使い方
- インストールコマンドをコピーし、自分の環境へ導入する前にソースを確認します。
- 必要なツール、認証情報、対応環境を原文で確認します。
- 入力と成果物を指定し、出力を資料と照合します。
npx skills add https://github.com/nphu0811/pmg-skills --skill pmg-diagram-motion開始前の確認
- 導入にはモデル契約、API枠、ファイルへのアクセス許可は含まれません。
- スター数や導入数は活動の指標であり、独立した品質・安全性試験ではありません。
見出し、例、コード、表、リンク、参照画像を含む原文を表示しています。
PMG Diagram Motion (pmg-diagram-motion)
Specialized motion intelligence for flowcharts, process pipelines, and system architecture diagrams in PowerPoint presentations.
Core Mental Model
Diagram animation must strictly follow causal topology. Never reveal nodes randomly or all at once: $$\text{Node A} \longrightarrow \text{Connector AB} \longrightarrow \text{Node B} \longrightarrow \text{Connector BC} \longrightarrow \text{Node C}$$ - Connectors wipe in the exact direction of causal flow (left-to-right, top-to-bottom). - Hierarchical architectures expand from central core roots outward to child modules. - Data packet indicators travel along pipelines to demonstrate real-time throughput.
Supported Patterns
| Pattern | Shared Primitive | Native PPTX Effect | Semantic Application |
|---|---|---|---|
| `process_flow` | SEQUENCE | fly_in / appear | Sequential node appearance following causality. |
| `connector_drawing` | DRAW | wipe (directional) | Connector vector drawn in flow direction. |
| `hierarchical_expansion` | FOCUS + STAGGER | zoom + wipe | Core root anchors $\to$ child modules cascade. |
| `data_flow` | MOVE | motion_path | Signal packet travels through architecture pipeline. |
| `cycle_flow` | SEQUENCE + DRAW | appear + wipe | Continuous feedback loop between circular stages. |
Python API Usage
from pmg.motion_planners import DiagramMotionPlanner
planner = DiagramMotionPlanner()
intents = planner.plan_motion({
"pattern": "process_flow",
"steps": [
{"id": "ingest_node", "label": "Data Ingestion"},
{"id": "process_node", "label": "Neural Processing"},
{"id": "output_node", "label": "Decision Output"},
],
"connectors": [
{"id": "conn_1_2", "source": "ingest_node", "direction": "right"},
{"id": "conn_2_3", "source": "process_node", "direction": "right"},
]
})
for intent in intents:
print(f"[{intent.target_id}] -> {intent.primitive.value} ({intent.semantic_purpose})")
