compose-spec 0.2.1

Python bindings for compose_spec — validated Docker Compose file parsing and serialization

compose-spec

Python bindings for the compose_spec Rust library — a fully validated, spec-compliant Docker Compose file parser and serializer.

Installation

pip install compose-spec

For development:

pip install maturin
maturin develop --release

Requires Rust toolchain (rustup) and Python >= 3.8.

Quick Start

from compose_spec import PyCompose

yaml = """
services:
  web:
    image: nginx:latest
    ports:
      - 80:80
    environment:
      FOO: bar
volumes:
  data:
"""

c = PyCompose.from_yaml(yaml)
print(c.service_names())  # ['web']
print(c.volumes)          # {'data': None}

# Round-trip back to YAML
print(c.to_yaml())

# Convert to Python dict
d = c.to_dict()

API Reference

PyCompose

The main class representing a parsed Docker Compose file.

Constructors

Method	Description
PyCompose.from_yaml(yaml_str)	Parse from a YAML string
PyCompose.from_json(json_str)	Parse from a JSON string
PyCompose.from_dict(d)	Parse from a Python dict

All constructors raise ValueError on invalid input.

Serialization

Method	Returns	Description
to_yaml()	str	Serialize to YAML
to_json()	str	Serialize to JSON
to_dict()	dict	Serialize to a Python dict

Properties (getters and setters)

All top-level Compose fields are exposed as Python properties. Getting a property returns a native Python object (dict, list, str, or None). Setting a property accepts the same types.

Property	Type	Description
name	str | None	Project name
version	str | None	Compose file version (legacy)
services	dict	Service definitions
networks	dict	Network definitions
volumes	dict	Volume definitions
configs	dict	Config definitions
secrets	dict	Secret definitions
include	list	Included sub-projects
extensions	dict	Extension values (x-* keys)

c.name = "my-project"
print(c.services)        # {'web': {'image': 'nginx:latest', ...}}
c.networks = {"frontend": {"driver": "bridge"}}

Validation

Checks that all resource references (networks, volumes, configs, secrets) used in services are defined at the top level. Raises ValueError with a descriptive message if validation fails.

Method	Description
validate()	Run all validation checks
validate_networks()	Check network references
validate_volumes()	Check volume references
validate_configs()	Check config references
validate_secrets()	Check secret references

c = PyCompose.from_yaml("""
services:
  web:
    image: nginx
    networks:
      - missing_net
""")
try:
    c.validate()
except ValueError as e:
    print(e)  # network "missing_net" is not defined

Service convenience methods

Method	Returns	Description
service_names()	list[str]	All service names in order
get_service(name)	dict | None	Get a single service by name
set_service(name, service_dict)	None	Add or replace a service
remove_service(name)	dict | None	Remove and return a service

web = c.get_service("web")
web["image"] = "nginx:alpine"
c.set_service("web", web)

c.set_service("redis", {"image": "redis:7"})
removed = c.remove_service("redis")

Dunder methods

Expression	Description
len(c)	Number of services
"web" in c	Check if a service exists
c1 == c2	Structural equality
repr(c)	String representation

---

PyOptions

Options builder for controlling YAML parsing behavior.

from compose_spec import PyOptions

opts = PyOptions(apply_merge=True)
# or:
opts = PyOptions()
opts.apply_merge = True

c = opts.from_yaml(yaml_string)

Constructor

PyOptions(apply_merge=False)

Properties

Property	Type	Default	Description
apply_merge	bool	False	Whether to merge YAML << keys

Methods

Method	Returns	Description
from_yaml(yaml_str)	PyCompose	Parse YAML with the configured options

When to use apply_merge: If your Compose file uses YAML anchors and the merge key (<<) to share configuration between services, set apply_merge=True to resolve them during parsing.

yaml = """
services:
  base:
    environment: &defaults
      LOG_LEVEL: info
  app:
    environment:
      <<: *defaults
      LOG_LEVEL: debug
"""

opts = PyOptions(apply_merge=True)
c = opts.from_yaml(yaml)
# app.environment is now {"LOG_LEVEL": "debug"} (merged from base)

---

parse_duration(s)

Parse a Compose-format duration string into seconds.

from compose_spec import parse_duration

parse_duration("1m30s")   # 90.0
parse_duration("2h")      # 7200.0
parse_duration("500ms")   # 0.5

Supported units: ns, us, ms, s, m, h

Raises ValueError for invalid format.

---

format_duration(secs)

Format a number of seconds as a Compose duration string.

from compose_spec import format_duration

format_duration(90.0)    # "1m30s"
format_duration(7200.0)  # "2h"
format_duration(0.5)     # "500ms"

Raises ValueError for negative values, NaN, or infinity.

---

How It Works

This library wraps the Rust compose_spec crate using PyO3. Parsing and validation happen in Rust, giving you:

• Full Compose spec compliance — every field validated against the spec
• Round-trip fidelity — parse YAML, modify in Python, serialize back without data loss
• Fast parsing — Rust-powered YAML/JSON deserialization

Python objects are converted via a JSON roundtrip: Rust structs serialize to JSON, then Python's json.loads produces native dicts/lists/strings. This keeps the binding layer thin while giving you full access to the data tree.

Building from Source

# Create a virtual environment
python -m venv .venv
source .venv/bin/activate

# Install maturin
pip install maturin

# Build and install in development mode
maturin develop

# Run the examples
python example.py

License

Apache 2