Update app.py

app.py

@@ -1,73 +1,38 @@
-import logging
+import io
 import os
 import shlex
 import subprocess
 import tempfile
-import time
+import zipfile
+from functools import partial
+
+from fastapi import FastAPI, File, UploadFile, Form, HTTPException
+from fastapi.responses import StreamingResponse
+from fastapi.middleware.cors import CORSMiddleware
 
-import gradio as gr
 import numpy as np
 import rembg
-import spaces
 import torch
 from PIL import Image
-from functools import partial
-
-subprocess.run(shlex.split('pip install wheel/torchmcubes-0.1.0-cp310-cp310-linux_x86_64.whl'))
 
 from tsr.system import TSR
 from tsr.utils import remove_background, resize_foreground, to_gradio_3d_orientation
 
+# ------------------------------------------------------------
+# 1. Model & utils initialization (runs at startup)
+# ------------------------------------------------------------
+# Install any local wheels (if needed)
+subprocess.run(
+    shlex.split(
+        "pip install wheel/torchmcubes-0.1.0-cp310-cp310-linux_x86_64.whl"
+    ),
+    check=False,
+)
 
-HEADER = """
-# TripoSR Demo
-<table bgcolor="#1E2432" cellspacing="0" cellpadding="0"  width="450">
-<tr style="height:50px;">
-<td style="text-align: center;">
-<a href="https://stability.ai">
-<img src="https://images.squarespace-cdn.com/content/v1/6213c340453c3f502425776e/6c9c4c25-5410-4547-bc26-dc621cdacb25/Stability+AI+logo.png" width="200" height="40" />
-</a>
-</td>
-<td style="text-align: center;">
-<a href="https://www.tripo3d.ai">
-<img src="https://tripo-public.cdn.bcebos.com/logo.png" width="40" height="40" />
-</a>
-</td>
-</tr>
-</table>
-<table bgcolor="#1E2432" cellspacing="0" cellpadding="0"  width="450">
-<tr style="height:30px;">
-<td style="text-align: center;">
-<a href="https://huggingface.co/stabilityai/TripoSR"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Model_Card-Huggingface-orange" height="20"></a>
-</td>
-<td style="text-align: center;">
-<a href="https://github.com/VAST-AI-Research/TripoSR"><img src="https://postimage.me/images/2024/03/04/GitHub_Logo_White.png" width="100" height="20"></a>
-</td>
-<td style="text-align: center; color: white;">
-<a href="https://arxiv.org/abs/2403.02151"><img src="https://img.shields.io/badge/arXiv-2403.02151-b31b1b.svg" height="20"></a>
-</td>
-</tr>
-</table>
-
-> Try our new model: **SF3D** with several improvements such as faster generation and more game-ready assets.
-> 
-> The model is available [here](https://huggingface.co/stabilityai/stable-fast-3d) and we also have a [demo](https://huggingface.co/spaces/stabilityai/stable-fast-3d).
- 
-
-**TripoSR** is a state-of-the-art open-source model for **fast** feedforward 3D reconstruction from a single image, developed in collaboration between [Tripo AI](https://www.tripo3d.ai/) and [Stability AI](https://stability.ai/).
-
-**Tips:**
-1. If you find the result is unsatisfied, please try to change the foreground ratio. It might improve the results.
-2. It's better to disable "Remove Background" for the provided examples since they have been already preprocessed.
-3. Otherwise, please disable "Remove Background" option only if your input image is RGBA with transparent background, image contents are centered and occupy more than 70% of image width or height.
-"""
-
-
-if torch.cuda.is_available():
-    device = "cuda:0"
-else:
-    device = "cpu"
+# device
+device = "cuda:0" if torch.cuda.is_available() else "cpu"
 
+# load model
 model = TSR.from_pretrained(
     "stabilityai/TripoSR",
     config_name="config.yaml",
@@ -76,20 +41,26 @@ model = TSR.from_pretrained(
 model.renderer.set_chunk_size(131072)
 model.to(device)
 
+# background removal
 rembg_session = rembg.new_session()
 
 
-def check_input_image(input_image):
-    if input_image is None:
-        raise gr.Error("No image uploaded!")
+def check_input_image(image: Image.Image):
+    if image is None:
+        raise HTTPException(status_code=400, detail="No image uploaded!")
 
