Merge pull request #1 from connerkward/claude/review-documentation-5ki2t

Reorganize project structure and improve documentation

Author: connerkward

PyBas3/TECHNICAL_REFERENCE.md

@@ -51,29 +51,42 @@ Modular system for interactive projection mapping with multi-person pose trackin
 PyBas3/
 ├── TECHNICAL_REFERENCE.md          # This file
 ├── requirements.txt
+├── orchestrator.py                 # Main entry point for running full system
+├── participants_db.json            # UUID persistence (atomic writes)
+├── zone_config.json                # Master zone configuration
 ├── agents/                         # Coordination docs
 │   ├── SESSION.md                  # Quick reference
-│   └── ...
+│   └── AGENT_0_SHARED.md           # Shared protocols
 ├── common/                         # Shared protocols & constants
 │   ├── __init__.py
 │   ├── protocols.py                # Data structures & constants
 │   └── shared_memory.py            # Binary protocol encoding/decoding
 ├── mediapipe/                      # Process 1
 │   ├── multi_person_detector.py
+│   ├── participant_tracker.py      # pHash UUID assignment
 │   ├── ndi_streamer.py
 │   ├── shared_memory_writer.py
-│   ├── zone_config.json
-│   └── participants_db.json
+│   ├── live_dashboard.py           # Real-time monitoring
+│   └── zone_config.json            # Runtime zone config
 ├── scoring/                        # Process 2
 │   ├── pose_scorer.py
 │   ├── reference_builder.py
 │   ├── shared_memory_reader.py
 │   └── output/participant_<uuid>_score.json
-├── pre_render/                     # Offline pipeline (migrated)
-│   ├── depth_blend_video.py
-│   └── frames_to_video.py
-└── td_scripts/                     # TouchDesigner (migrated)
-    └── ConnerTD/ConnerTD.toe
+├── pre_render/                     # Offline pipeline
+│   ├── depth_blend_video.py        # Main processing script
+│   ├── frames_to_video.py
+│   ├── generate_chronophoto_variations.py
+│   ├── pose_skeleton_render.py
+│   └── CHRONOPHOTO_CONTEXT.md
+├── td_scripts/                     # TouchDesigner integration
+│   ├── td_integration.py           # Main TD API
+│   ├── td_participant_manager.py
+│   ├── ndi_discovery.py
+│   ├── score_watcher.py
+│   └── ConnerTD/ConnerTD.toe
+└── tests/
+    └── test_integration.py
 ```
 
 ---

PyBas3/pre_render/CHRONOPHOTO_CONTEXT.md

@@ -37,7 +37,7 @@ Three blend modes are generated in parallel:
 
 1. **Initial approach**: Created chronophoto variations script with different frame counts (3, 5, 7, 10) and selection strategies (sequential, most different, high variance)
 2. **Depth map composites**: Initially tried compositing depth maps onto raw frames, but user wanted raw frame composites only
-3. **Pipeline integration**: Added as separate pass that runs after prerequisite passmes (depth maps, frame loading)
+3. **Pipeline integration**: Added as separate pass that runs after prerequisite passes (depth maps, frame loading)
 4. **Parallelization**: Process blend modes in parallel for efficiency
 5. **Full frame usage**: For long_exposure mode, use ALL frames to get continuous trail effect (not just subset)
 

openmemory.md

@@ -10,9 +10,6 @@
 - `PyBas3/scoring` reads shared memory poses and writes per-participant score JSON files.
 - `PyBas3/td_scripts` consumes NDI + score JSONs for TouchDesigner integration.
 
-## User Defined Namespaces
-- 
-
 ## Components
 - `MultiPersonDetector` (`PyBas3/mediapipe/multi_person_detector.py`): orchestrates MediaPipe detection, zone filtering, UUID assignment, shared memory writes, and NDI streaming.
 - `ParticipantTracker` (`PyBas3/mediapipe/participant_tracker.py`): computes face-based pHash, matches/creates UUIDs, persists to `participants_db.json`.