Skip to content

t27f lesson reels: course 0 lessons as data and a lesson reel generator, lesson 1 in EN and RU (Closes #10494) - #10495

Merged
gHashTag merged 1 commit into
masterfrom
claude/lesson-reels-10494
Oct 18, 2026
Merged

gHashTag merged 1 commit into
masterfrom
claude/lesson-reels-10494

Conversation

@gHashTag

Copy link
Copy Markdown
Owner

Closes #10494

The owner asked for a reel for every course and every lesson, in English and Russian, made by a generator from the topic rather than one at a time, with the source in t27. This PR is the first step. It adds course 0's lessons as data and a lesson reel generator, and uses them to render lesson 1.

What changes

  • reel_topics.t27 (new). The 27 lessons of course 0 (t27 basics), tool-written from gHashTag/trinity course.generated.ts at a pinned commit.
    • Each lesson has its title, goal, task, id and spec code, in EN and RU.
    • Text is stored as u16 code points, so the Russian survives an ASCII-only source. Spec code is stored as ASCII bytes with comment and blank lines dropped.
    • The provenance (stock sed + jq, no script) is in the header.
  • lesson_reel.t27 (new). Given a topic index LR_T and a language LR_LANG, it writes everything the token_floor template needs:
    • the tags ("LESSON 1 OF 27" plus the course title), with the holes in the course strings filled;
    • caption chunks of up to 14 characters, broken on word boundaries;
    • three code cards from the lesson's spec, never ending on a lone };
    • the page and the t27.ai CTA;
    • the ffmpeg graph for tall and wide: three voice phrases timed into the voice room, at most 1.2x tempo, under the music bed, with loudnorm.
    • Uppercasing covers Latin, Cyrillic and yo, and UTF-8 is written by hand.
  • token_floor_mux.t27. print_col, band_font, print_rise and print_fade become pub, so the lesson reel reuses them instead of copying them.

Measured (Railway lab)

  • t27c test-report: reel_topics 7/7, lesson_reel 6/6, token_floor_mux 16/16, with 0 FAIL and 0 vacuous.
  • Lesson 1, "What a spec is", was rendered as four reels: EN and RU, each 1080x1920 and 1920x1080. All four are 15.000 s at 30 fps. Integrated loudness is -14.4 LUFS (EN) and -13.9 LUFS (RU).
  • The EN voice is 11.84 s against a voice room of 12.25 s, so it plays at tempo 1.0; RU also plays at 1.0.
  • All three seals were written by t27c seal --save, and each spec_hash matches its spec.

Not claimed

  • The backdrop is the token_floor physics world, labels included. It is not a lesson-specific scene.
  • Only lesson 1 is rendered. The other 26 lessons are the same command with another LR_T, but they have not been rendered or watched.
  • The other eight courses and the blog posts are not in reel_topics yet.
  • The voice came from the internal voice-clone service. The public voice-clone URL is gone, and owner_voice_reel's recipe still names it.

🤖 Generated with Claude Code

…or, lesson 1 in EN and RU (Closes #10494)

reel_topics.t27 carries the 27 lessons of course 0 in EN and RU as code points:
title, goal, task, id and spec code. lesson_reel.t27 turns a topic index and a
language into caption chunks, code cards, tags, the CTA and the ffmpeg graph on
the token_floor template. token_floor_mux makes four text helpers pub.

Lab: reel_topics 7/7, lesson_reel 6/6, token_floor_mux 16/16, 0 vacuous.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-18 15:20:02 UTC

Summary

Status Count
Total Open PRs 50
PRs with Failing Checks 50
PRs with All Checks Green 0
READY 0
FAILING 50
PENDING 0
NO CHECKS YET 0

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=fad12492937a != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

@gHashTag
gHashTag enabled auto-merge (squash) October 18, 2026 15:20
@gHashTag
gHashTag merged commit a5a1d3e into master Oct 18, 2026
33 of 35 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

t27f lesson reels: course 0 lessons as data and a lesson reel generator, lesson 1 rendered in EN and RU

1 participant