Update app.py

app.py

@@ -1,154 +1,86 @@
+import os
+import base64
+import traceback
+import requests
+import fastapi
 import gradio as gr
-import numpy as np
-import random
-
-# import spaces #[uncomment to use ZeroGPU]
-from diffusers import DiffusionPipeline
-import torch
-
-device = "cuda" if torch.cuda.is_available() else "cpu"
-model_repo_id = "stabilityai/sdxl-turbo"  # Replace to the model you would like to use
-
-if torch.cuda.is_available():
-    torch_dtype = torch.float16
-else:
-    torch_dtype = torch.float32
-
-pipe = DiffusionPipeline.from_pretrained(model_repo_id, torch_dtype=torch_dtype)
-pipe = pipe.to(device)
-
-MAX_SEED = np.iinfo(np.int32).max
-MAX_IMAGE_SIZE = 1024
-
-
-# @spaces.GPU #[uncomment to use ZeroGPU]
-def infer(
-    prompt,
-    negative_prompt,
-    seed,
-    randomize_seed,
-    width,
-    height,
-    guidance_scale,
-    num_inference_steps,
-    progress=gr.Progress(track_tqdm=True),
-):
-    if randomize_seed:
-        seed = random.randint(0, MAX_SEED)
-
-    generator = torch.Generator().manual_seed(seed)
-
-    image = pipe(
-        prompt=prompt,
-        negative_prompt=negative_prompt,
-        guidance_scale=guidance_scale,
-        num_inference_steps=num_inference_steps,
-        width=width,
-        height=height,
-        generator=generator,
-    ).images[0]
-
-    return image, seed
-
-
-examples = [
-    "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k",
-    "An astronaut riding a green horse",
-    "A delicious ceviche cheesecake slice",
-]
-
-css = """
-#col-container {
-    margin: 0 auto;
-    max-width: 640px;
-}
-"""
-
-with gr.Blocks(css=css) as demo:
-    with gr.Column(elem_id="col-container"):
-        gr.Markdown(" # Text-to-Image Gradio Template")
-
-        with gr.Row():
-            prompt = gr.Text(
-                label="Prompt",
-                show_label=False,
-                max_lines=1,
-                placeholder="Enter your prompt",
-                container=False,
-            )
-
-            run_button = gr.Button("Run", scale=0, variant="primary")
-
-        result = gr.Image(label="Result", show_label=False)
-
-        with gr.Accordion("Advanced Settings", open=False):
-            negative_prompt = gr.Text(
-                label="Negative prompt",
-                max_lines=1,
-                placeholder="Enter a negative prompt",
-                visible=False,
-            )
-
-            seed = gr.Slider(
-                label="Seed",
-                minimum=0,
-                maximum=MAX_SEED,
-                step=1,
-                value=0,
-            )
-
-            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
-
-            with gr.Row():
-                width = gr.Slider(
-                    label="Width",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=1024,  # Replace with defaults that work for your model
-                )
-
-                height = gr.Slider(
-                    label="Height",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=1024,  # Replace with defaults that work for your model
-                )
-
-            with gr.Row():
-                guidance_scale = gr.Slider(
-                    label="Guidance scale",
-                    minimum=0.0,
-                    maximum=10.0,
-                    step=0.1,
-                    value=0.0,  # Replace with defaults that work for your model
-                )
-
-                num_inference_steps = gr.Slider(
-                    label="Number of inference steps",
-                    minimum=1,
-                    maximum=50,
-                    step=1,
-                    value=2,  # Replace with defaults that work for your model
-                )
-
-        gr.Examples(examples=examples, inputs=[prompt])
-    gr.on(
-        triggers=[run_button.click, prompt.submit],
-        fn=infer,
-        inputs=[
-            prompt,
-            negative_prompt,
-            seed,
-            randomize_seed,
-            width,
-            height,
-            guidance_scale,
-            num_inference_steps,
-        ],
-        outputs=[result, seed],
-    )
-
-if __name__ == "__main__":
-    demo.launch()
+from pydantic import BaseModel
+
+# 1. Configuración de URLs y Variables de Entorno
+API_URL = "https://api-inference.huggingface.co/models/joyfox/Qwen-Image-Edit-Cosplay"
+HF_TOKEN = os.getenv("HF_TOKEN")
+
+# 2. Inicializamos FastAPI
+app = fastapi.FastAPI()
+
+class CosplayPayload(BaseModel):
+    data: list[str]
+
+@app.post("/api/predict")
+async def fusionar_cosplay(payload: CosplayPayload):
+    if not HF_TOKEN:
+        print("[-] ERROR: El HF_TOKEN no está configurado en las variables (Secrets) del Space.")
