@marshulll/openclaw-wecom 0.1.8 → 0.1.10

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@marshulll/openclaw-wecom",
-  "version": "0.1.8",
+  "version": "0.1.10",
   "type": "module",
   "description": "OpenClaw WeCom channel plugin (intelligent bot + internal app)",
   "author": "OpenClaw",

package/wecom/src/wecom-bot.ts

@@ -1,5 +1,8 @@
 import type { IncomingMessage, ServerResponse } from "node:http";
 import crypto from "node:crypto";
+import { mkdir, readdir, rm, stat, writeFile } from "node:fs/promises";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
 
 import type { PluginRuntime } from "openclaw/plugin-sdk";
 
@@ -15,6 +18,8 @@ const STREAM_MAX_ENTRIES = 500;
 const DEDUPE_TTL_MS = 2 * 60 * 1000;
 const DEDUPE_MAX_ENTRIES = 2_000;
 
+const cleanupExecuted = new Set<string>();
+
 type StreamState = {
   streamId: string;
   msgid?: string;
@@ -26,6 +31,17 @@ type StreamState = {
   content: string;
 };
 
+type InboundMedia = {
+  path: string;
+  type: string;
+  url?: string;
+};
+
+type InboundBody = {
+  text: string;
+  media?: InboundMedia;
+};
+
 const streams = new Map<string, StreamState>();
 const msgidToStreamId = new Map<string, string>();
 const recentEncrypts = new Map<string, { ts: number; streamId?: string }>();
@@ -75,6 +91,67 @@ function truncateUtf8Bytes(text: string, maxBytes: number): string {
   return slice.toString("utf8");
 }
 
+function resolveExtFromContentType(contentType: string, fallback: string): string {
+  if (!contentType) return fallback;
+  if (contentType.includes("png")) return "png";
+  if (contentType.includes("gif")) return "gif";
+  if (contentType.includes("jpeg") || contentType.includes("jpg")) return "jpg";
+  if (contentType.includes("mp4")) return "mp4";
+  if (contentType.includes("amr")) return "amr";
+  if (contentType.includes("wav")) return "wav";
+  if (contentType.includes("mp3")) return "mp3";
+  return fallback;
+}
+
+async function cleanupMediaDir(
+  dir: string,
+  retentionHours?: number,
+  cleanupOnStart?: boolean,
+): Promise<void> {
+  if (cleanupOnStart === false) return;
+  if (!retentionHours || retentionHours <= 0) return;
+  if (cleanupExecuted.has(dir)) return;
+  cleanupExecuted.add(dir);
+  const cutoff = Date.now() - retentionHours * 3600 * 1000;
+  try {
+    const entries = await readdir(dir);
+    await Promise.all(entries.map(async (entry) => {
+      const full = join(dir, entry);
+      try {
+        const info = await stat(full);
+        if (info.isFile() && info.mtimeMs < cutoff) {
+          await rm(full, { force: true });
+        }
+      } catch {
+        // ignore
+      }
+    }));
+  } catch {
+    // ignore
+  }
+}
+
+function resolveMediaTempDir(target: WecomWebhookTarget): string {
+  return target.account.config.media?.tempDir?.trim()
+    || join(tmpdir(), "openclaw-wecom");
+}
+
+function resolveMediaMaxBytes(target: WecomWebhookTarget): number | undefined {
+  const maxBytes = target.account.config.media?.maxBytes;
+  return typeof maxBytes === "number" && maxBytes > 0 ? maxBytes : undefined;
+}
+
+function sanitizeFilename(name: string, fallback: string): string {
+  const base = name.split(/[/\\\\]/).pop() ?? "";
+  const trimmed = base.trim();
+  const safe = trimmed
+    .replace(/[^\w.\-() ]+/g, "_")
+    .replace(/\s+/g, " ")
+    .trim();
+  const finalName = safe.slice(0, 120);
+  return finalName || fallback;
+}
+
 function jsonOk(res: ServerResponse, body: unknown): void {
   res.statusCode = 200;
   res.setHeader("Content-Type", "application/json; charset=utf-8");
@@ -254,7 +331,8 @@ async function startAgentForStream(params: {
   const userid = msg.from?.userid?.trim() || "unknown";
   const chatType = msg.chattype === "group" ? "group" : "direct";
   const chatId = msg.chattype === "group" ? (msg.chatid?.trim() || "unknown") : userid;
-  const rawBody = buildInboundBody(msg);
+  const inbound = await buildInboundBody({ target, msg });
+  const rawBody = inbound.text;
 
