Upload 278 files

.gitattributes

@@ -142,3 +142,8 @@ frontend/src/assets/imagens-default-room/terrace/terrace2.png filter=lfs diff=lf
 frontend/src/assets/imagens-default-room/terrace/terrace3.png filter=lfs diff=lfs merge=lfs -text
 frontend/src/assets/imagens-default-room/terrace/terrace4.png filter=lfs diff=lfs merge=lfs -text
 frontend/src/assets/imagens-default-room/terrace/terrace5.png filter=lfs diff=lfs merge=lfs -text
+backend/uploads/demo-room_edit_ef2b9c2c_edit_32b3c807_edit_e6f30d22.jpg filter=lfs diff=lfs merge=lfs -text
+backend/uploads/demo-room_edit_ef2b9c2c_edit_32b3c807.jpg filter=lfs diff=lfs merge=lfs -text
+backend/uploads/demo-room_edit_ef2b9c2c.jpg filter=lfs diff=lfs merge=lfs -text
+backend/uploads/demo-room.jpg filter=lfs diff=lfs merge=lfs -text
+backend/uploads/proyecto-White-Houses_edit_d4a831c5_edit_c5a95477.jpg filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,131 +1,180 @@
----
-title: Hyper Reality Room Visualizer
-emoji: 🏠
-colorFrom: blue
-colorTo: indigo
-sdk: docker
-app_port: 7860
-pinned: false
----
-
-# PoC SaaS Iframe
-
-Este repositorio es una prueba de concepto de un SaaS que ofrece un visualizador embebido mediante iframe.
-
-## Visión general
-
-El proyecto está dividido en dos partes principales:
-
-- `backend/` - servidor en Python con FastAPI que expone APIs, sirve el build de React en producción y controla la seguridad.
-- `frontend/` - aplicación React creada con Vite que funciona como visualizador embebido.
-
-Además existen carpetas de ejemplos de cliente:
-
-- `cliente1/`, `cliente2/`, `cliente3/`
-
-Estas páginas muestran cómo integrar el widget del SaaS desde un host externo.
-
-## Documentación del proyecto
-
-- Documentación general: este archivo `README.md`
-- Documentación del backend: `backend/README.md`
-
-## Arquitectura
-
-1. El backend maneja:
-   - `POST /api/token` para generar tokens temporales
-   - `GET /config` para devolver datos del cliente
-   - `POST /session/start` para registrar sesiones activas
-   - `/app` para servir el frontend construido en producción
-   - `/widget.js` para el widget que inyecta el iframe en clientes externos
-
-2. El frontend React se ejecuta en dos modos:
-   - desarrollo (`npm run dev` en `frontend/`) para cambios rápidos y hot reload
-   - producción (`npm run build`) produce `frontend/dist`, que el backend sirve como estático
-
-3. El widget cliente solicita un token al backend y luego inyecta un iframe con `/app?token=...`.
-
-## Flujo de desarrollo
-
-Este proyecto tiene dos flujos claros de desarrollo:
-
-1. Desarrollo frontend con Vite
-2. Desarrollo backend con FastAPI
-
-Ambos pueden correr al mismo tiempo para que los devs trabajen en el app y en la API al mismo tiempo.
-
-### Flujo recomendado para desarrollo local
-
-1. Abrir una terminal para el backend:
-
-```cmd
-cd /d C:\Users\alane\OneDrive\Escritorio\Prueba-PoC\backend
-.venv\Scripts\activate
-pip install -r requirements.txt
-python -m uvicorn main:app --reload --host 0.0.0.0 --port 8000
-```
-
-2. Abrir otra terminal para el frontend:
-
-```cmd
-cd /d C:\Users\alane\OneDrive\Escritorio\Prueba-PoC\frontend
-npm install
-npm run dev
-```
-
-3. Cargar el app React en el navegador para ver hot reload:
-
-- `http://localhost:5173/app?token=...`
-
-4. Si necesitas probar la integración completa con el backend y los endpoints, usa también:
-
-- `http://localhost:8000/preview`
-- `http://localhost:8000/admin`
-
-### Qué significa este flujo
-
-- `localhost:5173` sirve el frontend en modo desarrollo con hot reload.
-- `localhost:8000` sirve el backend y, en producción, también el build estático de React.
-- Durante el desarrollo, no necesitas ejecutar `npm run build` cada vez que cambias `frontend/src`.
-- El build solo es necesario cuando quieres probar la app servida desde el backend (`/app`) o cuando subes a producción.
-
-### Desarrollo incremental
-
-- Cambia archivos en `frontend/src` y Vite recargará el navegador automáticamente.
-- Cambia `backend/main.py` y Uvicorn recargará el backend automáticamente.
-- Si quieres probar una ruta de frontend específica desde el backend, primero construye y luego carga `http://localhost:8000/app?token=...`.
-
-### Integración en producción
-
-Cuando ya estés listo para producción o para una prueba final:
-
-```cmd
-cd /d C:\Users\alane\OneDrive\Escritorio\Prueba-PoC\frontend
-npm run build
-```
-
-Luego abre:
-
-- `http://localhost:8000/app?token=...`
-
-En este modo, el backend sirve los archivos estáticos desde `frontend/dist`.
-
-## Comandos útiles
-
-- `npm run dev` - desarrollo React con Vite
-- `npm run build` - generar `frontend/dist`
-- `python -m uvicorn main:app --reload --host 0.0.0.0 --port 8000` - arrancar backend
-- `run_server.bat` - helper Windows para iniciar el backend con `.venv`
-
-## Estructura de carpetas
-
-- `backend/` - servidor FastAPI, documentos y páginas auxiliares
-- `frontend/` - app React + Vite
-- `frontend/dist/` - build de producción
-- `cliente1/`, `cliente2/`, `cliente3/` - ejemplos de integración
-
-## Importante
-
-- En desarrollo, usa `localhost:5173` para ver los cambios de frontend al instante.
-- En producción, el backend sirve todos los assets desde `frontend/dist` en `/app`.
-- `backend/README.md` contiene la documentación técnica detallada del backend.
+---
+title: Hyper Reality Room Visualizer
+emoji: 🏠
+colorFrom: blue
+colorTo: indigo
+sdk: docker
+app_port: 7860
+pinned: false
+---
+
+# Hyper Reality Room Visualizer — PoC SaaS
+
+Visualizador de habitaciones embebido como SaaS mediante iframe. Los clientes reciben una API key y un snippet HTML; el backend genera un token temporal y carga el visualizador en un iframe.
+
+## Estructura del repositorio
+
+```
+Prueba-PoC/
+├── backend/            # FastAPI — API, autenticación, SAM2, texturas
+│   ├── main.py
+│   ├── routers/
+│   ├── services/
+│   ├── requirements.txt
+│   ├── .env            # variables locales (no se sube a git)
+│   └── Dockerfile      # usado solo por docker-compose (desarrollo local)
+├── frontend/           # React + Vite — visualizador embebido
+│   ├── src/
+│   ├── package.json
+│   └── dist/           # generado por npm run build (ignorado en git)
+├── Dockerfile          # multi-stage para HuggingFace Spaces
+├── docker-compose.yml  # entorno local con Docker
+└── README.md
+```
+
+## Arquitectura
+
+```
+Cliente externo (HTML del cliente)
+  └─► widget.js          — se auto-descarga desde el servidor
+        └─► POST /api/token   — obtiene token temporal
+              └─► <iframe src="/app?token=...">
+                    └─► FastAPI sirve index.html del SPA React
+                          └─► SPA carga assets desde /assets/...
+```
+
+El backend expone:
+
+| Endpoint | Descripción |
+|---|---|
+| `POST /api/token` | Genera token temporal (TTL 5 min) |
+| `GET /config` | Devuelve datos del cliente validando token |
+| `POST /session/start` | Registra sesión activa |
+| `GET /widget.js` | Script que el cliente externo incluye en su HTML |
+| `GET /admin` | Panel de gestión de API keys |
+| `GET /app` | Punto de entrada del SPA React |
+| `POST /api/generate-key` | Crea nueva API key (persiste en MongoDB) |
+| `DELETE /api/keys/{id}` | Elimina API key |
+
+## Variables de entorno
+
+Crea el archivo `backend/.env` (ya está en `.gitignore`):
+
+```env
+MONGODB_URI=mongodb+srv://<usuario>:<password>@<cluster>.mongodb.net/
+```
+
+En **HuggingFace Spaces**: Settings → Repository secrets → `MONGODB_URI`.
+
+## Desarrollo local
+
+Hay dos opciones: con o sin Docker.
+
+### Opción A — Sin Docker (recomendado para desarrollo activo)
+
+Levanta backend y frontend en terminales separadas para obtener hot reload en ambos.
+
+**Terminal 1 — Backend:**
+
+```bash
+cd backend
+python -m venv .venv
+.venv\Scripts\activate          # Windows
+# source .venv/bin/activate     # Mac / Linux
+pip install -r requirements.txt
+python -m uvicorn main:app --reload --host 0.0.0.0 --port 8000
+```
+
+**Terminal 2 — Frontend:**
+
+```bash
+cd frontend
+npm install
+npm run dev
+```
+
+**URLs disponibles:**
+
+| URL | Qué muestra |
+|---|---|
+| `http://localhost:5173/app` | SPA React con hot reload (Vite dev server) |
+| `http://localhost:8000/admin` | Panel de API keys |
+| `http://localhost:8000/preview` | Vista previa del iframe |
+
+El frontend en `localhost:5173` proxea automáticamente `/api`, `/seg`, `/uploads` y `/widget.js` hacia `localhost:8000` (configurado en `vite.config.ts`).
+
+### Opción B — Con Docker Compose
+
+Levanta solo el backend en contenedor. Útil para probar la integración completa.
+
+```bash
+docker compose up --build
+```
+
+Accede en `http://localhost:8000`. El frontend se sirve desde `frontend/dist` montado como volumen — necesitas haber ejecutado `npm run build` antes.
+
+## Integración completa (backend sirve el frontend)
+
+Cuando quieres probar exactamente como quedaría en producción:
+
+```bash
+# 1. Construir el frontend
+cd frontend
+npm run build
+
+# 2. Levantar el backend (sirve frontend/dist en /app)
+cd ../backend
+python -m uvicorn main:app --reload --host 0.0.0.0 --port 8000
+```
+
+Abre `http://localhost:8000/app` — el backend sirve los assets estáticos desde `frontend/dist`.
+
+## Despliegue en HuggingFace Spaces
+
+El `Dockerfile` de la raíz es el que usa HuggingFace. Hace dos cosas:
+
+1. **Stage 1 (Node):** ejecuta `npm run build` y produce `frontend/dist`
+2. **Stage 2 (Python):** instala dependencias del backend y copia el dist del stage 1
+
+No necesitas subir `frontend/dist` al repositorio — Docker lo genera automáticamente en cada push.
+
+```bash
+git add .
+git commit -m "tu mensaje"
+git push
+```
+
+HuggingFace detecta el push y reconstruye el contenedor.
+
+**URL del Space:** `https://huggingface.co/spaces/eduardo4547/hyper-reality-visualizer`
+
+## Flujo del widget para clientes externos
+
+El cliente externo pega este snippet en su HTML (se obtiene desde `/admin`):
+
+```html
+<div id="contenedor-saas" data-client-id="CLIENTE_XXXXXXXX"></div>
+<script src="https://<tu-space>.hf.space/widget.js"></script>
+```
+
+El script:
+1. Lee `data-client-id` del div
+2. Llama `POST /api/token` con ese ID
+3. Inyecta un `<iframe>` apuntando a `/app?token=<token>`
+
+## Comandos útiles
+
+```bash
+# Backend
+python -m uvicorn main:app --reload --host 0.0.0.0 --port 8000
+
+# Frontend — desarrollo
+npm run dev
+
+# Frontend — producción
+npm run build
+
+# Docker local
+docker compose up --build
+docker compose down
+```

