Document pipeline checks and ASL dependencies

.gitignore

@@ -0,0 +1,4 @@
+__pycache__/
+*.py[cod]
+.pytest_cache/
+data/examples/*.mp4

README.md

@@ -11,7 +11,52 @@ pinned: false
 
 # ASL TTS Test
 
-MVP Gradio pour tester la brique TTS, llama.cpp et une interface custom avant de brancher le pipeline vidéo ASL.
+MVP Gradio pour tester la pipeline ASL video/camera -> intent JSON -> llama.cpp -> Qwen3-TTS.
+
+## Pipeline
+
+```text
+Video upload / camera capture
+-> ASL frame sampling
+-> MediaPipe landmarks, if installed
+-> TFLite ASL classifier, if data/models/asl/model.tflite is present
+-> DeepFace emotion aggregation, if installed
+-> llama.cpp subtitle + voice instruction
+-> Qwen3-TTS audio
+```
+
+The app is organized so each brick can fail independently with diagnostics instead of blocking
+the whole interface at startup.
+
+## Local checks
+
+Run unit tests:
+
+```bash
+python3 -m pytest -q
+```
+
+Run each brick directly:
+
+```bash
+python3 scripts/test_asl_brick.py
+python3 scripts/test_llm_brick.py
+python3 scripts/test_tts_brick.py
+python3 scripts/test_full_pipeline.py
+```
+
+`scripts/test_asl_brick.py` creates a tiny temporary demo clip when no video path is supplied.
+
+## ASL model files
+
+Place the ASL classifier assets here when available:
+
+```text
+data/models/asl/model.tflite
+data/models/asl/train.csv
+```
+
+Without these files, the ASL brick still samples frames and emits `model_missing` diagnostics.
 
 ## GPU dependencies
 
@@ -22,3 +67,12 @@ GPU runtime, install the GPU dependency set instead:
 ```bash
 pip install -r requirements-gpu.txt --no-build-isolation
 ```
+
+## Full ASL dependencies
+
+The default requirements keep the Space build lighter. For a full local ASL/emotion runtime with
+MediaPipe landmarks and DeepFace emotion detection, install:
+
+```bash
+pip install -r requirements-asl-full.txt
+```

pytest.ini

@@ -0,0 +1,4 @@
+[pytest]
+testpaths = tests
+asyncio_default_fixture_loop_scope = function
+

requirements-asl-full.txt

@@ -0,0 +1,7 @@
+-r requirements.txt
+
+protobuf>=4.25,<5.0
+mediapipe==0.10.14
+tensorflow-cpu==2.17.1
+deepface==0.0.100
+tf-keras==2.17.0

requirements.txt

@@ -7,3 +7,7 @@ soundfile
 torch
 huggingface-hub
 llama-cpp-python
+numpy>=1.26,<2.0
+pandas>=2.1,<3.0
+opencv-python-headless==4.11.0.86
+pytest>=8.0,<9.0