Spaces:
Sleeping
Sleeping
File size: 7,660 Bytes
afad319 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 |
"""
Hugging Face Spaces Configuration
================================
This module contains configuration settings optimized for deployment on
Hugging Face Spaces. It handles cache directories, permissions, and
environment-specific optimizations.
Key Features:
- Automatic cache directory setup in /tmp
- Permission handling for HF Spaces environment
- Model loading optimizations
- Resource usage monitoring
"""
import os
import logging
from pathlib import Path
# Configure logging for HF Spaces
logging.basicConfig(
level=logging.INFO, format="%(asctime)s - %(name)s - %(levelname)s - %(message)s"
)
logger = logging.getLogger(__name__)
class HFSpacesConfig:
"""
Configuration class for Hugging Face Spaces deployment
This class manages all environment-specific settings and ensures
the application works correctly in the HF Spaces environment.
"""
def __init__(self):
"""Initialize HF Spaces configuration"""
self.is_hf_spaces = self._detect_hf_spaces()
self.cache_dirs = self._setup_cache_directories()
self.env_vars = self._setup_environment_variables()
def _detect_hf_spaces(self) -> bool:
"""
Detect if running in Hugging Face Spaces environment
Returns:
bool: True if running in HF Spaces
"""
# Check for HF Spaces environment indicators
hf_indicators = [
"SPACE_ID" in os.environ,
"SPACE_HOST" in os.environ,
"HF_HUB_ENDPOINT" in os.environ,
os.path.exists("/tmp/huggingface"),
]
is_hf = any(hf_indicators)
logger.info(f"HF Spaces environment detected: {is_hf}")
return is_hf
def _setup_cache_directories(self) -> dict:
"""
Set up cache directories for HF Spaces
Returns:
dict: Cache directory paths
"""
if self.is_hf_spaces:
# Use /tmp for HF Spaces (writable)
cache_dirs = {
"hf_home": "/tmp/huggingface",
"transformers_cache": "/tmp/huggingface/transformers",
"torch_home": "/tmp/torch",
"hub_cache": "/tmp/huggingface/hub",
"xdg_cache": "/tmp",
"vector_store": "./vector_store",
}
else:
# Use standard locations for local development
cache_dirs = {
"hf_home": os.path.expanduser("~/.cache/huggingface"),
"transformers_cache": os.path.expanduser(
"~/.cache/huggingface/transformers"
),
"torch_home": os.path.expanduser("~/.cache/torch"),
"hub_cache": os.path.expanduser("~/.cache/huggingface/hub"),
"xdg_cache": os.path.expanduser("~/.cache"),
"vector_store": "./vector_store",
}
# Create directories
for name, path in cache_dirs.items():
try:
Path(path).mkdir(parents=True, exist_ok=True)
logger.info(f"Cache directory ready: {name} -> {path}")
except Exception as e:
logger.warning(f"Could not create cache directory {name}: {e}")
return cache_dirs
def _setup_environment_variables(self) -> dict:
"""
Set up environment variables for HF Spaces
Returns:
dict: Environment variable settings
"""
env_vars = {
"HF_HOME": self.cache_dirs["hf_home"],
"TRANSFORMERS_CACHE": self.cache_dirs["transformers_cache"],
"TORCH_HOME": self.cache_dirs["torch_home"],
"XDG_CACHE_HOME": self.cache_dirs["xdg_cache"],
"HF_HUB_CACHE": self.cache_dirs["hub_cache"],
"PYTHONPATH": "/app",
"STREAMLIT_SERVER_PORT": "8501",
"STREAMLIT_SERVER_ADDRESS": "0.0.0.0",
"STREAMLIT_SERVER_HEADLESS": "true",
"STREAMLIT_SERVER_ENABLE_CORS": "false",
"STREAMLIT_SERVER_ENABLE_XSRF_PROTECTION": "false",
"STREAMLIT_LOGGER_LEVEL": "info",
}
# Set environment variables
for key, value in env_vars.items():
os.environ[key] = value
logger.info(f"Set environment variable: {key}={value}")
return env_vars
def get_model_config(self) -> dict:
"""
Get optimized model configuration for HF Spaces
Returns:
dict: Model configuration settings
"""
return {
"embedding_model": "all-MiniLM-L6-v2",
"generative_model": "Qwen/Qwen2.5-1.5B-Instruct",
"fallback_model": "distilgpt2",
"chunk_sizes": [512, 1024, 2048],
"vector_store_path": self.cache_dirs["vector_store"],
"enable_guard_rails": True,
"cache_dir": self.cache_dirs["transformers_cache"],
}
def get_guard_rail_config(self) -> dict:
"""
Get guard rail configuration optimized for HF Spaces
Returns:
dict: Guard rail configuration settings
"""
return {
"max_query_length": 1000,
"max_response_length": 5000,
"min_confidence_threshold": 0.3,
"rate_limit_requests": 10,
"rate_limit_window": 60,
"enable_pii_detection": True,
"enable_prompt_injection_detection": True,
}
def get_resource_limits(self) -> dict:
"""
Get resource limits for HF Spaces environment
Returns:
dict: Resource limit settings
"""
return {
"max_memory_usage": 0.8, # 80% of available memory
"max_cpu_usage": 0.9, # 90% of available CPU
"max_concurrent_requests": 5,
"model_timeout": 30, # seconds
"cache_cleanup_interval": 3600, # 1 hour
}
def cleanup_cache(self):
"""
Clean up cache directories to free space
This is important for HF Spaces with limited storage.
"""
if not self.is_hf_spaces:
return
try:
import shutil
import time
# Remove old cache files (older than 1 hour)
current_time = time.time()
for cache_path in [
self.cache_dirs["transformers_cache"],
self.cache_dirs["torch_home"],
]:
if os.path.exists(cache_path):
for item in os.listdir(cache_path):
item_path = os.path.join(cache_path, item)
if os.path.isfile(item_path):
if current_time - os.path.getmtime(item_path) > 3600:
os.remove(item_path)
logger.info(f"Cleaned up old cache file: {item_path}")
logger.info("Cache cleanup completed")
except Exception as e:
logger.warning(f"Cache cleanup failed: {e}")
# Global configuration instance
hf_config = HFSpacesConfig()
def get_hf_config() -> HFSpacesConfig:
"""
Get the global HF Spaces configuration instance
Returns:
HFSpacesConfig: Configuration instance
"""
return hf_config
def is_hf_spaces() -> bool:
"""
Check if running in HF Spaces environment
Returns:
bool: True if in HF Spaces
"""
return hf_config.is_hf_spaces
def get_cache_dir() -> str:
"""
Get the appropriate cache directory for the current environment
Returns:
str: Cache directory path
"""
return hf_config.cache_dirs["transformers_cache"]
|