   const route = core.channel.routing.resolveAgentRoute({
     cfg: config,
@@ -301,6 +379,14 @@ async function startAgentForStream(params: {
     OriginatingTo: `wecom:${chatId}`,
   });
 
+  if (inbound.media) {
+    ctxPayload.MediaPath = inbound.media.path;
+    ctxPayload.MediaType = inbound.media.type;
+    if (inbound.media.url) {
+      ctxPayload.MediaUrl = inbound.media.url;
+    }
+  }
+
   await core.channel.session.recordInboundSession({
     storePath,
     sessionKey: ctxPayload.SessionKey ?? route.sessionKey,
@@ -392,20 +478,215 @@ async function startAgentForStream(params: {
   }
 }
 
-function buildInboundBody(msg: WecomInboundMessage): string {
+function pickString(...values: unknown[]): string {
+  for (const value of values) {
+    if (typeof value === "string" && value.trim()) return value.trim();
+  }
+  return "";
+}
+
+function resolveBotMediaUrl(msg: any, msgtype: "image" | "voice" | "video" | "file"): string {
+  if (!msg || typeof msg !== "object") return "";
+  const block = msg[msgtype] ?? {};
+  if (msgtype === "image") {
+    return pickString(
+      block.url,
+      block.picurl,
+      block.picUrl,
+      block.pic_url,
+      msg.picurl,
+      msg.picUrl,
+      msg.pic_url,
+      msg.url,
+    );
+  }
+  if (msgtype === "voice") {
+    return pickString(
+      block.url,
+      block.fileurl,
+      block.fileUrl,
+      block.file_url,
+      block.mediaUrl,
+      block.media_url,
+      msg.voiceUrl,
+      msg.voice_url,
+      msg.url,
+    );
+  }
+  if (msgtype === "video") {
+    return pickString(
+      block.url,
+      block.fileurl,
+      block.fileUrl,
+      block.file_url,
+      block.mediaUrl,
+      block.media_url,
+      msg.videoUrl,
+      msg.video_url,
+      msg.url,
+    );
+  }
+  return pickString(
+    block.url,
+    block.fileurl,
+    block.fileUrl,
+    block.file_url,
+    block.mediaUrl,
+    block.media_url,
+    msg.fileUrl,
+    msg.file_url,
+    msg.url,
+  );
+}
+
+function resolveBotMediaBase64(msg: any, msgtype: "image" | "voice" | "video" | "file"): string {
+  if (!msg || typeof msg !== "object") return "";
+  const block = msg[msgtype] ?? {};
+  return pickString(
+    block.base64,
+    block.data,
+    msg.base64,
+    msg.data,
+  );
+}
+
+function resolveBotMediaFilename(msg: any): string {
+  if (!msg || typeof msg !== "object") return "";
+  const block = msg.file ?? {};
+  return pickString(
+    block.filename,
+    block.fileName,
+    block.name,
+    block.file_name,
+    msg.filename,
+    msg.fileName,
+    msg.name,
+  );
+}
+
+async function buildBotMediaMessage(params: {
+  target: WecomWebhookTarget;
+  msgtype: "image" | "voice" | "video" | "file";
+  url?: string;
+  base64?: string;
+  filename?: string;
+}): Promise<InboundBody> {
+  const { target, msgtype, url, base64, filename } = params;
+
+  const fallbackLabel = msgtype === "image"
+    ? "[image]"
+    : msgtype === "voice"
+      ? "[voice]"
+      : msgtype === "video"
+        ? "[video]"
+        : "[file]";
+
+  if (!url && !base64) return { text: fallbackLabel };
+
+  try {
+    let buffer: Buffer | null = null;
+    let contentType = "";
+    if (base64) {
