Skip to content

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"。

FileKindSize
SKILL.mdskill md4.4 KB
build_naval_cards.pyscript6.7 KB
ingest_video_material.pyscript4.7 KB
kb_to_mpt.pyscript5.5 KB
transcribe_fw.pyscript2.3 KB

transcribe_fw.py

View on GitHub
"""用 faster-whisper(项目机器人同款 STT 引擎)转写音频 -> 字幕 vtt。
权重下载根目录与项目一致:backend/data/whisper_models
用法:
  python transcribe_fw.py <audio> --out <dir> --model base|small|medium|large-v3
"""
import argparse, os, sys, time, json

PROJECT_ROOT = os.path.dirname(os.path.abspath(__file__))
DL_ROOT = os.path.join(PROJECT_ROOT, "data", "whisper_models")


def main():
    ap = argparse.ArgumentParser()
    ap.add_argument("audio")
    ap.add_argument("--out", default=os.path.join(PROJECT_ROOT, "material", "audio"))
    ap.add_argument("--model", default="base")
    ap.add_argument("--lang", default="zh")
    args = ap.parse_args()

    os.makedirs(args.out, exist_ok=True)
    os.makedirs(DL_ROOT, exist_ok=True)

    from faster_whisper import WhisperModel
    # 直接以目录路径加载(已手动下载到 data/whisper_models/<size>/),
    # 绕开 huggingface_hub 的沙箱 safe-delete 限制,不触发任何下载。
    if os.path.isdir(args.model):
        model_path = args.model
    else:
        model_path = os.path.join(DL_ROOT, args.model)
    print(f"[fw] loading faster-whisper from {model_path}", flush=True)
    t0 = time.time()
    model = WhisperModel(model_path, device="cpu", compute_type="int8")
    print(f"[fw] model ready in {time.time()-t0:.1f}s", flush=True)

    segs, info = model.transcribe(args.audio, language=args.lang,
                                  vad_filter=True, word_timestamps=False)
    print(f"[fw] lang={info.language} dur={info.duration:.1f}s", flush=True)

    stem = os.path.splitext(os.path.basename(args.audio))[0]
    vtt = os.path.join(args.out, f"{stem}.vtt")
    txt = os.path.join(args.out, f"{stem}.txt")
    with open(vtt, "w", encoding="utf-8") as vf, open(txt, "w", encoding="utf-8") as tf:
        vf.write("WEBVTT\n\n")
        n = 0
        for s in segs:
            n += 1
            ts = f"{_t(s.start)} --> {_t(s.end)}"
            vf.write(f"{ts}\n{s.text.strip()}\n\n")
            tf.write(f"[{_t(s.start)}] {s.text.strip()}\n")
    print(f"[fw] done {n} segs -> {vtt}", flush=True)


def _t(sec):
    sec = max(0, float(sec))
    h = int(sec // 3600); m = int((sec % 3600) // 60); s = sec % 60
    return f"{h:02d}:{m:02d}:{s:06.3f}".replace(".", ".")


if __name__ == "__main__":
    main()
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: 198kUpdated Claude CodeCodexVerifiedInstall
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: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
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: 194kUpdated Claude CodeCodexVerifiedInstall