Updated Test Structure (markdown)

Test Structure.md

@@ -1,3 +0,0 @@
-# Test Structure
-
-_TODO: Add content here._

Test-Structure.md

@@ -0,0 +1,55 @@
+## 🧪 Test Structure
+
+This page documents how the LCS Pipeline is tested using local data, stubbed uploads, and simulated metadata inputs.
+
+---
+
+### 🗂️ Test Data Layout
+
+Test datasets mimic the live session structure:
+
+```
+/test-data/
+└── 2025.07.20/
+    ├── hits/
+    │   └── clutch-kill.mp4
+    │   └── clutch-kill.notes.json
+    ├── outtakes/
+    │   └── voice-marked-clip.mp4
+    ├── montages/
+    │   └── combined-moments.mp4
+    └── rendered/
+```
+
+---
+
+### 🧪 Test Types
+
+| Type                  | Description |
+|-----------------------|-------------|
+| Manual test runs      | Dev triggers pipeline using `main.py` with selected session folder |
+| CLI test harness      | Future unit-test scaffolding using `pytest` or similar |
+| Metadata validation   | JSON schema checks against notes files |
+| Upload dry-runs       | Bypasses YouTube/PeerTube API for local test |
+| Canvas syncing        | Ensures canvas content matches pushed page |
+
+---
+
+### 🧰 Local Testing Tools
+
+- `main.py --debug` → dry-run mode
+- `simulate_metadata()` → pre-fills mock `notes.json`
+- `render_engine.py` → can be run standalone on a single clip
+- `thumbnail_utils.py` → used for visual validation of thumb output
+
+---
+
+### 🧠 Notes
+
+- Real clips should never be modified in-place
+- Stubbed upload responses can be faked using mock response JSON files
+- You can disable cleanup to inspect rendered outputs
+
+---
+
+Next page: [`Thumbnail Prompting`](./Thumbnail-Prompting)