File size: 8,739 Bytes
0e805d4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
aa27972
0e805d4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
aa27972
 
0e805d4
aa27972
0e805d4
aa27972
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0e805d4
 
aa27972
 
 
 
 
 
 
 
 
 
 
 
0e805d4
aa27972
 
0e805d4
aa27972
 
0e805d4
 
 
aa27972
 
 
 
0e805d4
 
 
aa27972
 
 
 
 
 
0e805d4
 
aa27972
 
 
 
 
 
 
 
 
 
 
0e805d4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26f8b9a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0e805d4
26f8b9a
 
 
 
 
 
0e805d4
26f8b9a
0e805d4
26f8b9a
0e805d4
 
26f8b9a
 
 
 
 
 
0e805d4
 
26f8b9a
 
0e805d4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
"""Hunyuan3D-2.1 3D model generation."""

# CRITICAL: Import spaces BEFORE torch/CUDA packages
import spaces

import torch
from pathlib import Path
from gradio_client import Client, handle_file
import httpx
from tenacity import retry, stop_after_attempt, wait_exponential, retry_if_exception_type

from core.config import HUNYUAN_SETTINGS, QualityPreset
from utils.memory import MemoryManager


class HunyuanGenerator:
    """Generates 3D models using Hunyuan3D-2.1."""
    
    def __init__(self):
        self.memory_manager = MemoryManager()
    
    @retry(
        stop=stop_after_attempt(3),
        wait=wait_exponential(multiplier=1, min=4, max=10),
        retry=retry_if_exception_type((httpx.TimeoutException, httpx.NetworkError))
    )
    def _call_api(self, client: Client, **kwargs):
        """Call Hunyuan3D API with automatic retry."""
        return client.predict(**kwargs)
    
    @spaces.GPU(duration=90)
    def generate(
        self,
        image_path: Path,
        preset: QualityPreset,
        output_dir: Path
    ) -> Path:
        """Generate 3D model from 2D image."""
        try:
            print(f"[Hunyuan3D] Generating 3D model: {preset.name} quality")
            print(f"[Hunyuan3D] Input image: {image_path}")
            print(f"[Hunyuan3D] Settings: steps={preset.hunyuan_steps}, guidance={preset.hunyuan_guidance}, octree={preset.octree_resolution}")
            
            # Validate input image exists
            if not image_path.exists():
                raise FileNotFoundError(f"Input image not found: {image_path}")
            
            # Connect to API
            print(f"[Hunyuan3D] Connecting to {HUNYUAN_SETTINGS['space_id']}...")
            client = Client(
                HUNYUAN_SETTINGS["space_id"],
                httpx_kwargs={
                    "timeout": httpx.Timeout(
                        HUNYUAN_SETTINGS["timeout"],
                        connect=HUNYUAN_SETTINGS["connect_timeout"]
                    )
                }
            )
            print(f"[Hunyuan3D] Connected successfully")
            
            # Call API (returns tuple: file, output, mesh_stats, seed)
            print(f"[Hunyuan3D] Calling API with parameters...")
            result = self._call_api(
                client,
                image=handle_file(str(image_path)),
                mv_image_front=None,
                mv_image_back=None,
                mv_image_left=None,
                mv_image_right=None,
                steps=preset.hunyuan_steps,
                guidance_scale=preset.hunyuan_guidance,
                seed=1234,
                octree_resolution=preset.octree_resolution,
                check_box_rembg=True,
                num_chunks=preset.num_chunks,
                randomize_seed=True,
                api_name="/shape_generation"
            )
            print(f"[Hunyuan3D] API call completed")
            
            # Extract GLB path from tuple response
            # API returns: (file, output, mesh_stats, seed)
            print(f"[Hunyuan3D] Raw result type: {type(result)}")
            print(f"[Hunyuan3D] Raw result length: {len(result) if isinstance(result, (tuple, list)) else 'N/A'}")
            
            if not isinstance(result, tuple):
                raise ValueError(
                    f"Unexpected result type from Hunyuan3D API: {type(result)}. "
                    f"Expected tuple of (file, output, mesh_stats, seed)."
                )
            
            if len(result) != 4:
                raise ValueError(
                    f"Unexpected result length from Hunyuan3D API: {len(result)}. "
                    f"Expected 4 elements (file, output, mesh_stats, seed), got {len(result)}."
                )
            