+      buffer = Buffer.from(base64, "base64");
+      if (msgtype === "image") contentType = "image/jpeg";
+      else if (msgtype === "voice") contentType = "audio/amr";
+      else if (msgtype === "video") contentType = "video/mp4";
+      else contentType = "application/octet-stream";
+    } else if (url) {
+      const media = await fetchMediaFromUrl(url, target.account);
+      buffer = media.buffer;
+      contentType = media.contentType;
+    }
+
+    if (!buffer) return { text: fallbackLabel };
+
+    const maxBytes = resolveMediaMaxBytes(target);
+    if (maxBytes && buffer.length > maxBytes) {
+      if (msgtype === "image") return { text: "[图片过大，未处理]\n\n请发送更小的图片。" };
+      if (msgtype === "voice") return { text: "[语音消息过大，未处理]\n\n请发送更短的语音消息。" };
+      if (msgtype === "video") return { text: "[视频过大，未处理]\n\n请发送更小的视频。" };
+      return { text: "[文件过大，未处理]\n\n请发送更小的文件。" };
+    }
+
+    const tempDir = resolveMediaTempDir(target);
+    await mkdir(tempDir, { recursive: true });
+    await cleanupMediaDir(
+      tempDir,
+      target.account.config.media?.retentionHours,
+      target.account.config.media?.cleanupOnStart,
+    );
+
+    const fallbackExt = msgtype === "image"
+      ? "jpg"
+      : msgtype === "voice"
+        ? "amr"
+        : msgtype === "video"
+          ? "mp4"
+          : "bin";
+    const ext = resolveExtFromContentType(contentType, fallbackExt);
+
+    if (msgtype === "file") {
+      const safeName = sanitizeFilename(filename || "", `file-${Date.now()}.${ext}`);
+      const tempFilePath = join(tempDir, safeName);
+      await writeFile(tempFilePath, buffer);
+      return {
+        text: `[用户发送了一个文件: ${safeName}，已保存到: ${tempFilePath}]\n\n请根据文件内容回复用户。`,
+        media: { path: tempFilePath, type: contentType || "application/octet-stream", url },
+      };
+    }
+
+    const tempPath = join(
+      tempDir,
+      `${msgtype}-${Date.now()}-${Math.random().toString(36).slice(2)}.${ext}`,
+    );
+    await writeFile(tempPath, buffer);
+
+    if (msgtype === "image") {
+      return {
+        text: `[用户发送了一张图片，已保存到: ${tempPath}]\n\n请使用 Read 工具查看这张图片并描述内容。`,
+        media: { path: tempPath, type: contentType || "image/jpeg", url },
+      };
+    }
+    if (msgtype === "voice") {
+      return {
+        text: `[用户发送了一条语音消息，已保存到: ${tempPath}]\n\n请根据语音内容回复用户。`,
+        media: { path: tempPath, type: contentType || "audio/amr", url },
+      };
+    }
+    if (msgtype === "video") {
+      return {
+        text: `[用户发送了一个视频文件，已保存到: ${tempPath}]\n\n请根据视频内容回复用户。`,
+        media: { path: tempPath, type: contentType || "video/mp4", url },
+      };
+    }
+    return { text: fallbackLabel };
+  } catch (err) {
+    target.runtime.error?.(`wecom bot ${msgtype} download failed: ${String(err)}`);
+    if (msgtype === "image") return { text: "[用户发送了一张图片，但下载失败]\n\n请告诉用户图片处理暂时不可用。" };
+    if (msgtype === "voice") return { text: "[用户发送了一条语音消息，但下载失败]\n\n请告诉用户语音处理暂时不可用。" };
+    if (msgtype === "video") return { text: "[用户发送了一个视频，但下载失败]\n\n请告诉用户视频处理暂时不可用。" };
+    return { text: "[用户发送了一个文件，但下载失败]\n\n请告诉用户文件处理暂时不可用。" };
+  }
+}
+
