Spaces:
Sleeping
Sleeping
File size: 15,968 Bytes
1bd7131 b49d66f 1bd7131 7b9d520 ec0e527 7b9d520 1bd7131 b49d66f 1bd7131 16eca32 ec0e527 16eca32 1bd7131 16eca32 1bd7131 7b9d520 1bd7131 7b9d520 1bd7131 8c4055f 1bd7131 8c4055f 1bd7131 8c4055f 1bd7131 8c4055f 1bd7131 8c4055f 1bd7131 8c4055f 1bd7131 8c4055f 1bd7131 8c4055f 1bd7131 16eca32 ec0e527 16eca32 1bd7131 16eca32 ec0e527 16eca32 1bd7131 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 |
"""
Gemini AI Service for image and video generation.
Python port of the TypeScript geminiService.ts
Uses server-side API key from environment.
"""
import asyncio
import logging
import os
import uuid
import httpx
from typing import Optional, Literal
from google import genai
from google.genai import types
logger = logging.getLogger(__name__)
# Model names - easily configurable
MODELS = {
"text_generation": "gemini-2.5-flash",
"image_edit": "gemini-2.5-flash-image",
"video_generation": "veo-3.1-generate-preview"
}
# Type aliases
AspectRatio = Literal["16:9", "9:16"]
Resolution = Literal["720p", "1080p"]
# Video downloads directory
DOWNLOADS_DIR = os.path.join(os.path.dirname(os.path.dirname(__file__)), "downloads")
# Ensure downloads directory exists
os.makedirs(DOWNLOADS_DIR, exist_ok=True)
# Mock mode for local testing (set GEMINI_MOCK_MODE=true to skip real API calls)
MOCK_MODE = os.getenv("GEMINI_MOCK_MODE", "false").lower() == "true"
MOCK_MODE_SLEEP_TIME = os.getenv("GEMINI_MOCK_MODE_SLEEP_TIME", "0.5")
# Sample video URL for mock mode (a public test video)
MOCK_VIDEO_URL = "https://video.twimg.com/amplify_video/1994083297756848128/vid/avc1/576x576/ue31qU0xts8L9tXD.mp4?tag=21"
# Concurrency limits from environment (defaults)
MAX_CONCURRENT_VIDEOS = int(os.getenv("MAX_CONCURRENT_VIDEOS", "2"))
MAX_CONCURRENT_IMAGES = int(os.getenv("MAX_CONCURRENT_IMAGES", "5"))
MAX_CONCURRENT_TEXT = int(os.getenv("MAX_CONCURRENT_TEXT", "10"))
# Semaphores for concurrency control
_video_semaphore: Optional[asyncio.Semaphore] = None
_image_semaphore: Optional[asyncio.Semaphore] = None
_text_semaphore: Optional[asyncio.Semaphore] = None
def get_video_semaphore() -> asyncio.Semaphore:
"""Get or create video semaphore."""
global _video_semaphore
if _video_semaphore is None:
_video_semaphore = asyncio.Semaphore(MAX_CONCURRENT_VIDEOS)
logger.info(f"Video semaphore initialized with limit: {MAX_CONCURRENT_VIDEOS}")
return _video_semaphore
def get_image_semaphore() -> asyncio.Semaphore:
"""Get or create image semaphore."""
global _image_semaphore
if _image_semaphore is None:
_image_semaphore = asyncio.Semaphore(MAX_CONCURRENT_IMAGES)
logger.info(f"Image semaphore initialized with limit: {MAX_CONCURRENT_IMAGES}")
return _image_semaphore
def get_text_semaphore() -> asyncio.Semaphore:
"""Get or create text semaphore."""
global _text_semaphore
if _text_semaphore is None:
_text_semaphore = asyncio.Semaphore(MAX_CONCURRENT_TEXT)
logger.info(f"Text semaphore initialized with limit: {MAX_CONCURRENT_TEXT}")
return _text_semaphore
def get_gemini_api_key() -> str:
"""Get Gemini API key from environment."""
api_key = os.getenv("GEMINI_API_KEY")
if not api_key:
raise ValueError("Server Authentication Error with GEMINI")
return api_key
class GeminiService:
"""
Gemini AI Service for text, image, and video generation.
Uses server-side API key from environment.
