Skip to content

GUIDE.md

Latest commit

 

History

History
103 lines (73 loc) · 3.94 KB

GUIDE.md

File metadata and controls

103 lines (73 loc) · 3.94 KB
Raw

Creative Automation — Quick Guide

Turn any Notion textbook chapter into 80 images + 80 animated videos. One config block, two double-clicks.

First-time setup (do once)

  1. Double-click setup.bat (in D:\Creative_Automation\)
    • Installs Python dependencies
    • Launches two Chrome windows (one for ChatGPT, one for Gemini + Flow)
  2. Sign in inside each window:
    • Window 1 (ChatGPT) → log in to chatgpt.com
    • Window 2 (Gemini + Flow) → log in to gemini.google.com (tab 1) and labs.google/fx/tools/flow (tab 2)
  3. Leave both Chrome windows open. Sign-in is one-time — sessions persist.

Run a chapter

  1. Open D:\Creative_Automation\scripts\run_pipeline.cjs in any text editor

  2. Edit the CONFIG block at the top (around line 27):

    const CONFIG = {
  NOTION_URL: 'https://www.notion.so/PASTE-YOUR-CHAPTER-LINK-HERE',
  GRADE:   7,                  // 5–11
  LANG:    'uz',               // 'uz' or 'ru'
  SUBJECT: 'jahon tarixi',     // subject name
  CHAPTER: 1,                  // chapter number
};

    GRADE, LANG, SUBJECT, CHAPTER are what actually drive the fetch. The URL is for your reference (printed at the top of the run).

  3. Save the file.

  4. Double-click start.bat.

  5. Walk away. Total time ≈ 3–4 hours per chapter.

What runs (5 stages)

# Stage What happens Time
1 FETCH Pulls chapter text from Notion ~10 sec
2 REFINE ChatGPT rewrites the chapter 3–6 min
3 PROMPTS ChatGPT generates 80 image prompts (4 batches × 20) 10–20 min
4 IMAGES Gemini generates + saves 80 PNGs (10 parallel tabs) ~60 min
5 ANIMATE Flow turns each PNG into an 8-sec MP4 (4 in flight) ~2 hr

Each stage skips itself if its output already exists — re-running after a crash picks up exactly where it left off.

Where everything is saved

All paths under D:\Creative_Automation\:

Stage output Folder
Raw chapter (Notion) chapters\g{GRADE}-{LANG}\{subject-slug}\ch{NN}-{title}.md
Refined chapter (ChatGPT) refined\g{GRADE}-{LANG}\{subject-slug}\ch{NN}-{title}.md
80 prompts (the JSON the rest of the pipeline reads) prompts\g{GRADE}-{LANG}\{subject-slug}\ch{NN}-{title}.json
80 images images\g{GRADE}-{LANG}\{subject-slug}\ch{NN}-{title}\NNN-{slug}.png
80 videos videos\g{GRADE}-{LANG}\{subject-slug}\ch{NN}-{title}\NNN-{slug}.mp4

Concrete example (G7 Uzbek, jahon tarixi, ch10 Saljuqiylar):

chapters\g7-uz\jahon-tarixi\ch10-saljuqiylar-davlati.md
refined\g7-uz\jahon-tarixi\ch10-saljuqiylar-davlati.md
prompts\g7-uz\jahon-tarixi\ch10-saljuqiylar-davlati.json
images\g7-uz\jahon-tarixi\ch10-saljuqiylar-davlati\001-...png ... 080-...png
videos\g7-uz\jahon-tarixi\ch10-saljuqiylar-davlati\001-...mp4 ... 080-...mp4

Troubleshooting

Pipeline halts on stage 1 with "ModuleNotFoundError" — run setup.bat once (it installs Python deps).

Pipeline halts on pre-flight with "Chrome not reachable" — your Chrome window for that port is closed. Run setup.bat again (it skips windows that are already up).

A stage failed mid-run — fix whatever the error says, then double-click start.bat again. Completed stages skip automatically.

Want to redo a stage — delete its output file/folder, then re-run start.bat.

Want to run a different chapter — edit the CONFIG block, save, double-click start.bat.

File reference

File Purpose
setup.bat Run once to install deps and launch Chrome windows
start.bat Run every time to start the pipeline
scripts\run_pipeline.cjs The orchestrator — edit the CONFIG block here
.env API keys (don't share)
refine_prompt.txt The formula ChatGPT uses to refine chapters
80_prompt_formula.txt The formula ChatGPT uses to generate image prompts