fix: resolve ONNX CUDA fallback, Docker e2e infrastructure, and all test failures

- Add safe_onnx_session() to gpu.py with graceful CUDA EP → CPU fallback
- Replace bare ort.InferenceSession() calls across colorize, restore, inpaint, remove_bg
- Add libcublas-12-6 to production Dockerfile for ONNX Runtime CUDA EP
- Add skipIfFeatureNotInstalled guards to remove-bg, blur-faces, smart-crop, ocr, noise-removal e2e specs
- Add AI tool install prompt detection in tools-all.spec.ts
- Add smart-crop to PYTHON_SIDECAR_TOOLS so frontend shows install prompt correctly
- Create Dockerfile.test.dockerignore to include tests/ in test image builds
- Add libheif-examples and exiftool to Dockerfile.test for HEIC and metadata tests
- Regenerate visual regression baselines for Docker/Linux and skip on non-Docker platforms

packages/ai/python/remove_bg.py

@@ -76,7 +76,16 @@ def main():
 
         emit_progress(10, "Loading model")
 
-        session = new_session(model, providers=onnx_providers())
+        providers = onnx_providers()
+        try:
+            session = new_session(model, providers=providers)
+        except Exception as e:
+            if "CUDAExecutionProvider" in providers:
+                print(f"[remove-bg] GPU session failed ({e}), falling back to CPU",
+                      file=sys.stderr, flush=True)
+                session = new_session(model, providers=["CPUExecutionProvider"])
+            else:
+                raise
 
         emit_progress(25, "Model loaded")