Skip to main content

Fast YAML parser for Python with C++/SIMD backend

Project description

pyfastyaml

PyPI version

pyfastyaml — Fast YAML parser for Python with C++/SIMD backend. 12–18× faster than PyYAML (CSafeLoader/libyaml), 100–170× faster than PyYAML (SafeLoader), 150–220× faster than ruamel.yaml. API compatible with PyYAML for typical config use cases (Kubernetes, Ansible, CI, Docker Compose).

Features

  • High performance: C++17 implementation with AVX2 SIMD on x86 — orders of magnitude faster than pure-Python parsers
  • PyYAML compatible API: loads(s), load(fp), load_all(s), dump(), dump_all() — drop-in replacement for config loading
  • Types: dict, list, str, int, float, bool, null, nested structures
  • Indentation: 2-space (preferred), 4-space, tabs
  • Flow syntax: [a, b, c], {a: 1, b: 2} (incl. nested)
  • Multiline strings: | (literal), > (folded)
  • Anchors and aliases: &anchor, *alias
  • Merge key: <<: *anchor and <<: [*a, *b]
  • Document markers: ---, ..., %YAML directive

Installation

pip install pyfastyaml

Requires Python 3.10+ and a C++ compiler with C++17 support.

Usage

import pyfastyaml

# Parse string
data = pyfastyaml.loads("""
key: value
nested:
  inner: 42
items:
  - a
  - b
""")
# {'key': 'value', 'nested': {'inner': 42}, 'items': ['a', 'b']}

# Parse file (path or file-like object)
data = pyfastyaml.load("config.yaml")

# Serialize to YAML
yaml_str = pyfastyaml.dump(data)
pyfastyaml.dump(data, stream=open("out.yaml", "w"))

API matches PyYAML for common cases:

# Drop-in replacement
import pyfastyaml as yaml  # instead of import yaml
data = yaml.loads(yaml_string)
data = yaml.load("config.yaml")

Benchmark

Parsing time (μs, lower is better):

Size pyfastyaml PyYAML (C) PyYAML ruamel
Small (~200 B) ~2.3 ~32 ~229 ~376
Medium (~1.5 KB) ~11 ~134 ~1,014 ~1,689
Large (~15 KB) ~83 ~1,483 ~11,169 ~17,954
Realworld ~6.4 ~98 ~833 ~1,388

PyYAML (C) = CSafeLoader (libyaml extension). PyYAML = SafeLoader (pure Python).

Run benchmarks:

pip install pytest-benchmark PyYAML ruamel.yaml
pytest tests/test_benchmark.py -v --benchmark-only

PyYAML Compatibility

pyfastyaml produces the same result as yaml.safe_load() for block-style YAML typical of configs. Compatibility tests included in tests/test_pyyaml_compat.py.

Serialization (dump)

# Single document
yaml_str = pyfastyaml.dump({"a": 1, "b": [1, 2, 3]})

# Write to file (path or stream)
pyfastyaml.dump(data, "output.yaml")
pyfastyaml.dump(data, stream=open("out.yaml", "w"))

# Deterministic output (sorted keys)
yaml_str = pyfastyaml.dump(data, sort_keys=True)

# Multiple documents
yaml_str = pyfastyaml.dump_all([{"a": 1}, {"b": 2}])

Multi-document streams

docs = pyfastyaml.load_all("""
---
a: 1
---
b: 2
""")
# [{'a': 1}, {'b': 2}]

Limitations

Not yet supported:

  • Custom tags
  • Complex flow keys (? syntax)

Development

pip install -e ".[dev]"
python setup.py build_ext --inplace
pytest tests/

Benchmark commands:

pytest tests/test_benchmark.py -v --benchmark-only
pytest tests/test_benchmark.py -v --benchmark-only --benchmark-autosave
pytest tests/test_benchmark.py -v --benchmark-only --benchmark-compare

Publishing (maintainers)

pip install build twine
python -m build
python -m twine upload dist/*

License

MIT

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pyfastyaml-0.1.0b5.tar.gz (27.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

pyfastyaml-0.1.0b5-cp312-cp312-win_amd64.whl (90.9 kB view details)

Uploaded CPython 3.12Windows x86-64

File details

Details for the file pyfastyaml-0.1.0b5.tar.gz.

File metadata

  • Download URL: pyfastyaml-0.1.0b5.tar.gz
  • Upload date:
  • Size: 27.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.0

File hashes

Hashes for pyfastyaml-0.1.0b5.tar.gz
Algorithm Hash digest
SHA256 e6d7403b1943997b44906269e10d3759ca1d82a4a2be9ffcd2c48c97db5b4ef7
MD5 18085264c1771cedd577e2e5b1ef1147
BLAKE2b-256 0aebeaabebc1b0141be1d2024c869ff365c31683587ac1c11de8130d09dedf26

See more details on using hashes here.

File details

Details for the file pyfastyaml-0.1.0b5-cp312-cp312-win_amd64.whl.

File metadata

File hashes

Hashes for pyfastyaml-0.1.0b5-cp312-cp312-win_amd64.whl
Algorithm Hash digest
SHA256 e4722ddde735bab26c7e8e238bc2b4b3b66eb40c7b803060d48f38490757922d
MD5 509724c0fbaa5cd3898706ca871604ad
BLAKE2b-256 fe78c8552724c7c442764035e0620ce8470596b166bf96406f50b3103489482a

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page