anbeime/skill / skills/kb-to-mpt
kb-to-mpt
Verified知易知识库 → 短视频 流水线。把 knowledge_cards(蓝/绿/红卡)批量生成书摘/知识类短视频(复用 MoneyPrinterTurbo),并提供本地 faster-whisper 转写桥,把任意视频转文字入库成为脚本源。触发:用户想"把书/知识库做成视频""批量产出书摘视频""把视频转写进知识库"。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.4 KB |
| build_naval_cards.py | script | 6.7 KB |
| ingest_video_material.py | script | 4.7 KB |
| kb_to_mpt.py | script | 5.5 KB |
| transcribe_fw.py | script | 2.3 KB |
kb_to_mpt.py
View on GitHub#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
kb_to_mpt.py — 知易知识库 → MoneyPrinterTurbo 视频生成 桥接
复用已有的 MoneyPrinterTurbo(C:/E/MoneyPrinterTurbo-main,= 「书本视频 skill」),
把知易 knowledge_cards(蓝卡/绿卡)组装成 MPT 的自定义文案(video_script) + 素材关键词(video_terms),
推给 MPT 的本地 API 批量生成书摘/知识短视频。不重复造轮子。
用法(在已启动 MPT 的机器上运行):
# 用某专题的全部卡片生成一条视频
python kb_to_mpt.py --project-id 66 --mpt-url http://127.0.0.1:8080
# 按关键词搜卡片生成(如「纳瓦尔宝典」)
python kb_to_mpt.py --search "纳瓦尔" --subject "纳瓦尔宝典精读" --mpt-url http://127.0.0.1:8080
# 只生成文案不推(dry-run,打印 video_script)
python kb_to_mpt.py --project-id 66 --dry-run
依赖: requests(标准库 urllib 亦可,这里用 requests 简化)
"""
import argparse
import json
import sqlite3
import sys
import time
import urllib.request
import urllib.error
DB = "C:/D/zhiyi/backend/data/antinet.db"
DEFAULT_MPT = "http://127.0.0.1:8080"
def load_cards(project_id=None, search=None, limit=40):
c = sqlite3.connect(DB)
c.row_factory = sqlite3.Row
cur = c.cursor()
if project_id:
cur.execute(
"SELECT id,card_type,title,content,tags FROM knowledge_cards "
"WHERE project_id=? ORDER BY id LIMIT ?",
(project_id, limit),
)
elif search:
like = f"%{search}%"
cur.execute(
"SELECT id,card_type,title,content,tags FROM knowledge_cards "
"WHERE content LIKE ? OR title LIKE ? ORDER BY id LIMIT ?",
(like, like, limit),
)
else:
raise SystemExit("需要 --project-id 或 --search")
rows = cur.fetchall()
c.close()
return rows
def assemble_script(cards):
"""把卡片拼成一段连贯口播文案:蓝卡=事实点,绿卡=解读,红卡=行动号召。"""
lines = []
for r in cards:
t, title, content = r["card_type"], r["title"], (r["content"] or "")
# 取正文前 120 字,去掉换行的冗余
text = content.replace("\n", " ").strip()
text = text[:160]
if t == "blue":
prefix = "要点"
elif t == "green":
prefix = "解读"
elif t == "red":
prefix = "行动"
elif t == "yellow":
prefix = "提醒"
else:
prefix = "·"
lines.append(f"{prefix}:{title}。{text}")
return "\n".join(lines)
def extract_terms(cards, top_n=8):
"""优先用卡片 curated tags 作素材关键词;无 tag 时回退到标题 2 字非重叠 bigram。"""
import re, json
from collections import Counter
tags = []
for r in cards:
raw = r["tags"]
if raw:
try:
t = json.loads(raw) if isinstance(raw, str) else raw
if isinstance(t, list):
tags += [str(x) for x in t if len(str(x)) >= 2]
except Exception:
pass
if tags:
# 词频排序,去重保序
seen, out = set(), []
for w, _ in Counter(tags).most_common():
if w not in seen:
seen.add(w); out.append(w)
return out[:top_n]
# 回退:标题 2 字非重叠 bigram
stop = set("的 了 是 在 和 与 及 我们 你 我 他 这 那 也 都 就 而 把 被 让 从 到".split())
bigrams = []
for r in cards:
txt = re.sub(r"[^\u4e00-\u9fff]", "", r["title"] or "")
for i in range(0, len(txt) - 1, 2):
bg = txt[i:i+2]
if bg not in stop:
bigrams.append(bg)
return bigrams[:top_n]
def post_video(mpt_url, subject, script, terms, language="zh"):
url = f"{mpt_url.rstrip('/')}/api/v1/videos"
body = {
"video_subject": subject,
"video_script": script,
"video_terms": terms,
"video_language": language,
}
data = json.dumps(body, ensure_ascii=False).encode("utf-8")
req = urllib.request.Request(
url, data=data, headers={"Content-Type": "application/json"}, method="POST"
)
try:
with urllib.request.urlopen(req, timeout=30) as r:
return r.status, json.loads(r.read().decode("utf-8"))
except urllib.error.URLError as e:
return None, {"error": str(e)}
def main():
ap = argparse.ArgumentParser()
ap.add_argument("--project-id", type=int, default=None)
ap.add_argument("--search", type=str, default=None)
ap.add_argument("--subject", type=str, default=None)
ap.add_argument("--mpt-url", type=str, default=DEFAULT_MPT)
ap.add_argument("--limit", type=int, default=40)
ap.add_argument("--dry-run", action="store_true", help="只打印文案不上传")
a = ap.parse_args()
cards = load_cards(a.project_id, a.search, a.limit)
if not cards:
print("未命中任何卡片")
sys.exit(1)
print(f"命中卡片 {len(cards)} 张")
script = assemble_script(cards)
terms = extract_terms(cards)
subject = a.subject or (cards[0]["title"] if cards else "知识短视频")
print("\n===== 组装的 video_script =====")
print(script[:800])
print("\n===== 素材关键词 video_terms =====")
print(terms)
if a.dry_run:
print("\n[dry-run] 未上传")
return
status, resp = post_video(a.mpt_url, subject, script, terms)
print(f"\nMPT 响应: HTTP {status}")
print(json.dumps(resp, ensure_ascii=False, indent=2)[:500])
if __name__ == "__main__":
main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |