evOo-local/packages/core/src/attachments.ts

import type { AttachmentKind, AttachmentMimeType, MessageBlock } from "@rakazo/contracts";
import {
  ATTACHMENT_MAX_BASE64_LENGTH,
  ATTACHMENT_MAX_BYTES,
  attachmentKindOf,
  isAllowedAttachmentMimeType,
  isAttachmentImageMimeType,
} from "@rakazo/contracts";

export class AttachmentValidationError extends Error {
  constructor(message: string) {
    super(message);
    this.name = "AttachmentValidationError";
  }
}

export function decodeAttachmentBase64(contentBase64: string): Uint8Array {
  const normalized = contentBase64.trim();
  if (!normalized) throw new AttachmentValidationError("Attachment content is empty");
  if (normalized.length > ATTACHMENT_MAX_BASE64_LENGTH) {
    throw new AttachmentValidationError("Attachment exceeds the 10 MiB limit");
  }
  if (
    normalized.length % 4 !== 0 ||
    !/^(?:[A-Za-z0-9+/]{4})*(?:[A-Za-z0-9+/]{2}==|[A-Za-z0-9+/]{3}=)?$/.test(normalized)
  ) {
    throw new AttachmentValidationError("Attachment content is not valid base64");
  }
  let bytes: Buffer;
  try {
    bytes = Buffer.from(normalized, "base64");
  } catch {
    throw new AttachmentValidationError("Attachment content is not valid base64");
  }
  if (bytes.byteLength === 0) throw new AttachmentValidationError("Attachment content is empty");
  if (bytes.byteLength > ATTACHMENT_MAX_BYTES) {
    throw new AttachmentValidationError("Attachment exceeds the 10 MiB limit");
  }
  return new Uint8Array(bytes);
}

export function validateAttachmentMimeType(mimeType: string): void {
  if (!isAllowedAttachmentMimeType(mimeType)) {
    throw new AttachmentValidationError(`Unsupported attachment type: ${mimeType}`);
  }
}

export function attachmentKindForMimeType(mimeType: string): AttachmentKind {
  return attachmentKindOf(mimeType);
}

export function messageBlockForArtifact(artifact: {
  id: string;
  name: string;
  mimeType: string;
  size: number;
  width?: number | null;
  height?: number | null;
  durationMs?: number | null;
}): Extract<MessageBlock, { kind: "image" | "video" | "audio" | "file" }> {
  const kind = attachmentKindForMimeType(artifact.mimeType);
  if (kind === "image") {
    return {
      kind: "image",
      artifactId: artifact.id,
      mimeType: artifact.mimeType,
      name: artifact.name,
    };
  }
  if (kind === "video") {
    return {
      kind: "video",
      artifactId: artifact.id,
      mimeType: artifact.mimeType,
      name: artifact.name,
      size: artifact.size,
      ...(artifact.durationMs != null ? { durationMs: artifact.durationMs } : {}),
      ...(artifact.width != null ? { width: artifact.width } : {}),
      ...(artifact.height != null ? { height: artifact.height } : {}),
    };
  }
  if (kind === "audio") {
    return {
      kind: "audio",
      artifactId: artifact.id,
      mimeType: artifact.mimeType,
      name: artifact.name,
      size: artifact.size,
      ...(artifact.durationMs != null ? { durationMs: artifact.durationMs } : {}),
    };
  }
  return {
    kind: "file",
    artifactId: artifact.id,
    mimeType: artifact.mimeType,
    name: artifact.name,
    size: artifact.size,
  };
}

export function promptTextForAttachments(
  text: string | undefined,
  artifacts: Array<{ name: string; mimeType: string; size: number }>,
): string {
  const caption = text?.trim() ?? "";
  const notes = artifacts
    .filter((artifact) => !isAttachmentImageMimeType(artifact.mimeType))
    .map(
      (artifact) =>
        `User attached file ${JSON.stringify(artifact.name)} (${artifact.mimeType}, ${artifact.size} bytes).`,
    );
  return [caption, ...notes].filter(Boolean).join("\n\n") || "See attached files.";
}