backend/.env

@@ -1 +1,5 @@
 MONGODB_URI=mongodb+srv://alaneduardodelacruz407_db_user:YGUWdpXqUiGH104Q@naufar-cluster.n9htwoa.mongodb.net/
+# Para producción (HuggingFace):
+GRADIO_SPACE_URL=https://eduardo4547-hyper-reality-sam2-gpu.hf.space
+# Para desarrollo local (Gradio corriendo en puerto 7860):
+# GRADIO_SPACE_URL=http://localhost:7860

backend/core/config.py

@@ -51,6 +51,11 @@ SAM2_MODEL_DIR_CANDIDATES = ("models", "modelo")
 SAM2_UNLOAD_AFTER_USE = str(os.getenv("SAM2_UNLOAD_AFTER_USE", "0")).strip().lower() in {"1", "true", "yes"}
 FRONTEND_DEBUG = str(os.getenv("FRONTEND_DEBUG", "0")).strip().lower() in {"1", "true", "yes", "on"}
 
+# URL del Space de Gradio que procesa las imágenes con GPU.
+# Local:      http://localhost:7860
+# Producción: https://<tu-space>.hf.space
+GRADIO_SPACE_URL: str = os.getenv("GRADIO_SPACE_URL", "").rstrip("/")
+
 MAX_UPLOAD_WIDTH = 1024
 UPLOAD_JPEG_QUALITY = 82
 SD_JOB_STALE_SECONDS = 120