            # Extract GLB file path (first element)
            file_data, html_output, mesh_stats, used_seed = result
            
            print(f"[Hunyuan3D] file_data type: {type(file_data)}")
            print(f"[Hunyuan3D] mesh_stats: {mesh_stats}")
            print(f"[Hunyuan3D] used_seed: {used_seed}")
            
            # Extract path from file_data
            if file_data is None:
                raise ValueError(
                    "Hunyuan3D API returned None for file. "
                    "This usually means the generation failed on the server side. "
                    "Possible causes:\n"
                    "  - Invalid image input\n"
                    "  - API timeout\n"
                    "  - Server overload\n"
                    "Try again with a different image or quality setting."
                )
            
            # Handle different file_data formats
            if isinstance(file_data, dict):
                print(f"[Hunyuan3D] file_data is dict with keys: {file_data.keys()}")
                if 'path' in file_data:
                    glb_path = file_data['path']
                elif 'value' in file_data:
                    glb_path = file_data['value']
                elif 'name' in file_data:
                    glb_path = file_data['name']
                else:
                    raise ValueError(
                        f"Unexpected dict format from Hunyuan3D API. "
                        f"Keys: {list(file_data.keys())}"
                    )
            elif isinstance(file_data, str):
                glb_path = file_data
            else:
                raise ValueError(
                    f"Unexpected file_data type: {type(file_data)}. "
                    f"Expected dict or str."
                )
            
            print(f"[Hunyuan3D] Extracted GLB path: {glb_path}")
            
            # Validate path exists
            if not glb_path or glb_path == "None":
                raise ValueError(
                    "Hunyuan3D API returned invalid path. "
                    "The generation may have failed on the server side."
                )
            
            if not Path(glb_path).exists():
                raise ValueError(
                    f"GLB file not found at path: {glb_path}. "
                    f"The file may not have been generated or saved correctly."
                )
            
            print(f"[Hunyuan3D] Model generated: {glb_path}")
            
            # Cleanup
            del client
            import gc
            gc.collect()
            torch.cuda.empty_cache()
            
            return Path(glb_path)
            
        except Exception as e:
            import traceback
            error_details = traceback.format_exc()
            print(f"[Hunyuan3D] ERROR: {e}")
            print(f"[Hunyuan3D] Full traceback:\n{error_details}")
            
            # Provide helpful error message based on error type
            error_str = str(e).lower()
            
            if "quota" in error_str or "zerogpu" in error_str:
                raise RuntimeError(
                    f"⚠️ Hunyuan3D Space is out of GPU quota.\n"
                    f"This is a limitation of the free Hunyuan3D-2.1 Space.\n\n"
                    f"Solutions:\n"
                    f"1. Wait for quota reset (resets daily)\n"
                    f"2. Try again in a few hours\n"
                    f"3. Use a different time of day (less traffic)\n\n"
                    f"Note: Your L4 GPU is only used for FLUX generation.\n"
                    f"Hunyuan3D runs on an external space with quota limits."
                ) from e
            elif "list index out of range" in str(e) or "unexpected result" in error_str:
                raise ValueError(
                    f"❌ Hunyuan3D API returned empty result.\n"
                    f"This usually means:\n"
                    f"1. The Hunyuan3D Space is overloaded\n"
                    f"2. GPU quota exhausted\n"
                    f"3. Invalid image input\n\n"
                    f"Try again in a few minutes."
                ) from e
            elif "timeout" in error_str:
                raise TimeoutError(
                    f"⏱️ Hunyuan3D generation timed out.\n"
                    f"Try using a lower quality preset (Fast or Balanced)."
                ) from e
            elif "not found" in error_str or "404" in error_str:
                raise RuntimeError(
                    f"❌ Hunyuan3D Space not accessible.\n"
                    f"The tencent/Hunyuan3D-2.1 Space may be down or moved.\n"
                    f"Check: https://huggingface.co/spaces/tencent/Hunyuan3D-2.1"
                ) from e
            else:
                raise RuntimeError(
                    f"❌ Hunyuan3D generation failed: {e}\n"
                    f"Check the Hunyuan3D Space status and try again."
                ) from e