ip-hand-drawn-infographic-planner

name

ip-hand-drawn-infographic-planner

description

Plan concise hand-drawn infographic inputs for IP incubation, business cognition, and AI cognition whiteboard explainer videos. Use when Codex needs to convert polished_voiceover.md, voiceover_segments.json, or visual_beats.json into infographic/infographic_plan.json, board_specs/*.json, and image_prompts/*.prompt.md for hand-drawn-infographic-creator without rendering images, writing board coordinates, or producing video.

IP Hand-Drawn Infographic Planner

Overview

Turn polished short-video voiceover and visual beats into a compact hand-drawn infographic plan. The output prepares structured inputs for hand-drawn-infographic-creator and the later video board control layer.

Non-Goals

• Do not generate PNGs, videos, audio, subtitles, or HyperFrames projects.
• Do not replace hand-drawn-infographic-creator; only prepare its prompt-level inputs.
• Do not write board_manifest.json, motion_plan.json, bbox, camera, cursor, or annotation coordinates.
• Do not copy the voiceover sentence by sentence onto the board.

Required Inputs

Prefer a project package with:

script/
├── polished_voiceover.md
├── voiceover_segments.json
└── visual_beats.json

If only prose voiceover is available, first segment it into stable segment ids and visual beats before writing infographic outputs.

Workflow

1. Read references/schema.md before writing outputs.
2. Read references/business-cognition-diagram-patterns.md before deciding board count and diagram type.
3. Read references/hand-drawn-prompt-bridge.md before writing image prompts.
4. Extract the one core thesis and the few visual beats that actually need a board.
5. Decide board count:
   • Use one board for one thesis, one contrast, or up to 5 key objects.
   • Use two boards when the voiceover has two separate jobs, usually problem framing plus method/checklist.
   • Use three boards only when the script is longer than the normal 30-60 second package or has clearly separate concepts.
6. For each board, keep contentDensity as simple, set maxKeyObjects to 3-5, and give every key object a stable semantic id.
7. Write:

infographic/
├── infographic_plan.json
├── board_specs/
│   └── board-01.board_spec.json
└── image_prompts/
    └── board-01.prompt.md

8. Run the validator:

python3 ip-hand-drawn-infographic-planner/scripts/validate_infographic_plan.py <project-output>

Planning Rules

• Each board carries one main point, not a transcript.
• Each prompt must include this exact sentence: 内容简洁一点，不要逐字写满口播.
• Use style vocabulary from hand-drawn-infographic-creator: continuous line art, engineer's notebook sketch with annotations in margins, whiteboard explanation aesthetic, ink on parchment.
• Use the hand-drawn palette: parchment #faf8f3, charcoal #1a2332, ocean-blue annotations #2d5a7b, and only 1-2 semantic highlight colors per board.
• Use coral #e63946 for risk/problem, teal #4a9d9e for active/positive, and amber #f4a261 for progress/insight.
• Key object ids must be lowercase, stable, semantic, and reusable by downstream control-layer Skills, for example workflow_asset, not left_box_1.
• Keep board specs semantic. Use layout zones such as left_panel or bottom_takeaway; do not use exact coordinates.

Resources

• references/schema.md: output schema and validation expectations.
• references/business-cognition-diagram-patterns.md: board splitting and diagram selection rules.
• references/hand-drawn-prompt-bridge.md: prompt template aligned to hand-drawn-infographic-creator.
• scripts/validate_infographic_plan.py: deterministic contract validator.
• _examples/planner/ai-cognition-example/: complete commercial cognition / AI cognition sample.

Handoff

After validation, pass image_prompts/*.prompt.md to hand-drawn-infographic-creator for image generation. Pass board_specs/*.json plus generated PNGs to hand-drawn-infographic-video-board for controllable board manifests and motion plans.