t2m / src /app /core /cache.py
thanhkt's picture
implement core api
50a7bf0
"""
Redis caching strategies and decorators for FastAPI endpoints.
This module provides caching decorators, cache invalidation patterns,
cache warming strategies, and monitoring for the video generation API.
"""
import asyncio
import functools
import hashlib
import json
import logging
from datetime import datetime, timedelta
from typing import Any, Callable, Dict, List, Optional, Union, Tuple
from contextlib import asynccontextmanager
from fastapi import Request, Response
from redis.asyncio import Redis
from .redis import redis_manager, RedisKeyManager, safe_redis_operation
from .config import get_settings
logger = logging.getLogger(__name__)
settings = get_settings()
class CacheConfig:
"""Configuration for caching behavior."""
# Default TTL values (in seconds)
DEFAULT_TTL = 300 # 5 minutes
SHORT_TTL = 60 # 1 minute
MEDIUM_TTL = 900 # 15 minutes
LONG_TTL = 3600 # 1 hour
VERY_LONG_TTL = 86400 # 24 hours
# Cache key prefixes
ENDPOINT_CACHE = "endpoint_cache"
QUERY_CACHE = "query_cache"
USER_CACHE = "user_cache"
SYSTEM_CACHE = "system_cache"
# Cache warming settings
WARM_CACHE_BATCH_SIZE = 10
WARM_CACHE_DELAY = 0.1 # seconds between batch operations
class CacheKeyGenerator:
"""Utility class for generating consistent cache keys."""
@staticmethod
def endpoint_key(
method: str,
path: str,
query_params: Dict[str, Any] = None,
user_id: str = None,
additional_params: Dict[str, Any] = None
) -> str:
"""
Generate cache key for API endpoint responses.
Args:
method: HTTP method
path: Request path
query_params: Query parameters
user_id: User ID for user-specific caching
additional_params: Additional parameters for key generation
Returns:
Generated cache key
"""
key_parts = [method.upper(), path]
if user_id:
key_parts.append(f"user:{user_id}")
if query_params:
# Sort params for consistent key generation
sorted_params = sorted(query_params.items())
params_str = "&".join(f"{k}={v}" for k, v in sorted_params)
key_parts.append(f"params:{params_str}")
if additional_params:
sorted_additional = sorted(additional_params.items())
additional_str = "&".join(f"{k}={v}" for k, v in sorted_additional)
key_parts.append(f"extra:{additional_str}")
# Create hash for long keys
key_string = "|".join(key_parts)
if len(key_string) > 200: # Redis key length limit consideration
key_hash = hashlib.md5(key_string.encode()).hexdigest()
return RedisKeyManager.cache_key(CacheConfig.ENDPOINT_CACHE, key_hash)
return RedisKeyManager.cache_key(CacheConfig.ENDPOINT_CACHE, key_string)
@staticmethod
def query_key(query_name: str, params: Dict[str, Any] = None) -> str:
"""Generate cache key for database queries."""
key_parts = [query_name]
if params:
sorted_params = sorted(params.items())
params_str = "&".join(f"{k}={v}" for k, v in sorted_params)
key_parts.append(params_str)
key_string = "|".join(key_parts)
return RedisKeyManager.cache_key(CacheConfig.QUERY_CACHE, key_string)
@staticmethod
def user_key(user_id: str, data_type: str) -> str:
"""Generate cache key for user-specific data."""
return RedisKeyManager.cache_key(CacheConfig.USER_CACHE, f"{user_id}:{data_type}")
@staticmethod
def system_key(component: str, metric: str = None) -> str:
"""Generate cache key for system data."""
key = component
if metric:
key = f"{component}:{metric}"
return RedisKeyManager.cache_key(CacheConfig.SYSTEM_CACHE, key)
class CacheManager:
"""Advanced cache management with invalidation and warming strategies."""
def __init__(self):
self._cache_stats = {
"hits": 0,
"misses": 0,
"sets": 0,
"deletes": 0,
"invalidations": 0
}
async def get(
self,
key: str,
default: Any = None,
deserialize: bool = True
) -> Any:
"""
Get value from cache with statistics tracking.
Args:
key: Cache key
default: Default value if key doesn't exist
deserialize: Whether to deserialize JSON data
Returns:
Cached value or default
"""
try:
redis_client = redis_manager.redis
value = await redis_client.get(key)
if value is None:
self._cache_stats["misses"] += 1
return default
self._cache_stats["hits"] += 1
if deserialize:
try:
return json.loads(value)
except json.JSONDecodeError:
logger.warning(f"Failed to deserialize cached value for key: {key}")
return default
return value
except Exception as e:
logger.error(f"Cache get failed for key {key}: {e}")
self._cache_stats["misses"] += 1
return default
async def set(
self,
key: str,
value: Any,
ttl: int = CacheConfig.DEFAULT_TTL,
serialize: bool = True
) -> bool:
"""
Set value in cache with TTL.
