grantjenks/python-diskcache

Features

• Key-Value Persistence Stores - Provides a disk-backed key-value store for persisting mappings, sets, and queues to bypass memory limitations.
• Persistent Key-Value Caches - Provides a persistent key-value store that uses the local filesystem to cache data across threads and processes.
• Disk-Backed Data Structures - Maintains sorted lists, sets, and double-ended queues on the filesystem for permanent storage and fast retrieval.
• Disk-Persistent Data Structures - Implements queues and sorted sets that persist to the local filesystem to bypass memory limitations.
• Expensive Result Caches - Caches expensive function results and temporary data on disk to survive restarts and avoid redundant computations.
• SQLite Storage Adapters - Uses an embedded SQLite database to manage key-value mappings and metadata on the local filesystem.
• Distributed Locks - Coordinates shared resource access across threads and processes using distributed locks, reentrant locks, and semaphores.
• Process Synchronization Locks - Coordinates shared resource access across different OS processes using disk-based spin-locks and semaphores.
• Disk-Serialized Collection Libraries - Provides a collection of disk-serialized sorted lists, dictionaries, and deques that survive program restarts.
• Cross-Process Cache Coordination - Ensures thread-safe and process-safe access to cached data, allowing multiple execution units to read and write without corruption.
• Persistent Double-Ended Queues - Ships disk-serialized double-ended queues that support fast access and modifications at both endpoints.
• Pluggable Binary Serializers - Provides customizable binary conversion handlers and compression to optimize disk space and retrieval speed.
• Cache Eviction Policies - Implements LRU and LFU strategies for removing items when disk storage limits are reached.
• Tag-Based Eviction - Associates metadata tags with entries to enable grouped retrieval and targeted eviction of related items.
• Stampede Mitigation - Prevents redundant backend load by ensuring only one worker regenerates an expired item during concurrency bursts.
• Persistent Mappings - Provides mutable mappings and ordered dictionaries stored on disk that persist independently of process lifecycles.
• Least Recently Used Caches - Automatically removes the least recently used items when disk storage limits are reached.
• Data Sharding - Increases throughput by dividing the cache into independent storage buckets to reduce write contention.
• Custom Serialization Interfaces - Allows the definition of custom binary conversion formats for keys and values via custom disk handlers.
• Sharded Caches - Implements sharded data partitioning to divide the cache into multiple storage buckets, reducing write contention.
• Sorted - Provides key-value storage that maintains sorted keys for efficient range-based lookups and ordered iteration.
• Cache Expiration Management - Controls data lifecycles through time-to-live (TTL) values and expiration timestamp updates.
• Sorted Sets - Provides collections of unique elements that remain sorted on disk for range queries and ranking.
• Atomic Counters - Provides atomic increment and decrement operations on stored integer values to prevent race conditions.
• Disk-Backed Sorted Lists - Implements lists that maintain element order on the filesystem regardless of insertion sequence.
• FIFO Queues - Implements first-in-first-out queues persisted to the local filesystem to ensure task lists survive process restarts.
• Atomic File Writes - Ensures data consistency by grouping filesystem updates into atomic units to prevent corruption during crashes.
• Atomic Transaction Grouping - Groups multiple operations into single atomic units to ensure data consistency and improve write performance.
• Memoization Caches - Caches expensive function results to disk to optimize performance and prevent redundant computation.
• Caching - SQLite and file-backed cache backend.

Star-Verlauf