Skip to main content

nfsdict

NFS-safe persistent dictionary for Python (CPython & PyPy).

One file per key, atomic writes via temp + rename, in-memory read cache so reads never touch NFS after initial load. Corruption of one file loses that key only — never the whole store.

A best-effort key index speeds up init by avoiding expensive directory scans. The index is advisory — exact lookups always fall back to disk, so keys written by other processes are never lost.

Install

pip install nfsdict

Usage

from nfsdict import NfsDict

# default: pickle serializer, ~/.local/share/nfsdict/default/
d = NfsDict()

# custom name and directory
d = NfsDict("my_cache", base_dir="/shared/nfs/cache")

# json serializer (human-readable, but values must be JSON-serializable)
d = NfsDict("my_cache", serializer="json")

# lazy loading: only keys are loaded at init, values fetched on first access
d = NfsDict("my_cache", lazy=True)

# use it like a normal dict
d["key"] = {"nested": [1, 2, 3]}
print(d["key"])
del d["key"]
print(len(d))

# re-read from disk (full scan, call if another process may have written)
d.sync()

# context manager flushes index on exit
with NfsDict("my_cache") as d:
    d["key"] = "value"

How it works

  • Each key is stored as a separate file, named by its SHA-256 hash, inside a two-character prefix subdirectory for filesystem-friendly sharding.
  • Writes are atomic: data is written to a temp file, fsynced, then os.replaced into place.
  • An in-memory cache is populated once at init and updated on every write/delete — so reads are pure dict lookups with zero NFS calls.
  • If a key is not in the cache, __getitem__ falls back to reading its file directly from disk (covers keys written by other processes).
  • A best-effort key index avoids the expensive rglob directory scan on startup. If the index is missing or corrupted, a full scan runs automatically and rebuilds it.
  • Corrupted value files are silently skipped during loading: you lose that one key, not the whole store.

API

NfsDict(name="default", base_dir=None, serializer="pickle", lazy=False)

Parameter Description
name Namespace / subdirectory name
base_dir Storage root. Defaults to platformdirs.user_data_dir
serializer "pickle" or "json"
lazy If True, values are loaded on first access, not at init

Supports the full MutableMapping interface: __getitem__, __setitem__, __delitem__, __contains__, __iter__, __len__, plus:

  • .sync() — full re-read from disk
  • .close() — flush the key index to disk
  • Context manager (with NfsDict(...) as d:)

License

MIT

Metadata

Release files for nfsdict 0.2.2

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for nfsdict 0.2.2
File Size Uploaded Provenance Actions
nfsdict-0.2.2.tar.gz 5.9 kB Without provenance Details for nfsdict-0.2.2.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for nfsdict 0.2.2
File Interpreter ABI Platform Actions
nfsdict-0.2.2-py3-none-any.whl Python 3 none any Details for nfsdict-0.2.2-py3-none-any.whl

Total release size: 12.1 kB

Release files / nfsdict-0.2.2.tar.gz

Download URL nfsdict-0.2.2.tar.gz
Size 5.9 kB
Tags Source
SHA-256 checksum
How to use checksums
8c9bfb7b4877bbdc0eda37ff7ae345a48485c8fe399eeabeb28bc722c3fd5145
BLAKE2b-256 checksum
How to use checksums
43cf26a567016af75dd15b6a706c4456fc5bfa456200bda96cb423e0cff1af56
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.12.10

Release files / nfsdict-0.2.2-py3-none-any.whl

Download URL nfsdict-0.2.2-py3-none-any.whl
Size 6.2 kB
Tags Python 3
SHA-256 checksum
How to use checksums
1c0687389108a645c7789146fbca52fd863495bc5073393ceb9d0d5f0aed0fde
BLAKE2b-256 checksum
How to use checksums
fe88522e5ef860dd019d8ad632671221f7b685489d13387aafd6737fad5223ce
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.12.10

Security for nfsdict 0.2.2

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page