+async function buildInboundBody(params: { target: WecomWebhookTarget; msg: WecomInboundMessage }): Promise<InboundBody> {
+  const { target, msg } = params;
   const msgtype = String(msg.msgtype ?? "").toLowerCase();
   if (msgtype === "text") {
     const content = (msg as any).text?.content;
-    return typeof content === "string" ? content : "";
+    return { text: typeof content === "string" ? content : "" };
   }
   if (msgtype === "voice") {
     const content = (msg as any).voice?.content;
-    return typeof content === "string" ? content : "[voice]";
+    if (typeof content === "string" && content.trim()) return { text: content.trim() };
+    const url = resolveBotMediaUrl(msg as any, "voice");
+    const base64 = resolveBotMediaBase64(msg as any, "voice");
+    return await buildBotMediaMessage({ target, msgtype: "voice", url, base64 });
   }
   if (msgtype === "mixed") {
     const items = (msg as any).mixed?.msg_item;
     if (Array.isArray(items)) {
-      return items
+      const text = items
         .map((item: any) => {
           const t = String(item?.msgtype ?? "").toLowerCase();
           if (t === "text") return String(item?.text?.content ?? "");
@@ -414,30 +695,35 @@ function buildInboundBody(msg: WecomInboundMessage): string {
         })
         .filter((part: string) => Boolean(part && part.trim()))
         .join("\n");
+      return { text };
     }
-    return "[mixed]";
+    return { text: "[mixed]" };
   }
   if (msgtype === "image") {
-    const url = String((msg as any).image?.url ?? "").trim();
-    return url ? `[image] ${url}` : "[image]";
+    const url = resolveBotMediaUrl(msg as any, "image");
+    const base64 = resolveBotMediaBase64(msg as any, "image");
+    return await buildBotMediaMessage({ target, msgtype: "image", url, base64 });
   }
   if (msgtype === "file") {
-    const url = String((msg as any).file?.url ?? "").trim();
-    return url ? `[file] ${url}` : "[file]";
+    const url = resolveBotMediaUrl(msg as any, "file");
+    const base64 = resolveBotMediaBase64(msg as any, "file");
+    const filename = resolveBotMediaFilename(msg as any);
+    return await buildBotMediaMessage({ target, msgtype: "file", url, base64, filename });
   }
   if (msgtype === "video") {
-    const url = String((msg as any).video?.url ?? "").trim();
-    return url ? `[video] ${url}` : "[video]";
+    const url = resolveBotMediaUrl(msg as any, "video");
+    const base64 = resolveBotMediaBase64(msg as any, "video");
+    return await buildBotMediaMessage({ target, msgtype: "video", url, base64 });
   }
   if (msgtype === "event") {
     const eventtype = String((msg as any).event?.eventtype ?? "").trim();
-    return eventtype ? `[event] ${eventtype}` : "[event]";
+    return { text: eventtype ? `[event] ${eventtype}` : "[event]" };
   }
   if (msgtype === "stream") {
     const id = String((msg as any).stream?.id ?? "").trim();
-    return id ? `[stream_refresh] ${id}` : "[stream_refresh]";
+    return { text: id ? `[stream_refresh] ${id}` : "[stream_refresh]" };
   }
-  return msgtype ? `[${msgtype}]` : "";
+  return { text: msgtype ? `[${msgtype}]` : "" };
 }
 
 function normalizeMediaType(raw?: string): "image" | "voice" | "video" | "file" | null {