_ERROR_CODE || isContentModerationError(metadata?.errorDetails);
+    if (!moderationFailure) return false;
+    if (!metadata?.failedPromptFingerprint) return true;
+    return metadata.failedPromptFingerprint === generationPromptFingerprint(prompt);
+}
+
+// 指纹只用于识别“原样重试”，不是安全或鉴权用途。
+export function generationPromptFingerprint(value: string) {
+    const normalized = value.trim().replace(/\s+/g, " ");
+    let hash = 2166136261;
+    for (let index = 0; index < normalized.length; index += 1) {
+        hash ^= normalized.charCodeAt(index);
+        hash = Math.imul(hash, 16777619);
+    }
+    return `${normalized.length}:${(hash >>> 0).toString(36)}`;
+}