Args:
key: Cache key
value: Value to cache
ttl: Time to live in seconds
serialize: Whether to serialize value as JSON
Returns:
True if successful
"""
try:
redis_client = redis_manager.redis
if serialize:
value = json.dumps(value, default=str)
result = await redis_client.setex(key, ttl, value)
self._cache_stats["sets"] += 1
return result
except Exception as e:
logger.error(f"Cache set failed for key {key}: {e}")
return False
async def delete(self, key: str) -> bool:
"""Delete key from cache."""
try:
redis_client = redis_manager.redis
result = await redis_client.delete(key)
self._cache_stats["deletes"] += 1
return bool(result)
except Exception as e:
logger.error(f"Cache delete failed for key {key}: {e}")
return False
async def delete_pattern(self, pattern: str) -> int:
"""
Delete all keys matching pattern.
Args:
pattern: Redis key pattern (supports wildcards)
Returns:
Number of keys deleted
"""
try:
redis_client = redis_manager.redis
keys = await redis_client.keys(pattern)
if not keys:
return 0
deleted = await redis_client.delete(*keys)
self._cache_stats["deletes"] += deleted
return deleted
except Exception as e:
logger.error(f"Cache pattern delete failed for pattern {pattern}: {e}")
return 0
async def invalidate_user_cache(self, user_id: str) -> int:
"""Invalidate all cache entries for a specific user."""
pattern = RedisKeyManager.cache_key(CacheConfig.USER_CACHE, f"{user_id}:*")
deleted = await self.delete_pattern(pattern)
self._cache_stats["invalidations"] += 1
logger.info(f"Invalidated {deleted} cache entries for user {user_id}")
return deleted
async def invalidate_endpoint_cache(self, path_pattern: str) -> int:
"""Invalidate cache entries for specific endpoint patterns."""
pattern = RedisKeyManager.cache_key(CacheConfig.ENDPOINT_CACHE, f"*{path_pattern}*")
deleted = await self.delete_pattern(pattern)
self._cache_stats["invalidations"] += 1
logger.info(f"Invalidated {deleted} cache entries for pattern {path_pattern}")
return deleted
async def warm_cache(
self,
warm_functions: List[Tuple[Callable, Dict[str, Any]]],
batch_size: int = CacheConfig.WARM_CACHE_BATCH_SIZE
) -> Dict[str, Any]:
"""
Warm cache with predefined data.
Args:
warm_functions: List of (function, kwargs) tuples to execute
batch_size: Number of operations per batch
Returns:
Warming results and statistics
"""
results = {
"total_functions": len(warm_functions),
"successful": 0,
"failed": 0,
"errors": []
}
for i in range(0, len(warm_functions), batch_size):
batch = warm_functions[i:i + batch_size]
for func, kwargs in batch:
try:
await func(**kwargs)
results["successful"] += 1
except Exception as e:
results["failed"] += 1
results["errors"].append({
"function": func.__name__,
"error": str(e)
})
logger.error(f"Cache warming failed for {func.__name__}: {e}")
# Small delay between batches to avoid overwhelming the system
if i + batch_size < len(warm_functions):
await asyncio.sleep(CacheConfig.WARM_CACHE_DELAY)
logger.info(f"Cache warming completed: {results}")
return results
def get_stats(self) -> Dict[str, Any]:
"""Get cache statistics."""
total_operations = sum(self._cache_stats.values())
hit_rate = (
self._cache_stats["hits"] / (self._cache_stats["hits"] + self._cache_stats["misses"])
if (self._cache_stats["hits"] + self._cache_stats["misses"]) > 0
else 0
)
return {
**self._cache_stats,
"total_operations": total_operations,
"hit_rate": round(hit_rate * 100, 2),
"timestamp": datetime.utcnow().isoformat()
}
def reset_stats(self) -> None:
"""Reset cache statistics."""
self._cache_stats = {
"hits": 0,
"misses": 0,
"sets": 0,
"deletes": 0,
"invalidations": 0
}
# Global cache manager instance
cache_manager = CacheManager()
def cache_response(
ttl: int = CacheConfig.DEFAULT_TTL,
key_generator: Optional[Callable] = None,
user_specific: bool = False,
skip_cache_header: str = "X-Skip-Cache",
vary_on: List[str] = None
):
"""
Decorator for caching FastAPI endpoint responses.