backend/logs/app.log

@@ -0,0 +1,213 @@
+2026-04-29 20:12:40,375 INFO backend.segmentation: [LIFESPAN] GRADIO_SPACE_URL set — skipping local SAM2 load.
+2026-04-29 20:13:22,434 INFO pymongo.serverSelection: {"message": "Waiting for suitable server to become available", "selector": "Primary()", "operation": "find", "topologyDescription": "<TopologyDescription id: 69f2bac20e0544b45d45e6f1, topology_type: ReplicaSetNoPrimary, servers: [<ServerDescription ('ac-tynjizz-shard-00-00.n9htwoa.mongodb.net', 27017) server_type: Unknown, rtt: None>, <ServerDescription ('ac-tynjizz-shard-00-01.n9htwoa.mongodb.net', 27017) server_type: Unknown, rtt: None>, <ServerDescription ('ac-tynjizz-shard-00-02.n9htwoa.mongodb.net', 27017) server_type: Unknown, rtt: None>]>", "clientId": {"$oid": "69f2bac20e0544b45d45e6f1"}, "remainingTimeMS": 30}
+2026-04-29 20:13:53,491 INFO backend.segmentation: [UPLOAD_BG] START at 2026-04-30T02:13:53.491443+00:00
+2026-04-29 20:13:53,492 INFO backend.segmentation: [JOB 667a8562af1044a6926d53449b625dd6] preparing_image progress=12
+2026-04-29 20:13:53,743 INFO backend.segmentation: [JOB 667a8562af1044a6926d53449b625dd6] segmenting_with_sam2 progress=30 estimated_seconds=51.939840000000004
+2026-04-29 20:13:56,153 INFO httpx: HTTP Request: GET https://eduardo4547-hyper-reality-sam2-gpu.hf.space/config "HTTP/1.1 200 OK"
+2026-04-29 20:13:56,625 INFO httpx: HTTP Request: GET https://eduardo4547-hyper-reality-sam2-gpu.hf.space/gradio_api/info?serialize=False "HTTP/1.1 200 OK"
+2026-04-29 20:13:56,675 WARNING services.gradio_client_service: Gradio Space call failed, falling back to local SAM2: Cannot find a function with `api_name`: /segment.
+2026-04-29 20:13:56,675 INFO backend.segmentation: [SAM2_GENERATE] START at 2026-04-30T02:13:56.675848+00:00
+2026-04-29 20:13:56,687 INFO backend.segmentation: [SAM2] Mask generator not loaded; attempting on-demand load.
+2026-04-29 20:13:56,687 WARNING backend.segmentation: [SAM2] torch not installed — SAM2 unavailable (using Gradio Space)
+2026-04-29 20:13:56,687 ERROR backend.segmentation: [JOB 667a8562af1044a6926d53449b625dd6] failed: 500: torch not installed — SAM2 unavailable (using Gradio Space)
+Traceback (most recent call last):
+  File "C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\services\image_service.py", line 145, in run_upload_job
+    label_map, mask_count = generate_label_map(image_rgb)
+                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\services\scene_service.py", line 134, in generate_label_map
+    generator = get_sam2_mask_generator()
+                ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\services\sam2_service.py", line 184, in get_sam2_mask_generator
+    raise HTTPException(status_code=500, detail=detail)
+fastapi.exceptions.HTTPException: 500: torch not installed — SAM2 unavailable (using Gradio Space)
+2026-04-29 20:13:56,689 INFO backend.segmentation: [UPLOAD_BG] DONE 3.196s at 2026-04-30T02:13:56.689353+00:00
+2026-04-29 20:13:56,689 INFO backend.segmentation: Releasing resources (full_unload=False)
+2026-04-29 20:13:57,044 INFO httpx: HTTP Request: HEAD https://huggingface.co/api/telemetry/py_client/initiated "HTTP/1.1 200 OK"
+2026-04-29 20:13:57,163 INFO httpx: HTTP Request: GET https://eduardo4547-hyper-reality-sam2-gpu.hf.space/gradio_api/heartbeat/68121945-8a97-442e-8cbb-59cd817f9df3 "HTTP/1.1 200 OK"
+2026-04-29 20:16:38,951 INFO backend.segmentation: [LIFESPAN] GRADIO_SPACE_URL set — skipping local SAM2 load.
+2026-04-29 20:16:55,032 INFO backend.segmentation: [UPLOAD_BG] START at 2026-04-30T02:16:55.032702+00:00
+2026-04-29 20:16:55,032 INFO backend.segmentation: [JOB 50d2907ecb844a3fafc2f8a5247620ea] preparing_image progress=12
+2026-04-29 20:16:55,089 INFO backend.segmentation: [JOB 50d2907ecb844a3fafc2f8a5247620ea] segmenting_with_sam2 progress=30 estimated_seconds=51.939840000000004
+2026-04-29 20:16:57,866 INFO httpx: HTTP Request: GET http://localhost:7860/config "HTTP/1.1 200 OK"
+2026-04-29 20:17:00,098 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/info?serialize=False "HTTP/1.1 200 OK"
+2026-04-29 20:17:00,613 INFO httpx: HTTP Request: HEAD https://huggingface.co/api/telemetry/py_client/initiated "HTTP/1.1 200 OK"
+2026-04-29 20:17:02,540 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/heartbeat/06d3302a-135c-423c-8294-e950bff0a17a "HTTP/1.1 200 OK"
+2026-04-29 20:17:02,542 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/upload "HTTP/1.1 200 OK"
+2026-04-29 20:17:04,774 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/queue/join "HTTP/1.1 200 OK"
+2026-04-29 20:17:07,079 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/queue/data?session_hash=06d3302a-135c-423c-8294-e950bff0a17a "HTTP/1.1 200 OK"
+2026-04-29 20:17:26,593 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/file=C:\Users\alane\AppData\Local\Temp\gradio\b17b68f6472ad2c6c0bb7499f231e242c50b344237d319bd3b45d56ef2d488aa\image.webp "HTTP/1.1 200 OK"
+2026-04-29 20:17:26,607 INFO services.gradio_client_service: Gradio Space segmentation: entorno=🌳 Terraza / Patio / Jardín motor=Híbrido Arquitectura (Cityscapes Grande + DINO Pequeño) mask_count=20
+2026-04-29 20:17:26,607 INFO backend.segmentation: [JOB 50d2907ecb844a3fafc2f8a5247620ea] saving_masks progress=92
+2026-04-29 20:17:26,690 INFO backend.segmentation: [JOB 50d2907ecb844a3fafc2f8a5247620ea] segments_meta saved (20 segments)
+2026-04-29 20:17:26,690 INFO backend.segmentation: [JOB 50d2907ecb844a3fafc2f8a5247620ea] done mask_count=20
+2026-04-29 20:17:26,691 INFO backend.segmentation: [UPLOAD_BG] DONE 31.658s at 2026-04-30T02:17:26.691033+00:00
+2026-04-29 20:17:26,691 INFO backend.segmentation: Releasing resources (full_unload=False)
+2026-04-29 20:17:45,867 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:17:45.867306+00:00
+2026-04-29 20:17:46,402 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\proyecto-White-Houses.jpg
+2026-04-29 20:17:46,702 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.835s at 2026-04-30T02:17:46.702156+00:00
+2026-04-29 20:17:50,496 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:17:50.496976+00:00
+2026-04-29 20:17:50,819 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\proyecto-White-Houses.jpg
+2026-04-29 20:17:50,985 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.488s at 2026-04-30T02:17:50.985122+00:00
