@9000ai/cli 0.5.5 → 0.6.0

package/dist/commands/transcribe.js

@@ -1,15 +1,54 @@
 import { request, printJson, pollUntilDone } from "../client.js";
 import { loadJsonFile } from "../utils/format.js";
+/** Extract unique video_ids from monitor task output */
+function extractVideoIdsFromMonitorOutput(output) {
+    const data = (output.data ?? output);
+    const creators = (data.creators ?? []);
+    const seen = new Set();
+    for (const creator of creators) {
+        const vids = (creator.videos ?? []);
+        for (const v of vids) {
+            const vid = v.video_id;
+            if (vid)
+                seen.add(vid);
+        }
+    }
+    return [...seen];
+}
 export function registerTranscribeCommands(parent) {
     const cmd = parent.command("transcribe").description("Video / audio transcription");
     cmd
         .command("submit")
-        .description("Submit batch video-to-text task")
-        .requiredOption("--json-file <path>", "JSON file with video task details")
+        .description("Submit video-to-text task (by video IDs or JSON file)")
+        .option("--video-ids <ids>", "Comma-separated video IDs (backend resolves URLs from cache)")
+        .option("--json-file <path>", "JSON file with video task details (legacy, includes URLs)")
+        .option("--group-label <label>", "Batch group label for tracking")
+        .option("--webhook <url>", "Webhook URL for completion callback")
         .option("--wait", "Poll until task completes then print results")
         .option("--fields <list>", "Comma-separated fields to extract (used with --wait)")
         .option("--compact", "One JSON object per line (used with --wait)")
         .action(async (opts) => {
+        if (!opts.videoIds && !opts.jsonFile) {
+            console.error("Error: must provide --video-ids or --json-file");
+            process.exit(1);
+        }
+        // ── --video-ids: 新流程，后端自己查 URL + 去重 ──
+        if (opts.videoIds) {
+            const ids = opts.videoIds.split(",").map((s) => s.trim()).filter(Boolean);
+            const payload = { video_ids: ids };
+            if (opts.groupLabel)
+                payload.group_label = opts.groupLabel;
+            if (opts.webhook)
+                payload.webhook = opts.webhook;
+            const data = await request({
+                method: "POST",
+                path: "/api/v1/media/transcribe-by-video-ids",
+                payload,
+            });
+            printJson(data, { fields: opts.fields, compact: opts.compact });
+            return;
+        }
+        // ── --json-file: 兼容老流程 ──
         const payload = loadJsonFile(opts.jsonFile);
         const data = await request({
             method: "POST",
@@ -23,7 +62,6 @@ export function registerTranscribeCommands(parent) {
             printJson(data);
             return;
         }
-        // --wait: poll then fetch results
         console.error(`Batch submitted → ${batchId}, waiting...`);
         await pollUntilDone(`/api/v1/tasks/batch/${batchId}`);
         const resultData = await request({
@@ -73,4 +111,71 @@ export function registerTranscribeCommands(parent) {
             clearTimeout(timer);
         }
     });
+    // ────────────────────────────────────────────
+    // from-monitor: 监控结果 → 提取 video_id → 后端处理
+    // ────────────────────────────────────────────
+    cmd
+        .command("from-monitor")
+        .description("Extract video IDs from monitor results, send to backend for dedup + transcribe")
+        .option("--task-id <id>", "Monitor task ID to extract videos from")
+        .option("--sec-user <ids>", "Comma-separated sec_user_id list (fetches videos, caches URLs)")
+        .option("--group-label <label>", "Batch group label for tracking")
+        .option("--count <n>", "Videos per user (with --sec-user)", "20")
+        .option("--webhook <url>", "Webhook URL for completion callback")
+        .action(async (opts) => {
+        if (!opts.taskId && !opts.secUser) {
+            console.error("Error: must provide --task-id or --sec-user");
+            process.exit(1);
+        }
+        // ── Step 1: Collect video_ids ──
+        console.error("Step 1: Collecting video IDs...");
+        const allVideoIds = new Set();
+        if (opts.taskId) {
+            const taskResp = await request({ method: "GET", path: `/api/v1/tasks/${opts.taskId}` });
+            const taskData = (taskResp.data ?? taskResp);
+            if (taskData.status !== "SUCCESS") {
+                console.error(`Task status: ${taskData.status ?? "unknown"} (need SUCCESS)`);
+                process.exit(1);
+            }
+            const output = (typeof taskData.output === "string" ? JSON.parse(taskData.output) : taskData.output);
+            for (const vid of extractVideoIdsFromMonitorOutput(output)) {
+                allVideoIds.add(vid);
+            }
+        }
+        if (opts.secUser) {
+            // fetch 会同时写入后端 video_url_cache，所以只需收集 video_id
+            const secUsers = opts.secUser.split(",").map((s) => s.trim()).filter(Boolean);
+            for (const secId of secUsers) {
+                const params = new URLSearchParams({ sec_user_id: secId, count: opts.count ?? "20", sort_type: "0" });
+                console.error(`  Fetching ${secId}...`);
+                const resp = await request({ method: "GET", path: `/api/v1/douyin/monitor/user-posts?${params}` });
+                const data = (resp.data ?? resp);
+                const items = (data.items ?? data.videos ?? []);
+                for (const v of items) {
+                    const vid = v.video_id;
+                    if (vid)
+                        allVideoIds.add(vid);
+                }
+            }
+        }
+        const videoIds = [...allVideoIds];
+        console.error(`  Collected ${videoIds.length} unique video IDs`);
+        if (videoIds.length === 0) {
+            printJson({ total_videos: 0, already_transcribed: 0, newly_submitted: 0 });
+            return;
+        }
+        // ── Step 2: 调后端统一处理（去重 + 查缓存 + 转写） ──
+        console.error("Step 2: Sending to backend for dedup + transcribe...");
+        const payload = { video_ids: videoIds };
+        if (opts.groupLabel)
+            payload.group_label = opts.groupLabel;
+        if (opts.webhook)
+            payload.webhook = opts.webhook;
+        const data = await request({
+            method: "POST",
+            path: "/api/v1/media/transcribe-by-video-ids",
+            payload,
+        });
+        printJson(data);
+    });
 }

package/dist/index.js

@@ -14,7 +14,7 @@ const program = new Command();
 program
     .name("9000ai")
     .description("9000AI Toolbox CLI — unified interface for 9000AI platform")
-    .version("0.5.5");
+    .version("0.6.0");
 registerConfigCommands(program);
 registerAuthCommands(program);
 registerSearchCommands(program);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@9000ai/cli",
-  "version": "0.5.5",
+  "version": "0.6.0",
   "description": "9000AI Toolbox CLI — unified command-line interface for 9000AI platform",
   "type": "module",
   "bin": {