 
-def preprocess(input_image, do_remove_background, foreground_ratio):
-    def fill_background(image):
-        image = np.array(image).astype(np.float32) / 255.0
-        image = image[:, :, :3] * image[:, :, 3:4] + (1 - image[:, :, 3:4]) * 0.5
-        image = Image.fromarray((image * 255.0).astype(np.uint8))
-        return image
+def preprocess(
+    input_image: Image.Image, do_remove_background: bool, foreground_ratio: float
+) -> Image.Image:
+    """
+    Mimics the Gradio preprocess(...) function.
+    """
+    def fill_background(image: Image.Image) -> Image.Image:
+        arr = np.array(image).astype(np.float32) / 255.0
+        arr = arr[:, :, :3] * arr[:, :, 3:4] + (1 - arr[:, :, 3:4]) * 0.5
+        out = (arr * 255.0).astype(np.uint8)
+        return Image.fromarray(out)
 
     if do_remove_background:
         image = input_image.convert("RGB")
@@ -100,97 +71,101 @@ def preprocess(input_image, do_remove_background, foreground_ratio):
         image = input_image
         if image.mode == "RGBA":
             image = fill_background(image)
+
     return image
 
 
-@spaces.GPU
-def generate(image, mc_resolution, formats=["obj", "glb"]):
+def generate(
+    image: Image.Image, mc_resolution: int, formats=["obj", "glb"]
+) -> tuple[str, str]:
+    """
+    Mimics the Gradio generate(...) function.
+    Returns paths to .obj and .glb on disk.
+    """
+    # 1. inference
     scene_codes = model(image, device=device)
     mesh = model.extract_mesh(scene_codes, resolution=mc_resolution)[0]
     mesh = to_gradio_3d_orientation(mesh)
 