+2026-04-29 20:17:54,307 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:17:54.307539+00:00
+2026-04-29 20:17:54,768 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\proyecto-White-Houses.jpg
+2026-04-29 20:17:54,931 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.624s at 2026-04-30T02:17:54.931489+00:00
+2026-04-29 20:17:56,247 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:17:56.247254+00:00
+2026-04-29 20:17:56,728 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\proyecto-White-Houses.jpg
+2026-04-29 20:17:56,898 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.651s at 2026-04-30T02:17:56.898353+00:00
+2026-04-29 20:18:00,146 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:18:00.146989+00:00
+2026-04-29 20:18:00,616 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\proyecto-White-Houses.jpg
+2026-04-29 20:18:00,782 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.636s at 2026-04-30T02:18:00.782444+00:00
+2026-04-29 20:18:05,560 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:18:05.560412+00:00
+2026-04-29 20:18:06,116 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\proyecto-White-Houses.jpg
+2026-04-29 20:18:06,272 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.712s at 2026-04-30T02:18:06.272748+00:00
+2026-04-29 20:18:11,868 INFO backend.segmentation: [UPLOAD_BG] START at 2026-04-30T02:18:11.868080+00:00
+2026-04-29 20:18:11,868 INFO backend.segmentation: [JOB 6897b1b65f7c4201ab0a5c35bf18b368] preparing_image progress=12
+2026-04-29 20:18:11,941 INFO backend.segmentation: [JOB 6897b1b65f7c4201ab0a5c35bf18b368] segmenting_with_sam2 progress=30 estimated_seconds=49.28768
+2026-04-29 20:18:14,341 INFO httpx: HTTP Request: GET http://localhost:7860/config "HTTP/1.1 200 OK"
+2026-04-29 20:18:16,543 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/info?serialize=False "HTTP/1.1 200 OK"
+2026-04-29 20:18:16,695 INFO httpx: HTTP Request: HEAD https://huggingface.co/api/telemetry/py_client/initiated "HTTP/1.1 200 OK"
+2026-04-29 20:18:18,825 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/heartbeat/e332d34b-376c-4071-9631-656973e262a0 "HTTP/1.1 200 OK"
+2026-04-29 20:18:18,828 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/upload "HTTP/1.1 200 OK"
+2026-04-29 20:18:21,032 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/queue/join "HTTP/1.1 200 OK"
+2026-04-29 20:18:23,304 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/queue/data?session_hash=e332d34b-376c-4071-9631-656973e262a0 "HTTP/1.1 200 OK"
+2026-04-29 20:18:36,010 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/file=C:\Users\alane\AppData\Local\Temp\gradio\7b5a7a453887652d3284bb9eabe628e95982e15d5e0f16ff718d22beb5ed2a72\image.webp "HTTP/1.1 200 OK"
+2026-04-29 20:18:36,020 INFO services.gradio_client_service: Gradio Space segmentation: entorno=🛁 Baño / Cocina motor=SegFormer (SegFormer ADE20K+ DINO) + SAM 2.1 mask_count=18
+2026-04-29 20:18:36,021 INFO backend.segmentation: [JOB 6897b1b65f7c4201ab0a5c35bf18b368] saving_masks progress=92
+2026-04-29 20:18:36,063 INFO backend.segmentation: [JOB 6897b1b65f7c4201ab0a5c35bf18b368] segments_meta saved (14 segments)
+2026-04-29 20:18:36,063 INFO backend.segmentation: [JOB 6897b1b65f7c4201ab0a5c35bf18b368] done mask_count=18
+2026-04-29 20:18:36,063 INFO backend.segmentation: [UPLOAD_BG] DONE 24.195s at 2026-04-30T02:18:36.063054+00:00
+2026-04-29 20:18:36,063 INFO backend.segmentation: Releasing resources (full_unload=False)
+2026-04-29 20:18:44,885 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:18:44.885155+00:00
+2026-04-29 20:18:45,431 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:18:45,591 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.706s at 2026-04-30T02:18:45.591016+00:00
+2026-04-29 20:18:49,898 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:18:49.898853+00:00
+2026-04-29 20:18:50,427 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:18:50,573 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.674s at 2026-04-30T02:18:50.573101+00:00
+2026-04-29 20:18:52,643 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:18:52.643494+00:00
+2026-04-29 20:18:53,101 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:18:53,259 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.615s at 2026-04-30T02:18:53.259095+00:00
+2026-04-29 20:18:55,884 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:18:55.884791+00:00
+2026-04-29 20:18:56,336 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:18:56,483 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.599s at 2026-04-30T02:18:56.483159+00:00
+2026-04-29 20:18:58,689 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:18:58.689702+00:00
+2026-04-29 20:18:59,139 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:18:59,285 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.596s at 2026-04-30T02:18:59.285688+00:00
+2026-04-29 20:19:01,587 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:19:01.587113+00:00
+2026-04-29 20:19:02,057 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:19:02,202 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.616s at 2026-04-30T02:19:02.202912+00:00
+2026-04-29 20:19:08,881 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:19:08.881813+00:00
+2026-04-29 20:19:09,450 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:19:09,598 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.716s at 2026-04-30T02:19:09.598660+00:00
+2026-04-29 20:19:14,081 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:19:14.081047+00:00
+2026-04-29 20:19:14,613 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:19:14,765 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.684s at 2026-04-30T02:19:14.765513+00:00
+2026-04-29 20:19:20,588 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:19:20.588691+00:00
+2026-04-29 20:19:21,001 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:19:21,148 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.560s at 2026-04-30T02:19:21.148963+00:00
+2026-04-29 20:19:26,717 INFO backend.segmentation: [UPLOAD_BG] START at 2026-04-30T02:19:26.717015+00:00
+2026-04-29 20:19:26,717 INFO backend.segmentation: [JOB 0e77f0bbab294a828b0dbfbb826c3805] preparing_image progress=12