+        raise fastapi.HTTPException(status_code=500, detail="Falta el token HF_TOKEN")
+
+    if len(payload.data) < 2:
+        raise fastapi.HTTPException(status_code=400, detail="Se requieren exactamente 2 imágenes.")
+
+    try:
+        print("[+] Petición recibida desde Android. Procesando cadenas Base64...")
+        
+        # Aseguramos el prefijo Data URI para el JSON de Qwen
+        data_uri_real = payload.data[0] if payload.data[0].startswith("data:") else f"data:image/jpeg;base64,{payload.data[0]}"
+        data_uri_ref = payload.data[1] if payload.data[1].startswith("data:") else f"data:image/jpeg;base64,{payload.data[1]}"
+        
+        prompt_texto = "把图2中动漫人物中的发型，服饰和装扮道具迁移到图1的真实人物，图1真实人物保持高度人脸一致性"
+        
+        # Payload JSON limpio para la API de Inferencia Serverless
+        json_payload = {
+            "inputs": {
+                "text": prompt_texto,
+                "images": [data_uri_real, data_uri_ref]
+            },
+            "parameters": {
+                "max_new_tokens": 512
+            }
+        }
+        
+        headers = {
+            "Authorization": f"Bearer {HF_TOKEN}",
+            "Content-Type": "application/json"
+        }
+        
+        print("[+] Enviando petición a la API de Inferencia de Hugging Face mediante el puente de Gradio...")
+        
+        # Al estar bajo el SDK de Gradio, esta petición ya NO lanzará NameResolutionError
+        respuesta = requests.post(API_URL, headers=headers, json=json_payload, timeout=110)
+        print(f"[+] Respuesta recibida. Código de estado: {respuesta.status_code}")
+        
+        if respuesta.status_code == 200:
+            content_type = respuesta.headers.get("content-type", "").lower()
+            
+            if "application/json" in content_type:
+                return respuesta.json()
+            else:
+                imagen_procesada_base64 = base64.b64encode(respuesta.content).decode("utf-8")
+                return {"data": f"data:image/jpeg;base64,{imagen_procesada_base64}"}
+        else:
+            print(f"[-] Error de la API de Inferencia: {respuesta.text}")
+            raise fastapi.HTTPException(status_code=502, detail=f"HF API Error: {respuesta.status_code}")
+
+    except fastapi.HTTPException as status_error:
+        raise status_error
+    except Exception as e:
+        print("\n================= CRASH EN EL PROXY =================")
+        traceback.print_exc()
+        print("=====================================================\n")
+        raise fastapi.HTTPException(status_code=500, detail=f"Crash interno: {str(e)}")
+
+# 3. Interfaz de Gradio Dummy (Obligatoria para que Hugging Face valide el Space)
+with gr.Blocks() as demo:
+    gr.Markdown("# Cosplay API Bridge Active 🎭")
+    gr.Markdown("Este Space está corriendo un servidor FastAPI en segundo plano optimizado para peticiones desde Android.")
+
+# 4. ¡MAGIA! Montamos FastAPI dentro del entorno de Gradio
+# Esto hace que FastAPI herede todas las configuraciones y accesos de red del contenedor de Gradio.
+app = gr.mount_gradio_app(app, demo, path="/")