-    mesh_path_glb = tempfile.NamedTemporaryFile(suffix=f".glb", delete=False)
-    mesh.export(mesh_path_glb.name)
-
-    mesh_path_obj = tempfile.NamedTemporaryFile(suffix=f".obj", delete=False)
-    mesh.apply_scale([-1, 1, 1])  # Otherwise the visualized .obj will be flipped
-    mesh.export(mesh_path_obj.name)
-    
-    return mesh_path_obj.name, mesh_path_glb.name
-
-def run_example(image_pil):
-    preprocessed = preprocess(image_pil, False, 0.9)
-    mesh_name_obj, mesh_name_glb = generate(preprocessed, 256, ["obj", "glb"])
-    return preprocessed, mesh_name_obj, mesh_name_glb
-
-with gr.Blocks() as demo:
-    gr.Markdown(HEADER)
-    with gr.Row(variant="panel"):
-        with gr.Column():
-            with gr.Row():
-                input_image = gr.Image(
-                    label="Input Image",
-                    image_mode="RGBA",
-                    sources="upload",
-                    type="pil",
-                    elem_id="content_image",
-                )
-                processed_image = gr.Image(label="Processed Image", interactive=False)
-            with gr.Row():
-                with gr.Group():
-                    do_remove_background = gr.Checkbox(
-                        label="Remove Background", value=True
-                    )
-                    foreground_ratio = gr.Slider(
-                        label="Foreground Ratio",
-                        minimum=0.5,
-                        maximum=1.0,
-                        value=0.85,
-                        step=0.05,
-                    )
-                    mc_resolution = gr.Slider(
-                        label="Marching Cubes Resolution",
-                        minimum=32,
-                        maximum=320,
-                        value=256,
-                        step=32
-                     )
-            with gr.Row():
-                submit = gr.Button("Generate", elem_id="generate", variant="primary")
-        with gr.Column():
-            with gr.Tab("OBJ"):
-                output_model_obj = gr.Model3D(
-                    label="Output Model (OBJ Format)",
-                    interactive=False,
-                )
-                gr.Markdown("Note: Downloaded object will be flipped in case of .obj export. Export .glb instead or manually flip it before usage.")
-            with gr.Tab("GLB"):
-                output_model_glb = gr.Model3D(
-                    label="Output Model (GLB Format)",
-                    interactive=False,
-                )
-                gr.Markdown("Note: The model shown here has a darker appearance. Download to get correct results.")
-    with gr.Row(variant="panel"):
-        gr.Examples(
-            examples=[
-                os.path.join("examples", img_name) for img_name in sorted(os.listdir("examples"))
-            ],
-            inputs=[input_image],
-            outputs=[processed_image, output_model_obj, output_model_glb],
-            cache_examples=True,
-            fn=partial(run_example),
-            label="Examples",
-            examples_per_page=20
-        )
-    submit.click(fn=check_input_image, inputs=[input_image]).success(
-        fn=preprocess,
-        inputs=[input_image, do_remove_background, foreground_ratio],
-        outputs=[processed_image],
-    ).success(
-        fn=generate,
-        inputs=[processed_image, mc_resolution],
-        outputs=[output_model_obj, output_model_glb],
-    )
-
-demo.queue(max_size=10)
-demo.launch()
+    # 2. export GLB
+    glb_tmp = tempfile.NamedTemporaryFile(suffix=".glb", delete=False)
+    mesh.export(glb_tmp.name)
+
+    # 3. export OBJ (flip x-axis so OBJ is not mirrored)
+    obj_tmp = tempfile.NamedTemporaryFile(suffix=".obj", delete=False)
+    mesh.apply_scale([-1, 1, 1])
+    mesh.export(obj_tmp.name)
+
+    return obj_tmp.name, glb_tmp.name
+
+
+# ------------------------------------------------------------
+# 2. FastAPI app
+# ------------------------------------------------------------
+app = FastAPI(title="TripoSR FastAPI Demo")
+
+# If you need CORS (e.g. calling from a browser-based front-end)
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["POST", "GET", "OPTIONS"],
+    allow_headers=["*"],
+)
+
+
+@app.post("/generate", response_class=StreamingResponse)
+async def generate_endpoint(
+    image_file: UploadFile = File(...),
+    do_remove_background: bool = Form(True),
+    foreground_ratio: float = Form(0.85),
+    mc_resolution: int = Form(256),
+):
+    """
+    1. Read & validate image
+    2. Preprocess
+    3. Generate mesh
+    4. Package processed image + .obj + .glb into a ZIP
+    """
+    # 1) Read image bytes
+    contents = await image_file.read()
+    try:
+        pil_img = Image.open(io.BytesIO(contents))
+    except Exception:
+        raise HTTPException(status_code=400, detail="Invalid image file")
+
+    check_input_image(pil_img)
+
+    # 2) Preprocess
+    processed = preprocess(pil_img, do_remove_background, foreground_ratio)
+
+    # 3) Generate mesh
+    obj_path, glb_path = generate(processed, mc_resolution)
+
+    # 4) Create in-memory ZIP
+    zip_buffer = io.BytesIO()
+    with zipfile.ZipFile(zip_buffer, mode="w") as zf:
+        # processed image
+        buf = io.BytesIO()
+        processed.save(buf, format="PNG")
+        zf.writestr("processed.png", buf.getvalue())
+
+        # .obj
+        with open(obj_path, "rb") as f:
+            zf.writestr(os.path.basename(obj_path), f.read())
+
+        # .glb
+        with open(glb_path, "rb") as f:
+            zf.writestr(os.path.basename(glb_path), f.read())
+
+    zip_buffer.seek(0)
+
+    # Cleanup temp files
+    os.remove(obj_path)
+    os.remove(glb_path)
+
+    headers = {
+        "Content-Disposition": 'attachment; filename="tripo_output.zip"'
+    }
+    return StreamingResponse(
+        zip_buffer, media_type="application/zip", headers=headers
+    )