+2026-04-29 20:19:26,784 INFO backend.segmentation: [JOB 0e77f0bbab294a828b0dbfbb826c3805] segmenting_with_sam2 progress=30 estimated_seconds=49.28768
+2026-04-29 20:19:29,222 INFO httpx: HTTP Request: GET http://localhost:7860/config "HTTP/1.1 200 OK"
+2026-04-29 20:19:31,597 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/info?serialize=False "HTTP/1.1 200 OK"
+2026-04-29 20:19:31,734 INFO httpx: HTTP Request: HEAD https://huggingface.co/api/telemetry/py_client/initiated "HTTP/1.1 200 OK"
+2026-04-29 20:19:34,073 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/heartbeat/19c5143b-64ec-4b5d-942b-098f7096a07d "HTTP/1.1 200 OK"
+2026-04-29 20:19:34,076 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/upload "HTTP/1.1 200 OK"
+2026-04-29 20:19:36,338 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/queue/join "HTTP/1.1 200 OK"
+2026-04-29 20:19:38,618 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/queue/data?session_hash=19c5143b-64ec-4b5d-942b-098f7096a07d "HTTP/1.1 200 OK"
+2026-04-29 20:19:50,817 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/file=C:\Users\alane\AppData\Local\Temp\gradio\627367a96ce0d7cea77d51e3b57be3a803ddd9b06e2c8a9547129f322756dd05\image.webp "HTTP/1.1 200 OK"
+2026-04-29 20:19:50,828 INFO services.gradio_client_service: Gradio Space segmentation: entorno=🛁 Baño / Cocina motor=SegFormer (SegFormer ADE20K+ DINO) + SAM 2.1 mask_count=13
+2026-04-29 20:19:50,828 INFO backend.segmentation: [JOB 0e77f0bbab294a828b0dbfbb826c3805] saving_masks progress=92
+2026-04-29 20:19:50,860 INFO backend.segmentation: [JOB 0e77f0bbab294a828b0dbfbb826c3805] segments_meta saved (13 segments)
+2026-04-29 20:19:50,861 INFO backend.segmentation: [JOB 0e77f0bbab294a828b0dbfbb826c3805] done mask_count=13
+2026-04-29 20:19:50,861 INFO backend.segmentation: [UPLOAD_BG] DONE 24.145s at 2026-04-30T02:19:50.861597+00:00
+2026-04-29 20:19:50,861 INFO backend.segmentation: Releasing resources (full_unload=False)
+2026-04-29 20:19:58,632 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:19:58.632422+00:00
+2026-04-29 20:19:58,950 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:19:59,096 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.464s at 2026-04-30T02:19:59.096073+00:00
+2026-04-29 20:20:03,025 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:20:03.025850+00:00
+2026-04-29 20:20:03,491 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:20:03,640 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.615s at 2026-04-30T02:20:03.640105+00:00
+2026-04-29 20:20:20,501 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:20:20.501928+00:00
+2026-04-29 20:20:21,076 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:20:21,227 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.727s at 2026-04-30T02:20:21.227737+00:00
+2026-04-29 20:20:28,266 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:20:28.266563+00:00
+2026-04-29 20:20:28,768 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:20:28,935 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.669s at 2026-04-30T02:20:28.935807+00:00
+2026-04-29 20:20:35,728 INFO backend.segmentation: [UPLOAD_BG] START at 2026-04-30T02:20:35.728058+00:00
+2026-04-29 20:20:35,729 INFO backend.segmentation: [JOB f44cfa950b0546ffbcd7fa090bad49db] preparing_image progress=12
+2026-04-29 20:20:35,790 INFO backend.segmentation: [JOB f44cfa950b0546ffbcd7fa090bad49db] segmenting_with_sam2 progress=30 estimated_seconds=49.28768
+2026-04-29 20:20:38,252 INFO httpx: HTTP Request: GET http://localhost:7860/config "HTTP/1.1 200 OK"
+2026-04-29 20:20:40,610 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/info?serialize=False "HTTP/1.1 200 OK"
+2026-04-29 20:20:40,810 INFO httpx: HTTP Request: HEAD https://huggingface.co/api/telemetry/py_client/initiated "HTTP/1.1 200 OK"
+2026-04-29 20:20:43,106 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/heartbeat/31bdf8c4-3e85-4561-b055-cb540bff6222 "HTTP/1.1 200 OK"
+2026-04-29 20:20:43,110 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/upload "HTTP/1.1 200 OK"
+2026-04-29 20:20:45,435 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/queue/join "HTTP/1.1 200 OK"
+2026-04-29 20:20:47,968 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/queue/data?session_hash=31bdf8c4-3e85-4561-b055-cb540bff6222 "HTTP/1.1 200 OK"
+2026-04-29 20:21:09,199 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/file=C:\Users\alane\AppData\Local\Temp\gradio\de39b2da096587a3b0e0d7fb57541b12755192b377f4914a8047c40356bfbf4a\image.webp "HTTP/1.1 200 OK"
+2026-04-29 20:21:09,216 INFO services.gradio_client_service: Gradio Space segmentation: entorno=🛁 Baño / Cocina motor=SegFormer (SegFormer ADE20K+ DINO) + SAM 2.1 mask_count=30
+2026-04-29 20:21:09,217 INFO backend.segmentation: [JOB f44cfa950b0546ffbcd7fa090bad49db] saving_masks progress=92
+2026-04-29 20:21:09,330 INFO backend.segmentation: [JOB f44cfa950b0546ffbcd7fa090bad49db] segments_meta saved (18 segments)
+2026-04-29 20:21:09,330 INFO backend.segmentation: [JOB f44cfa950b0546ffbcd7fa090bad49db] done mask_count=30
+2026-04-29 20:21:09,331 INFO backend.segmentation: [UPLOAD_BG] DONE 33.603s at 2026-04-30T02:21:09.331315+00:00
+2026-04-29 20:21:09,331 INFO backend.segmentation: Releasing resources (full_unload=False)
+2026-04-29 20:21:19,563 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:21:19.563482+00:00
+2026-04-29 20:21:19,875 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:21:20,037 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.474s at 2026-04-30T02:21:20.037544+00:00
+2026-04-29 20:21:27,716 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:21:27.716717+00:00
+2026-04-29 20:21:28,240 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:21:28,388 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.671s at 2026-04-30T02:21:28.388007+00:00
+2026-04-29 20:21:33,770 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:21:33.770053+00:00
+2026-04-29 20:21:34,297 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:21:34,445 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.676s at 2026-04-30T02:21:34.445668+00:00
+2026-04-29 20:21:45,556 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:21:45.556315+00:00
+2026-04-29 20:21:46,063 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:21:46,214 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.658s at 2026-04-30T02:21:46.214366+00:00
