Skip to main content

didactic-settings

PyPI Python License CI Docs

didactic-settings composes a validated dx.Model from layered configuration: a primary file and the config-group fragments its defaults: list selects, a profile, overlay files, environment variables, dotenv files, command-line arguments and dotted overrides. Every layer is checked against the schema at every depth, tagged-union fields are descended into the variant their discriminator selects, ${...} expressions are resolved against the composed tree, and the instance records which layer wrote each leaf.

Install

pip install didactic-settings
pip install 'didactic-settings[yaml]'

JSON and TOML are read through the standard library; the yaml extra installs PyYAML for .yaml and .yml files.

Compose a configuration

from typing import Literal

import didactic.api as dx
from didactic.settings import compose_traced


class OptimizerSpec(dx.TaggedUnion, discriminator="kind", extra="forbid"):
    lr: float = 1e-3


class Adam(OptimizerSpec):
    kind: Literal["adam"] = "adam"
    betas: tuple[float, ...] = (0.9, 0.999)


class Sgd(OptimizerSpec):
    kind: Literal["sgd"] = "sgd"
    momentum: float = 0.9


class TrainerSpec(dx.Model, extra="forbid"):
    epochs: int = 1
    out_dir: str = "runs"
    log_dir: str = "${.out_dir}/logs"


class RunSpec(dx.Model, extra="forbid"):
    optimizer: OptimizerSpec = dx.field(default_factory=Adam)
    trainer: TrainerSpec = dx.field(default_factory=TrainerSpec)


run = compose_traced(
    "conf/run.yaml",
    schema=RunSpec,
    profile="dev",
    overlays=["conf/big.toml"],
    groups={"optimizer": "sgd"},
    overrides=["optimizer.momentum=0.5", ("trainer.epochs", 3)],
)
run.value.trainer.log_dir
run.provenance["trainer.epochs"].label      # "override:trainer.epochs=3"
run.provenance["optimizer.momentum"].label  # "override:optimizer.momentum=0.5"

conf/run.yaml may carry a defaults: list selecting root fragments and config-group fragments (optimizer: adam loads conf/optimizer/adam.yaml); groups= replaces the file's choice for a slot; profile="dev" loads conf/profiles/dev.yaml; an override whose key contains / (model/type_encoder=lstm) selects a nested group and one without sets a field, decoded by the leaf's annotation. A key the schema does not declare, at any depth and in any layer, is refused with its dotted path and the layer that set it, and so is a value of the wrong type.

Class-based settings

from didactic.settings import EnvSource, FileSource, Settings


class RunSettings(Settings, RunSpec):
    __sources__ = (FileSource("local.toml"), EnvSource(prefix="APP_"))


settings = RunSettings.load("conf/run.yaml", profile="dev", trainer__epochs=3)
settings.__provenance__["trainer.epochs"].label

Settings.load() composes the primary file, its groups, the profile, overlays, the declared sources and the overrides in that precedence. EnvSource reads APP_TRAINER__EPOCHS for trainer.epochs and decodes the text by the field's annotation; APP_OPTIMIZER='{"kind": "sgd"}' sets a whole slot.

Source Input
EnvSource(prefix="APP_") environment variables
DotEnvSource(path=".env", prefix="APP_") a dotenv file
FileSource(path="config.toml") one JSON, TOML, or YAML document
CliSource(args=namespace) an argparse.Namespace or mapping

Because Settings extends dx.Model, the composed tree receives the same type checks, axioms, field validators, and indexed-field checks as any other Didactic model.

Documentation

See the settings guide for the composition ladder, config groups, union descent, interpolation and the resolver registry, and per-leaf provenance.

License

Released under the MIT License.

Metadata

Release files for didactic-settings 0.18.1

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

Source distribution (sdist)

Source distribution for didactic-settings 0.18.1
File Size Uploaded Provenance Actions
didactic_settings-0.18.1.tar.gz 51.3 kB With provenance Details for didactic_settings-0.18.1.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for didactic-settings 0.18.1
File Interpreter ABI Platform Actions
didactic_settings-0.18.1-py3-none-any.whl Python 3 none any Details for didactic_settings-0.18.1-py3-none-any.whl

Total release size: 112.5 kB

Release files / didactic_settings-0.18.1.tar.gz

Download URL didactic_settings-0.18.1.tar.gz
Size 51.3 kB
Tags Source
SHA-256 checksum
How to use checksums
e8c4cd31d0873f66dcd1f0d9f7c8c707374c56016c7181486c9651d06c099140
BLAKE2b-256 checksum
How to use checksums
d6767a00fa41738ebaeb42ce6e57daa28a97f62abd513ac35390b2d636e792db
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

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.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 30, 2026.

Transparency log

Release files / didactic_settings-0.18.1-py3-none-any.whl

Download URL didactic_settings-0.18.1-py3-none-any.whl
Size 61.2 kB
Tags Python 3
SHA-256 checksum
How to use checksums
1904f0d21875ca88f66b15afa0beebd1e39544303a9aa4d500cc3d92e2f2850c
BLAKE2b-256 checksum
How to use checksums
152d69d9b05ab2945e720af83276a8791efdd4266feb8c8164cebd88761b9c15
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

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.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 30, 2026.

Transparency log

Security for didactic-settings 0.18.1

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.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for didactic_settings-0.18.1.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 30, 2026.

Transparency log

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.

Release history Release notifications | RSS feed

This release

0.18.1 This release

2 release files Full provenance

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