Args:
ttl: Time to live in seconds
key_generator: Custom key generation function
user_specific: Whether to include user ID in cache key
skip_cache_header: Header name to skip cache
vary_on: List of headers/params to vary cache on
Usage:
@router.get("/api/v1/jobs")
@cache_response(ttl=300, user_specific=True)
async def get_jobs(request: Request, user_id: str = Depends(get_current_user_id)):
return await job_service.get_jobs(user_id)
"""
def decorator(func: Callable) -> Callable:
@functools.wraps(func)
async def wrapper(*args, **kwargs):
# Extract request object
request = None
for arg in args:
if isinstance(arg, Request):
request = arg
break
if not request:
# If no request object, execute function without caching
return await func(*args, **kwargs)
# Check if cache should be skipped
if request.headers.get(skip_cache_header):
return await func(*args, **kwargs)
# Generate cache key
if key_generator:
cache_key = await key_generator(request, *args, **kwargs)
else:
user_id = None
if user_specific:
# Try to extract user_id from kwargs or dependencies
user_id = kwargs.get("user_id") or kwargs.get("current_user_id")
vary_params = {}
if vary_on:
for header in vary_on:
if header in request.headers:
vary_params[header] = request.headers[header]
cache_key = CacheKeyGenerator.endpoint_key(
method=request.method,
path=request.url.path,
query_params=dict(request.query_params),
user_id=user_id,
additional_params=vary_params
)
# Try to get from cache
cached_result = await cache_manager.get(cache_key)
if cached_result is not None:
logger.debug(f"Cache hit for key: {cache_key}")
return cached_result
# Execute function and cache result
result = await func(*args, **kwargs)
# Cache the result
await cache_manager.set(cache_key, result, ttl)
logger.debug(f"Cached result for key: {cache_key}")
return result
return wrapper
return decorator
def cache_query(
ttl: int = CacheConfig.DEFAULT_TTL,
key_prefix: str = "query"
):
"""
Decorator for caching database query results.
Args:
ttl: Time to live in seconds
key_prefix: Prefix for cache key
Usage:
@cache_query(ttl=600, key_prefix="user_jobs")
async def get_user_jobs(user_id: str, status: str = None):
# Database query logic
return results
"""
def decorator(func: Callable) -> Callable:
@functools.wraps(func)
async def wrapper(*args, **kwargs):
# Generate cache key from function name and parameters
cache_key = CacheKeyGenerator.query_key(
f"{key_prefix}:{func.__name__}",
{**dict(zip(func.__code__.co_varnames, args)), **kwargs}
)
# Try to get from cache
cached_result = await cache_manager.get(cache_key)
if cached_result is not None:
logger.debug(f"Query cache hit for key: {cache_key}")
return cached_result
# Execute function and cache result
result = await func(*args, **kwargs)
# Cache the result
await cache_manager.set(cache_key, result, ttl)
logger.debug(f"Cached query result for key: {cache_key}")
return result
return wrapper
return decorator
class CacheInvalidationManager:
"""Manages cache invalidation patterns and strategies."""
@staticmethod
async def invalidate_job_related_cache(job_id: str, user_id: str = None):
"""Invalidate all cache entries related to a specific job."""
patterns_to_invalidate = [
f"*jobs*{job_id}*",
f"*job_status*{job_id}*",
f"*videos*{job_id}*"
]
if user_id:
patterns_to_invalidate.extend([
f"*user:{user_id}*jobs*",
f"*user_jobs*{user_id}*"
])
total_deleted = 0
for pattern in patterns_to_invalidate:
deleted = await cache_manager.delete_pattern(pattern)
total_deleted += deleted
logger.info(f"Invalidated {total_deleted} cache entries for job {job_id}")
return total_deleted
@staticmethod
async def invalidate_user_related_cache(user_id: str):
"""Invalidate all cache entries related to a specific user."""
return await cache_manager.invalidate_user_cache(user_id)
@staticmethod
async def invalidate_system_cache():
"""Invalidate system-wide cache entries."""
patterns = [
f"*{CacheConfig.SYSTEM_CACHE}*",
"*health*",
"*metrics*",
"*queue*"
]
total_deleted = 0
for pattern in patterns:
deleted = await cache_manager.delete_pattern(pattern)
total_deleted += deleted
logger.info(f"Invalidated {total_deleted} system cache entries")
return total_deleted
# Cache warming functions
async def warm_common_queries():
"""Warm cache with commonly accessed data."""
warming_functions = [
# Add warming functions here as they're implemented
# (get_system_health, {}),
# (get_queue_status, {}),
# (get_popular_jobs, {"limit": 10}),
]
if warming_functions:
return await cache_manager.warm_cache(warming_functions)
return {"message": "No warming functions configured"}
# Utility functions
async def get_cache_info() -> Dict[str, Any]:
"""Get comprehensive cache information and statistics."""
try:
redis_client = redis_manager.redis
redis_info = await redis_client.info("memory")
return {
"cache_stats": cache_manager.get_stats(),
"redis_memory": {
"used_memory": redis_info.get("used_memory_human", "unknown"),
"used_memory_peak": redis_info.get("used_memory_peak_human", "unknown"),
"memory_fragmentation_ratio": redis_info.get("memory_fragmentation_ratio", 0)
},
"connection_info": await redis_manager.get_connection_info(),
"timestamp": datetime.utcnow().isoformat()
}
except Exception as e:
logger.error(f"Failed to get cache info: {e}")
return {
"error": str(e),
"cache_stats": cache_manager.get_stats(),
"timestamp": datetime.utcnow().isoformat()
}