+2026-04-29 20:21:54,885 INFO backend.segmentation: [UPLOAD_BG] START at 2026-04-30T02:21:54.885450+00:00
+2026-04-29 20:21:54,886 INFO backend.segmentation: [JOB ec618977b0b7478ca5d30d251fa15945] preparing_image progress=12
+2026-04-29 20:21:55,003 INFO backend.segmentation: [JOB ec618977b0b7478ca5d30d251fa15945] segmenting_with_sam2 progress=30 estimated_seconds=138.4576
+2026-04-29 20:21:57,389 INFO httpx: HTTP Request: GET http://localhost:7860/config "HTTP/1.1 200 OK"
+2026-04-29 20:21:59,757 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/info?serialize=False "HTTP/1.1 200 OK"
+2026-04-29 20:21:59,891 INFO httpx: HTTP Request: HEAD https://huggingface.co/api/telemetry/py_client/initiated "HTTP/1.1 200 OK"
+2026-04-29 20:22:02,312 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/upload "HTTP/1.1 200 OK"
+2026-04-29 20:22:02,322 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/heartbeat/3f8ed7a1-4b28-46c2-9e01-7e4358b38e76 "HTTP/1.1 200 OK"
+2026-04-29 20:22:04,682 INFO httpx: HTTP Request: POST http://localhost:7860/gradio_api/queue/join "HTTP/1.1 200 OK"
+2026-04-29 20:22:07,179 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/queue/data?session_hash=3f8ed7a1-4b28-46c2-9e01-7e4358b38e76 "HTTP/1.1 200 OK"
+2026-04-29 20:22:33,186 INFO httpx: HTTP Request: GET http://localhost:7860/gradio_api/file=C:\Users\alane\AppData\Local\Temp\gradio\80421cbcd599451a04044e3ece45f375843574c97db6199cd4e3f022f90b9311\image.webp "HTTP/1.1 200 OK"
+2026-04-29 20:22:33,213 INFO services.gradio_client_service: Gradio Space segmentation: entorno=🛁 Baño / Cocina motor=SegFormer (SegFormer ADE20K+ DINO) + SAM 2.1 mask_count=16
+2026-04-29 20:22:33,214 INFO backend.segmentation: [JOB ec618977b0b7478ca5d30d251fa15945] saving_masks progress=92
+2026-04-29 20:22:33,419 INFO backend.segmentation: [JOB ec618977b0b7478ca5d30d251fa15945] segments_meta saved (16 segments)
+2026-04-29 20:22:33,419 INFO backend.segmentation: [JOB ec618977b0b7478ca5d30d251fa15945] done mask_count=16
+2026-04-29 20:22:33,419 INFO backend.segmentation: [UPLOAD_BG] DONE 38.534s at 2026-04-30T02:22:33.419503+00:00
+2026-04-29 20:22:33,420 INFO backend.segmentation: Releasing resources (full_unload=False)
+2026-04-29 20:22:38,594 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:22:38.593824+00:00
+2026-04-29 20:22:38,977 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:22:39,402 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.808s at 2026-04-30T02:22:39.402689+00:00
+2026-04-29 20:22:44,189 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:22:44.189800+00:00
+2026-04-29 20:22:44,726 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:22:45,179 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.990s at 2026-04-30T02:22:45.179303+00:00
+2026-04-29 20:22:50,087 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:22:50.087424+00:00
+2026-04-29 20:22:50,751 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\demo-room.jpg
+2026-04-29 20:22:51,466 INFO backend.segmentation: [APPLY_TEXTURE] DONE 1.379s at 2026-04-30T02:22:51.466435+00:00
+2026-04-29 20:26:39,065 INFO backend.segmentation: [LIFESPAN] GRADIO_SPACE_URL set — skipping local SAM2 load.
+2026-04-29 20:26:39,964 INFO backend.segmentation: [LIFESPAN] GRADIO_SPACE_URL set — skipping local SAM2 load.
+2026-04-29 20:33:56,025 INFO backend.segmentation: [LIFESPAN] GRADIO_SPACE_URL set — skipping local SAM2 load.
+2026-04-29 20:34:02,828 INFO backend.segmentation: [UPLOAD_BG] START at 2026-04-30T02:34:02.828129+00:00
+2026-04-29 20:34:02,830 INFO backend.segmentation: [JOB ccecb5fbdf744d9eae57eb2b04e0af13] preparing_image progress=12
+2026-04-29 20:34:02,900 INFO backend.segmentation: [JOB ccecb5fbdf744d9eae57eb2b04e0af13] segmenting_with_sam2 progress=30 estimated_seconds=51.939840000000004
+2026-04-29 20:34:04,067 INFO httpx: HTTP Request: GET https://eduardo4547-hyper-reality-sam2-gpu.hf.space/config "HTTP/1.1 200 OK"
+2026-04-29 20:34:04,727 INFO httpx: HTTP Request: GET https://eduardo4547-hyper-reality-sam2-gpu.hf.space/gradio_api/info?serialize=False "HTTP/1.1 200 OK"
+2026-04-29 20:34:05,153 INFO httpx: HTTP Request: HEAD https://huggingface.co/api/telemetry/py_client/initiated "HTTP/1.1 200 OK"
+2026-04-29 20:34:05,258 INFO httpx: HTTP Request: GET https://eduardo4547-hyper-reality-sam2-gpu.hf.space/gradio_api/heartbeat/b306737b-bf28-4f02-8963-83b74032d449 "HTTP/1.1 200 OK"
+2026-04-29 20:34:05,354 INFO httpx: HTTP Request: POST https://eduardo4547-hyper-reality-sam2-gpu.hf.space/gradio_api/upload "HTTP/1.1 200 OK"
+2026-04-29 20:34:05,788 INFO httpx: HTTP Request: POST https://eduardo4547-hyper-reality-sam2-gpu.hf.space/gradio_api/queue/join "HTTP/1.1 200 OK"
+2026-04-29 20:34:06,344 INFO httpx: HTTP Request: GET https://eduardo4547-hyper-reality-sam2-gpu.hf.space/gradio_api/queue/data?session_hash=b306737b-bf28-4f02-8963-83b74032d449 "HTTP/1.1 200 OK"
+2026-04-29 20:34:24,624 INFO httpx: HTTP Request: GET https://eduardo4547-hyper-reality-sam2-gpu.hf.space/gradio_api/file=/tmp/gradio/edcd82066c73a2359a9133192b4091c873fb9fca768c0f377a7a519aeb9f9b3a/image.webp "HTTP/1.1 200 OK"
+2026-04-29 20:34:24,678 INFO services.gradio_client_service: Gradio Space segmentation: entorno=🌳 Terraza / Patio / Jardín motor=Híbrido Arquitectura (Cityscapes Grande + DINO Pequeño) mask_count=20
+2026-04-29 20:34:24,684 INFO backend.segmentation: [JOB ccecb5fbdf744d9eae57eb2b04e0af13] saving_masks progress=92
+2026-04-29 20:34:24,781 INFO backend.segmentation: [JOB ccecb5fbdf744d9eae57eb2b04e0af13] segments_meta saved (20 segments)
+2026-04-29 20:34:24,782 INFO backend.segmentation: [JOB ccecb5fbdf744d9eae57eb2b04e0af13] done mask_count=20
+2026-04-29 20:34:24,782 INFO backend.segmentation: [UPLOAD_BG] DONE 21.955s at 2026-04-30T02:34:24.782746+00:00
+2026-04-29 20:34:24,783 INFO backend.segmentation: Releasing resources (full_unload=False)
+2026-04-29 20:34:27,987 INFO backend.segmentation: [APPLY_TEXTURE] START at 2026-04-30T02:34:27.987941+00:00
+2026-04-29 20:34:28,292 INFO backend.segmentation: [APPLY_TEXTURE] cleared mask from original source: C:\Users\alane\OneDrive\Escritorio\Trabajo\Prueba-PoC\backend\uploads\proyecto-White-Houses.jpg
+2026-04-29 20:34:28,570 INFO backend.segmentation: [APPLY_TEXTURE] DONE 0.582s at 2026-04-30T02:34:28.570140+00:00

