ZJU-REAL/Easel / skills/builtin/clipify
clipify
Verified从长视频中自动提取精彩片段,切成独立短视频,支持 16:9→9:16 竖版转制和逐字字幕烧录。 当用户说"视频切片""提取精彩片段""长视频切短""切成短视频""高光剪辑""逐字字幕""转竖版短视频"时使用。 和 video-highlights 的区别:clipify 专做英文口播找笑点+动态人脸 pan;video-highlights 更通用(中文/直播皆可),静态转竖版更稳。
| File | Kind | Size |
|---|---|---|
| EASEL-META.md | markdown | 623 B |
| SKILL.md | skill md | 8.9 KB |
| assets/preview.png | binary | 273.8 KB |
| scripts/analyze.py | script | 2.2 KB |
| scripts/audio_align.py | script | 951 B |
| scripts/build_ass.py | script | 2.6 KB |
| scripts/build_pan.py | script | 524 B |
scripts/analyze.py
View on GitHub#!/usr/bin/env python3
"""Build speaker timeline from two ROI motion files.
Usage: analyze.py LEFT_MOTION.txt RIGHT_MOTION.txt [MIN_DUR]
Stdout: JSON segments. Stderr: count summary.
"""
import re, json, sys
def parse(path):
times, vals = [], []
cur_t = None
with open(path) as f:
for line in f:
m = re.match(r"frame:\d+\s+pts:\d+\s+pts_time:([0-9.]+)", line)
if m:
cur_t = float(m.group(1)); continue
m = re.match(r"lavfi\.signalstats\.YAVG=([0-9.]+)", line)
if m and cur_t is not None:
times.append(cur_t); vals.append(float(m.group(1))); cur_t = None
return times, vals
LZ = sys.argv[1]
DZ = sys.argv[2]
MIN_DUR = float(sys.argv[3]) if len(sys.argv) > 3 else 1.0
t_l, v_l = parse(LZ)
t_d, v_d = parse(DZ)
assert len(v_l) == len(v_d), f"len mismatch: {len(v_l)} vs {len(v_d)}"
def norm(v):
m = sum(v) / max(len(v), 1)
return [x / m if m > 0 else 0 for x in v]
n_l = norm(v_l); n_d = norm(v_d)
WIN = 15
def smooth(v):
out = []
for i in range(len(v)):
a = max(0, i - WIN // 2); b = min(len(v), i + WIN // 2 + 1)
out.append(sum(v[a:b]) / (b - a))
return out
s_l = smooth(n_l); s_d = smooth(n_d)
MARGIN = 1.15
speaker = []
cur = 0 if s_l[0] >= s_d[0] else 1
for i in range(len(s_l)):
if cur == 0 and s_d[i] > s_l[i] * MARGIN: cur = 1
elif cur == 1 and s_l[i] > s_d[i] * MARGIN: cur = 0
speaker.append(cur)
fps = 30.0
segments = []
i = 0
while i < len(speaker):
j = i
while j + 1 < len(speaker) and speaker[j + 1] == speaker[i]:
j += 1
segments.append({"start": i / fps, "end": (j + 1) / fps,
"speaker": "left" if speaker[i] == 0 else "right"})
i = j + 1
merged = []
for seg in segments:
if merged and (seg["end"] - seg["start"]) < MIN_DUR:
merged[-1]["end"] = seg["end"]
else:
merged.append(seg)
collapsed = []
for seg in merged:
if collapsed and collapsed[-1]["speaker"] == seg["speaker"]:
collapsed[-1]["end"] = seg["end"]
else:
collapsed.append(seg)
print(json.dumps(collapsed, indent=2))
print(f"{len(collapsed)} segments, total {collapsed[-1]['end']:.2f}s", file=sys.stderr)
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 |