"""
def __init__(self, api_key: Optional[str] = None):
"""Initialize the Gemini client with API key from env or provided."""
self.api_key = api_key or get_gemini_api_key()
self.client = genai.Client(api_key=self.api_key)
def _handle_api_error(self, error: Exception, context: str):
"""Handle API errors with descriptive messages."""
msg = str(error)
if "404" in msg or "NOT_FOUND" in msg or "Requested entity was not found" in msg or "[5," in msg:
raise ValueError(
f"Model not found ({context}). Ensure your API key project has access to this model. "
"Veo requires a paid account."
)
raise error
async def generate_animation_prompt(
self,
base64_image: str,
mime_type: str,
custom_prompt: Optional[str] = None
) -> str:
"""
Analyzes the image to generate a suitable animation prompt.
"""
# Mock mode for testing
if MOCK_MODE:
logger.info("[MOCK MODE] Generating animation prompt")
await asyncio.sleep(GEMINI_MOCK_MODE_SLEEP_TIME) # Simulate API delay
return "A gentle breeze rustles through the scene as soft light dances across the surface. The camera slowly zooms in with a subtle parallax effect, creating depth and movement."
default_prompt = custom_prompt or "Describe how this image could be subtly animated with cinematic movement."
async with get_text_semaphore():
try:
response = await asyncio.to_thread(
self.client.models.generate_content,
model=MODELS["text_generation"],
contents=types.Content(
parts=[
types.Part.from_bytes(
data=base64_image,
mime_type=mime_type
),
types.Part.from_text(text=default_prompt)
]
)
)
return response.text or "Cinematic subtle movement"
except Exception as error:
self._handle_api_error(error, MODELS["text_generation"])
async def edit_image(
self,
base64_image: str,
mime_type: str,
prompt: str
) -> str:
"""
Edit an image using Gemini image model.
Returns base64 data URI of the edited image.
"""
# Mock mode for testing - return a sample image
if MOCK_MODE:
logger.info(f"[MOCK MODE] Editing image with prompt: {prompt}")
await asyncio.sleep(1) # Simulate API delay
# Return a small red placeholder image (1x1 pixel)
return "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mP8z8DwHwAFBQIAX8jx0gAAAABJRU5ErkJggg=="
async with get_image_semaphore():
try:
response = await asyncio.to_thread(
self.client.models.generate_content,
model=MODELS["image_edit"],
contents=types.Content(
parts=[
types.Part.from_bytes(
data=base64_image,
mime_type=mime_type
),
types.Part.from_text(text=prompt or "Enhance this image")
]
)
)
candidates = response.candidates
if not candidates:
raise ValueError("No candidates returned from Gemini.")
for part in candidates[0].content.parts:
if hasattr(part, 'inline_data') and part.inline_data and part.inline_data.data:
result_mime = part.inline_data.mime_type or 'image/png'
return f"data:{result_mime};base64,{part.inline_data.data}"
raise ValueError("No image data found in the response.")
except Exception as error:
self._handle_api_error(error, MODELS["image_edit"])
async def start_video_generation(
self,
base64_image: str,
mime_type: str,
prompt: str,
aspect_ratio: AspectRatio = "16:9",
resolution: Resolution = "720p",
number_of_videos: int = 1
) -> dict:
"""
Start video generation using Veo model.
Returns operation details for polling.
"""
# Mock mode for testing without API credits
if MOCK_MODE:
import uuid
mock_operation_name = f"mock_operation_{uuid.uuid4().hex[:16]}"
logger.info(f"[MOCK MODE] Starting video generation: {mock_operation_name}")
return {
"gemini_operation_name": mock_operation_name,
"done": False,
"status": "pending"
}
async with get_video_semaphore():
try:
# Start video generation
operation = await asyncio.to_thread(
self.client.models.generate_videos,
model=MODELS["video_generation"],
prompt=prompt,
image=types.Image(
image_bytes=base64_image,
mime_type=mime_type
),
config=types.GenerateVideosConfig(
number_of_videos=number_of_videos,
resolution=resolution,
aspect_ratio=aspect_ratio
)
)
# Return operation details
return {
"gemini_operation_name": operation.name,
"done": operation.done,
"status": "completed" if operation.done else "pending"
}
except Exception as error:
self._handle_api_error(error, MODELS["video_generation"])
async def check_video_status(self, gemini_operation_name: str) -> dict:
"""
Check the status of a video generation operation.
Returns status and video URL if complete.