backend/requirements.txt

@@ -1,15 +1,11 @@
-fastapi==0.111.1
-python-dotenv
-uvicorn[standard]==0.23.2
-motor==3.4.0
-pymongo==4.7.2
-transformers==4.36.2
-numpy
-Pillow
-pydantic
-hydra-core
-omegaconf
-git+https://github.com/facebookresearch/sam2.git
-python-multipart
-torch
-opencv-python-headless
+fastapi==0.111.1
+python-dotenv
+uvicorn[standard]==0.23.2
+motor==3.4.0
+pymongo==4.7.2
+numpy
+Pillow
+pydantic
+python-multipart
+gradio_client
+opencv-python-headless

backend/routers/catalog.py

@@ -123,132 +123,6 @@ _CATALOG = [
             },
         ],
     },
-    {
-        "id": "pisos",
-        "nombre": "Pisos",
-        "descripcion": "Pisos de madera, laminado y mármol",
-        "productos": [
-            {
-                "id": "piso_madera",
-                "nombre": "Madera Natural",
-                "textura": "wood_floor_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/wood_floor_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "piso_tablones",
-                "nombre": "Tablones de Madera",
-                "textura": "wood_planks_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/wood_planks_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "piso_laminado",
-                "nombre": "Laminado",
-                "textura": "laminate_floor_02_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/laminate_floor_02_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "piso_duela",
-                "nombre": "Duela",
-                "textura": "plank_flooring_04_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/plank_flooring_04_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "piso_marmol",
-                "nombre": "Mármol",
-                "textura": "marble_01_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/marble_01_diff_2k.jpg",
-                "dimensiones": [],
-            },
-        ],
-    },
-    {
-        "id": "piedra_concreto",
-        "nombre": "Piedra y Concreto",
-        "descripcion": "Materiales de piedra, adoquín y granito",
-        "productos": [
-            {
-                "id": "piedra_ladrillo",
-                "nombre": "Ladrillo / Pavimento",
-                "textura": "brick_pavement_03_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/brick_pavement_03_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "piedra_adoquin",
-                "nombre": "Adoquín",
-                "textura": "cobblestone_floor_08_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/cobblestone_floor_08_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "piedra_roca",
-                "nombre": "Roca Natural",
-                "textura": "dry_riverbed_rock_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/dry_riverbed_rock_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "piedra_granito",
-                "nombre": "Granito",
-                "textura": "granite_tile_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/granite_tile_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "piedra_cartago",
-                "nombre": "Cartago Gris",
-                "textura": "grey_cartago_03_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/grey_cartago_03_diff_2k.jpg",
-                "dimensiones": [],
-            },
-        ],
-    },
-    {
-        "id": "metales",
-        "nombre": "Metales",
-        "descripcion": "Acabados metálicos y estructurales",
-        "productos": [
-            {
-                "id": "metal_azul",
-                "nombre": "Metal Azul",
-                "textura": "blue_metal_plate_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/blue_metal_plate_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "metal_placa",
-                "nombre": "Placa Metálica",
-                "textura": "metal_plate_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/metal_plate_diff_2k.jpg",
-                "dimensiones": [],
-            },
-            {
-                "id": "metal_oxido",
-                "nombre": "Metal Oxidado",
-                "textura": "rusty_metal_grid_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/rusty_metal_grid_diff_2k.jpg",
-                "dimensiones": [],
-            },
-        ],
-    },
-    {
-        "id": "madera",
-        "nombre": "Madera",
-        "descripcion": "Acabados en madera para interiores",
-        "productos": [
-            {
-                "id": "madera_gabinete",
-                "nombre": "Gabinete de Madera",
-                "textura": "wood_cabinet_worn_long_diff_2k.jpg",
-                "url_preview": "/seg/texture-preview/wood_cabinet_worn_long_diff_2k.jpg",
-                "dimensiones": [],
-            },
-        ],
-    },
 ]
 
 

backend/services/gradio_client_service.py

@@ -0,0 +1,93 @@
+"""
+Cliente para llamar al Gradio Space de ZeroGPU (SAM2 + SegFormer + DINO).
+Se activa solo si GRADIO_SPACE_URL está definido en el entorno.
+"""
+import asyncio
+import base64
+import io
+import json
+import logging
+from pathlib import Path
+
+import numpy as np
+from PIL import Image
+
+from core.config import GRADIO_SPACE_URL
+
+logger = logging.getLogger(__name__)
+
+
+def is_gradio_enabled() -> bool:
+    return bool(GRADIO_SPACE_URL)
+
+
+def _call_gradio_sync(image_path: Path) -> tuple[np.ndarray, int]:
+    """
+    Synchronous Gradio call — safe to invoke from a background thread.
+    Returns (label_map, mask_count).
+    Raises on any error so the caller can handle fallback.
+    """
+    from gradio_client import Client, file  # type: ignore
+
+    client = Client(GRADIO_SPACE_URL)
+
+    # segment_for_backend returns (overlay_image, combined_json_str)
+    _overlay_file, combined_json_str = client.predict(
+        file(str(image_path)),
+        api_name="/segment",
+    )
+
+    if not isinstance(combined_json_str, str):
+        raise ValueError(f"Unexpected response type from Gradio Space: {type(combined_json_str)}")
+
+    combined: dict = json.loads(combined_json_str)
+
+    if "error" in combined:
+        raise RuntimeError(f"Gradio Space error: {combined['error'][:500]}")
+
+    label_map_b64: str = combined.get("label_map_b64", "")
+    if not label_map_b64:
+        return np.zeros((1, 1), dtype=np.uint8), 0
+
+    # Decode PNG-encoded label map (lossless uint8 grayscale)
+    label_map_bytes = base64.b64decode(label_map_b64)
+    pil_label = Image.open(io.BytesIO(label_map_bytes))
+    label_map = np.array(pil_label, dtype=np.uint8)
+    mask_count = int(label_map.max())
+
+    entorno = combined.get("entorno", "?")
+    motor = combined.get("motor", "?")
+    logger.info(
+        "Gradio Space segmentation: entorno=%s motor=%s mask_count=%d",
+        entorno, motor, mask_count,
+    )
+
+    return label_map, mask_count
+
+
+def segment_via_gradio_sync(image_path: Path) -> tuple[np.ndarray, int] | None:
+    """
+    Blocking call to the Gradio Space from a sync context (background task thread).
+    Returns (label_map, mask_count) or None on failure / not configured.
+    """
+    if not is_gradio_enabled():
+        return None
+    try:
+        return _call_gradio_sync(image_path)
+    except Exception as exc:
+        logger.warning("Gradio Space call failed, falling back to local SAM2: %s", exc)
+        return None
+
+
+async def segment_via_gradio(image_path: Path) -> tuple[np.ndarray, int] | None:
+    """
+    Async wrapper — offloads the blocking Gradio call to a thread.
+    Returns (label_map, mask_count) or None on failure / not configured.
+    """
+    if not is_gradio_enabled():
+        return None
+    try:
+        return await asyncio.to_thread(_call_gradio_sync, image_path)
+    except Exception as exc:
+        logger.warning("Gradio Space call failed: %s", exc)
+        return None

