SciQLop Cache

A fast, persistent key-value cache for Python and C++20. Built on SQLite with hybrid blob/file storage, it stays flat from 100 to 1M+ entries while being 2x faster than diskcache on writes and up to 6x faster in batched transactions.

pip install sciqlop-cache

Quick Start (Python)

from pysciqlop_cache import Cache

cache = Cache("/tmp/my-cache")
cache["sensor/temperature"] = {"ts": 1710000000, "values": [21.3, 21.5, 21.4]}
print(cache["sensor/temperature"])

Any picklable Python object works out of the box. For better performance with structured data, use the msgspec serializer:

from pysciqlop_cache import Cache, MsgspecSerializer

cache = Cache("/tmp/my-cache", serializer=MsgspecSerializer())

Expiration and Tags

cache.set("session/abc", token, expire=3600)          # expires in 1 hour
cache.set("sensor/temp", data, tag="sensor")           # tagged entry
cache.set("sensor/hum", data, expire=600, tag="sensor")

cache.evict_tag("sensor")   # bulk-remove all "sensor" entries
cache.expire()               # remove all expired entries

LRU Eviction

cache = Cache("/tmp/bounded", max_size=1_000_000_000)  # 1 GB limit
# Least-recently-used entries are evicted automatically in the background

Atomic Transactions

with cache.transact():
    cache["balance"] = cache.get("balance", 0) - amount
    cache["log"] = f"withdrew {amount}"
# Automatically commits on success, rolls back on exception

Sharded Concurrency with FanoutCache

For write-heavy concurrent workloads, FanoutCache shards keys across N independent stores:

from pysciqlop_cache import FanoutCache

cache = FanoutCache("/tmp/sharded", shard_count=8, max_size=1_000_000_000)
cache["key"] = "value"         # routed to shard via hash(key) % 8

with cache.transact("key"):   # transaction scoped to key's shard
    cache["key"] = transform(cache["key"])

Lightweight Index

Index and FanoutIndex are bare key-value stores with no expiration, eviction, tags, or stats overhead:

from pysciqlop_cache import Index

idx = Index("/tmp/my-index")
idx["dataset/v2"] = metadata

Distributed Locking

# Cross-process lock backed by atomic add()
with cache.lock("my-resource", expire=30):
    # exclusive access across processes
    do_work()

Memoization

@cache.memoize(expire=300, tag="compute")
def expensive(x, y):
    return heavy_computation(x, y)

expensive(1, 2)  # computed
expensive(1, 2)  # served from cache

Atomic Counters

cache.incr("page_views")             # 1
cache.incr("page_views", delta=5)    # 6
cache.decr("page_views")             # 5

Hit/Miss Statistics

cache = Cache("/tmp/stats")
cache.set("k", "v")
cache.get("k")          # hit
cache.get("missing")    # miss
print(cache.stats())    # {"hits": 1, "misses": 1}
cache.reset_stats()

Integrity Checking

result = cache.check()       # verify structural integrity
result = cache.check(True)   # verify and fix (orphaned files, counter drift)
print(result.ok, result.orphaned_files, result.dangling_rows)

Disk Usage

print(cache.volume())  # total bytes on disk (SQLite DB + file-backed values)
print(cache.size())    # total bytes of stored values only

Dict-like Interface

All store types support the standard Python dict interface:

cache["key"] = value           # set
value = cache["key"]           # get
del cache["key"]               # delete
"key" in cache                 # exists
for key in cache: ...          # iterate keys
len(cache)                     # count

Migrating from diskcache

# Migrate an existing diskcache Cache or FanoutCache
python -m pysciqlop_cache.migrate /old/diskcache /new/sciqlop-cache

# Migrate a diskcache Index
python -m pysciqlop_cache.migrate --type index /old/index /new/index

# Migrate and delete entries from source as they are copied
python -m pysciqlop_cache.migrate --drop /old/diskcache /new/sciqlop-cache

Auto-detects Cache vs FanoutCache, preserves expiration TTLs and tags. Use --type index for Index sources (creates a lightweight sciqlop-cache Index). Also usable as a library:

from pysciqlop_cache.migrate import migrate
result = migrate("/old/cache", "/new/cache", drop=True)
result = migrate("/old/index", "/new/index", store_type="index")
print(result)  # {"migrated": 1234, "skipped": 0, "errors": 0, "elapsed_secs": 1.5}

C++ API

#include "sciqlop_cache/sciqlop_cache.hpp"

// Full-featured cache with expiration, LRU eviction, tags, and stats
Cache cache(".cache/", /*max_size=*/1'000'000'000);

cache.set("key", std::vector<char>{'a', 'b', 'c'});
cache.set("key", data, 60s);                  // with expiration
cache.set("key", data, "mytag");               // with tag
cache.set("key", data, 60s, "mytag");          // both

auto value = cache.get("key");                 // std::optional<Buffer>
cache.del("key");
cache.pop("key");                              // get + delete
cache.add("key", data);                        // set only if absent
cache.touch("key", 120s);                      // update expiration
cache.evict_tag("mytag");                      // bulk remove by tag

// Bare key-value store (no expiration/eviction/tags overhead)
Index index(".index/");

// Atomic counters
cache.incr("counter", 1, /*default=*/0);
cache.decr("counter");

// Integrity and diagnostics
auto cr = cache.check();         // structural integrity
auto cr2 = cache.check(true);    // verify and fix
cache.volume();                  // total disk usage in bytes
cache.stats();                   // {hits, misses}

// Sharded variants for write concurrency
FanoutCache fc(".fc/", /*shard_count=*/8, /*max_size=*/0);
FanoutIndex fi(".fi/", /*shard_count=*/8);

Concurrency

• Thread-safe: per-instance mutex + per-instance SQLite connection. Multiple threads can share a single Cache instance.
• Multi-process safe: SQLite WAL mode with 600s busy timeout. Multiple processes can open the same cache directory.
• FanoutCache/FanoutIndex: shard keys across N independent stores for write concurrency. Each shard has its own database and lock.

Performance

Latency scaling (100 to 1M entries, 256-byte values)

Latency vs value size (64B to 1MB)

Batched transactions (per-op cost)

Amortized per-op latency drops significantly with larger batches, especially for small values:

Reproduce

# Scaling benchmarks
PYTHONPATH=build python benchmark/scaling.py --max-entries 1000000 --backend both > results.csv
python benchmark/plot_scaling.py results.csv -o benchmark/scaling_chart.png

PYTHONPATH=build python benchmark/scaling.py --max-entries 1000000 --raw --backend both > raw.csv
python benchmark/plot_scaling.py raw.csv --violin -o benchmark/scaling_violin.png

# Value-size and batch benchmarks
PYTHONPATH=build python benchmark/bench_valuesize.py > benchmark/valuesize_results.csv
python benchmark/plot_valuesize.py benchmark/valuesize_results.csv -o benchmark

Building from Source

pip install meson-python numpy
meson setup build -Dwith_tests=true
meson compile -C build
meson test -C build

License

MIT