"""
# Mock mode for testing without API credits
if MOCK_MODE:
# Simulate processing time: complete after 2 checks (track via a simple mechanism)
# For simplicity, always return completed with mock video URL
logger.info(f"[MOCK MODE] Checking video status: {gemini_operation_name}")
await asyncio.sleep(2) # Simulate API delay
return {
"gemini_operation_name": gemini_operation_name,
"done": True,
"status": "completed",
"video_url": MOCK_VIDEO_URL
}
try:
# Get operation status using the operation object
# First, we need to recreate the operation from the name
from google.genai.types import GenerateVideosOperation
operation = await asyncio.to_thread(
self.client.operations.get,
GenerateVideosOperation(name=gemini_operation_name, done=False)
)
if not operation.done:
return {
"gemini_operation_name": gemini_operation_name,
"done": False,
"status": "pending"
}
# Check for error - handle both string and object types
if operation.error:
error_msg = operation.error
if hasattr(operation.error, 'message'):
error_msg = operation.error.message
return {
"gemini_operation_name": gemini_operation_name,
"done": True,
"status": "failed",
"error": str(error_msg) or "Unknown error"
}
# Extract video URI from result
result = operation.result
if result and hasattr(result, 'generated_videos') and result.generated_videos:
video = result.generated_videos[0]
if hasattr(video, 'video') and video.video and hasattr(video.video, 'uri'):
video_uri = video.video.uri
return {
"gemini_operation_name": gemini_operation_name,
"done": True,
"status": "completed",
"video_url": f"{video_uri}&key={self.api_key}"
}
return {
"gemini_operation_name": gemini_operation_name,
"done": True,
"status": "failed",
"error": "No video URI returned. May be due to safety filters."
}
except Exception as error:
msg = str(error)
if "404" in msg or "NOT_FOUND" in msg or "Requested entity was not found" in msg:
return {
"gemini_operation_name": gemini_operation_name,
"done": True,
"status": "failed",
"error": "Operation not found (404). It may have expired."
}
raise error
async def download_video(self, video_url: str, operation_id: str) -> str:
"""
Download video from Gemini to local storage.
Returns the local filename.
"""
filename = f"{operation_id}.mp4"
filepath = os.path.join(DOWNLOADS_DIR, filename)
try:
# follow_redirects=True is required as Gemini returns 302 redirects
async with httpx.AsyncClient(timeout=120.0, follow_redirects=True) as client:
response = await client.get(video_url)
response.raise_for_status()
with open(filepath, 'wb') as f:
f.write(response.content)
logger.info(f"Downloaded video to {filepath}")
return filename
except Exception as e:
logger.error(f"Failed to download video: {e}")
raise ValueError(f"Failed to download video: {e}")
async def generate_text(
self,
prompt: str,
model: Optional[str] = None
) -> str:
"""
Simple text generation with Gemini.
"""
# Mock mode for testing
if MOCK_MODE:
logger.info(f"[MOCK MODE] Generating text for prompt: {prompt[:50]}...")
await asyncio.sleep(MOCK_MODE_SLEEP_TIME) # Simulate API delay
return f"This is a mock response for your prompt: '{prompt[:100]}...'. In production, this would be generated by Gemini AI."
model_name = model or MODELS["text_generation"]
async with get_text_semaphore():
try:
response = await asyncio.to_thread(
self.client.models.generate_content,
model=model_name,
contents=types.Content(
parts=[types.Part.from_text(text=prompt)]
)
)
return response.text or ""
except Exception as error:
self._handle_api_error(error, model_name)
async def analyze_image(
self,
base64_image: str,
mime_type: str,
prompt: str
) -> str:
"""
Analyze image with custom prompt.
"""
# Mock mode for testing
if MOCK_MODE:
logger.info(f"[MOCK MODE] Analyzing image with prompt: {prompt[:50]}...")
await asyncio.sleep(MOCK_MODE_SLEEP_TIME) # Simulate API delay
return f"Mock analysis result: The image appears to show a scene that matches your query '{prompt[:50]}...'. This is placeholder content for testing."
async with get_text_semaphore():
try:
response = await asyncio.to_thread(
self.client.models.generate_content,
model=MODELS["text_generation"],
contents=types.Content(
parts=[
types.Part.from_bytes(
data=base64_image,
mime_type=mime_type
),
types.Part.from_text(text=prompt)
]
)
)
return response.text or ""
except Exception as error:
self._handle_api_error(error, MODELS["text_generation"])
|