export function blocksToAgentHistoryText(blocks: MessageBlock[]): string {
  return blocks
    .map((block) => {
      if (block.kind === "text") return block.text;
      if (block.kind === "chart") return `[chart: ${block.name}]`;
      if (block.kind === "image") return `[image: ${block.name}]`;
      if (block.kind === "file") {
        return `[file: ${block.name} (${block.mimeType}, ${block.size} bytes)]`;
      }
      // Media keeps the artifact id so a later turn can edit or export what it refers to.
      if (block.kind === "video" || block.kind === "audio") {
        return `[${block.kind}: ${block.name} (artifact ${block.artifactId})]`;
      }
      if (block.kind === "media") {
        const outputs = block.outputs
          .map((output) => `${output.kind} ${output.name} (artifact ${output.artifactId})`)
          .join(", ");
        return `[media ${block.operation} ${block.status}: ${block.title}${outputs ? ` -> ${outputs}` : ""}]`;
      }
      // Keep attribution on peer messages: without it a later turn cannot tell
      // which lines came from another bot rather than the user.
      if (block.kind === "bot_message_received") return `[from ${block.fromBotName}] ${block.text}`;
      if (block.kind === "bot_message_sent") return `[to ${block.toBotName}] ${block.text}`;
      if (block.kind === "task") {
        return `[task for ${block.assigneeBotName}: ${block.title} (${block.status})] ${block.text}`;
      }
      if (block.kind === "handoff") {
        return `[handoff ${block.fromBotId} -> ${block.toBotId}] ${block.text}`;
      }
      if ("text" in block && typeof block.text === "string") return block.text;
      return "";
    })
    .filter(Boolean)
    .join("\n");
}

const EXTENSION_MIME_TYPES: Record<string, AttachmentMimeType> = {
  ".jpg": "image/jpeg",
  ".jpeg": "image/jpeg",
  ".png": "image/png",
  ".webp": "image/webp",
  ".gif": "image/gif",
  ".pdf": "application/pdf",
  ".txt": "text/plain",
  ".md": "text/markdown",
  ".markdown": "text/markdown",
  ".csv": "text/csv",
  ".json": "application/json",
  ".mp4": "video/mp4",
  ".m4v": "video/mp4",
  ".webm": "video/webm",
  ".mov": "video/quicktime",
  ".mp3": "audio/mpeg",
  ".m4a": "audio/mp4",
  ".wav": "audio/wav",
  ".ogg": "audio/ogg",
  ".oga": "audio/ogg",
  ".vtt": "text/vtt",
  ".srt": "application/x-subrip",
};

const MIME_TYPE_EXTENSIONS: Record<AttachmentMimeType, string> = {
  "image/jpeg": ".jpg",
  "image/png": ".png",
  "image/webp": ".webp",
  "image/gif": ".gif",
  "application/pdf": ".pdf",
  "text/plain": ".txt",
  "text/markdown": ".md",
  "text/csv": ".csv",
  "application/json": ".json",
  "video/mp4": ".mp4",
  "video/webm": ".webm",
  "video/quicktime": ".mov",
  "audio/mpeg": ".mp3",
  "audio/mp4": ".m4a",
  "audio/wav": ".wav",
  "audio/webm": ".weba",
  "audio/ogg": ".ogg",
  "text/vtt": ".vtt",
  "application/x-subrip": ".srt",
};

export function inferAttachmentMimeType(
  name: string,
  reportedType?: string,
): AttachmentMimeType | null {
  const dot = name.lastIndexOf(".");
  const extensionType = dot < 0 ? undefined : EXTENSION_MIME_TYPES[name.slice(dot).toLowerCase()];
  // Some browsers and native document pickers report Markdown as text/plain.
  if (extensionType === "text/markdown" && (!reportedType || reportedType === "text/plain")) {
    return extensionType;
  }
  if (reportedType && isAllowedAttachmentMimeType(reportedType)) return reportedType;
  return extensionType ?? null;
}

export function attachmentExtensionForMimeType(mimeType: string): string {
  return isAllowedAttachmentMimeType(mimeType) ? MIME_TYPE_EXTENSIONS[mimeType] : "";
}

export function attachmentsForThread<T extends { threadKey: string }>(
  attachments: readonly T[],
  threadKey: string | undefined,
): T[] {
  if (!threadKey) return [];
  return attachments.filter((attachment) => attachment.threadKey === threadKey);
}

export function attachmentsForBot<T extends { botId: string }>(
  attachments: readonly T[],
  botId: string | undefined,
): T[] {
  if (!botId) return [];
  return attachments.filter((attachment) => attachment.botId === botId);
}

export function userTurnBlocksForRun(
  trigger: string,
  runId: string,
  messages: Array<{
    id?: string;
    role: string;
    runId?: string | null;
    blocks: MessageBlock[];
  }>,
  sourceMessageId?: string | null,
): MessageBlock[] | undefined {
  if (trigger !== "user") return undefined;
  return messages.find(
    (message) =>
      message.role === "user" &&
      (sourceMessageId ? message.id === sourceMessageId : message.runId === runId),
  )?.blocks;
}