backend/services/image_service.py

@@ -18,6 +18,7 @@ from core.config import (
     logger,
     utc_now_iso,
 )
+from services.gradio_client_service import is_gradio_enabled, segment_via_gradio_sync
 from services.sam2_service import (
     jobs,
     jobs_lock,
@@ -131,8 +132,17 @@ def run_upload_job(job_id: str, content: bytes, original_name: str) -> None:
             })
         logger.info(f"[JOB {job_id}] segmenting_with_sam2 progress=30 estimated_seconds={estimated_seconds}")
 
-        generate_label_map = _get_generate_label_map()
-        label_map, mask_count = generate_label_map(image_rgb)
+        image_path = UPLOAD_DIR / safe_name
+        if is_gradio_enabled():
+            gradio_result = segment_via_gradio_sync(image_path)
+        else:
+            gradio_result = None
+
+        if gradio_result is not None:
+            label_map, mask_count = gradio_result
+        else:
+            generate_label_map = _get_generate_label_map()
+            label_map, mask_count = generate_label_map(image_rgb)
 
         with jobs_lock:
             job = jobs.setdefault(job_id, {})

backend/services/sam2_service.py

@@ -6,7 +6,13 @@ from contextlib import asynccontextmanager
 from pathlib import Path
 from typing import Any
 
-import torch
+try:
+    import torch
+    _TORCH_AVAILABLE = True
+except ImportError:
+    torch = None  # type: ignore[assignment]
+    _TORCH_AVAILABLE = False
+
 from fastapi import FastAPI, HTTPException
 
 from core.config import (
@@ -107,6 +113,11 @@ def find_sam2_model_path() -> Path:
 def load_sam2_model() -> None:
     global sam2_mask_generator, sam2_image_predictor, sam2_load_error
 
+    if not _TORCH_AVAILABLE:
+        sam2_load_error = "torch not installed — SAM2 unavailable (using Gradio Space)"
+        logger.warning(f"[SAM2] {sam2_load_error}")
+        return
+
     if not _SAM2_AVAILABLE:
         sam2_load_error = "SAM2 library not installed"
         logger.error(f"[SAM2] ERROR: {sam2_load_error}")
@@ -226,29 +237,30 @@ def release_resources(full_unload: bool = False) -> None:
             sam2_image_predictor = None
             sam2_load_error = None
 
-        try:
-            if torch.cuda.is_available():
-                try:
-                    torch.cuda.synchronize()
-                except Exception:
-                    pass
-                try:
-                    torch.cuda.empty_cache()
-                except Exception:
-                    pass
-                try:
-                    if hasattr(torch.cuda, "ipc_collect"):
-                        torch.cuda.ipc_collect()
-                except Exception:
-                    pass
-                try:
-                    reserved = torch.cuda.memory_reserved()
-                    allocated = torch.cuda.memory_allocated()
-                    logger.info(f"CUDA memory reserved={reserved} allocated={allocated}")
-                except Exception:
-                    pass
-        except Exception:
-            pass
+        if _TORCH_AVAILABLE:
+            try:
+                if torch.cuda.is_available():
+                    try:
+                        torch.cuda.synchronize()
+                    except Exception:
+                        pass
+                    try:
+                        torch.cuda.empty_cache()
+                    except Exception:
+                        pass
+                    try:
+                        if hasattr(torch.cuda, "ipc_collect"):
+                            torch.cuda.ipc_collect()
+                    except Exception:
+                        pass
+                    try:
+                        reserved = torch.cuda.memory_reserved()
+                        allocated = torch.cuda.memory_allocated()
+                        logger.info(f"CUDA memory reserved={reserved} allocated={allocated}")
+                    except Exception:
+                        pass
+            except Exception:
+                pass
 
         gc.collect()
     except Exception as exc:
@@ -258,6 +270,11 @@ def release_resources(full_unload: bool = False) -> None:
 @asynccontextmanager
 async def lifespan(app: FastAPI) -> AsyncIterator[None]:
     import asyncio
+    from services.gradio_client_service import is_gradio_enabled
+    if is_gradio_enabled():
+        logger.info("[LIFESPAN] GRADIO_SPACE_URL set — skipping local SAM2 load.")
+        yield
+        return
     try:
         load_sam2_model()
         yield

backend/services/segmentation_service.py

@@ -3,7 +3,12 @@ from typing import Any, cast
 
 import cv2
 import numpy as np
-import torch
+try:
+    import torch
+    _TORCH_AVAILABLE = True
+except ImportError:
+    torch = None  # type: ignore[assignment]
+    _TORCH_AVAILABLE = False
 from fastapi import HTTPException
 from PIL import Image
 

backend/uploads/masks/demo-room.jpg_labels_meta.json

@@ -0,0 +1 @@
+{"segments": [{"index": 1, "type": "object", "label": "Objeto 1", "area_ratio": 0.0007}, {"index": 2, "type": "object", "label": "Objeto 2", "area_ratio": 0.0}, {"index": 3, "type": "object", "label": "Objeto 3", "area_ratio": 0.0001}, {"index": 4, "type": "door", "label": "Puerta 1", "area_ratio": 0.0217}, {"index": 5, "type": "wall", "label": "Pared 1", "area_ratio": 0.1139}, {"index": 6, "type": "door", "label": "Puerta 2", "area_ratio": 0.0518}, {"index": 7, "type": "object", "label": "Objeto 4", "area_ratio": 0.0027}, {"index": 8, "type": "wall", "label": "Pared 2", "area_ratio": 0.0064}, {"index": 9, "type": "object", "label": "Objeto 5", "area_ratio": 0.0017}, {"index": 10, "type": "window", "label": "Ventana 1", "area_ratio": 0.0334}, {"index": 11, "type": "wall", "label": "Pared 3", "area_ratio": 0.3694}, {"index": 12, "type": "window", "label": "Ventana 2", "area_ratio": 0.0205}, {"index": 13, "type": "window", "label": "Ventana 3", "area_ratio": 0.0059}, {"index": 14, "type": "window", "label": "Ventana 4", "area_ratio": 0.0051}, {"index": 15, "type": "window", "label": "Ventana 5", "area_ratio": 0.0047}, {"index": 16, "type": "window", "label": "Ventana 6", "area_ratio": 0.0919}]}