modelo

modelo is a Rust multi-model validation library and a Python bindings extension intended as a drop-in replacement for python-ort.

Namespaces

Each model family lives in its own namespace, so families can be added without disturbing the others. Currently there is one: ort, the OSS Review Toolkit model.

	ORT model	Shared validation engine
Rust	`modelo::models::ort::*` (`src/models/ort/`)	`modelo::models::{Model, ValidationError}`
Python	`modelo.ort`	—

Adding a family means a new src/models/<name>/ directory whose types implement models::Model, plus a modelo.<name> submodule in src/python/mod.rs. The engine and the bindings' plumbing stay as they are.

Workspace layout

This repository is a Cargo workspace with two crates, so the library stays free of CLI/TUI dependencies (clap, ratatui, crossterm):

Crate	Contents	Dependencies
`modelo`	The validation engine and models (`src/`), plus the optional Python bindings	`serde`, `serde_yaml`, `thiserror`, `purl`, `pyo3` (feature-gated)
`modelo-cli`	The `modelo` binary: CLI subcommands and the TUI	`modelo`, `clap`, `ratatui`, `crossterm`, `anyhow`

Build

cargo build --release --workspace  # library + `modelo` binary
cargo test --workspace              # unit + integration tests (203 tests)
cargo clippy --workspace --all-targets # lints
cargo doc --workspace --open        # API documentation

The Python extension module lives in its own modelo-python crate, so the modelo library and the modelo binary never link against libpython:

cargo build --release -p modelo-python

Validating files: the `modelo` CLI

The modelo binary lives in the modelo-cli crate:

cargo install modelo-cli

modelo license-classifications  tests/data/license-classifications.yml
modelo repository-configuration tests/data/repo_config/curations.yml
modelo ort-result               tests/data/evaluation-result.yml

The subcommands are still flat rather than grouped per family (modelo ort ort-result ...); that regrouping is worth doing when a second family lands, not before.

Each subcommand prints valid <kind> file. and exits 0, or reports the first parse/validation error and exits non-zero. --debug additionally pretty-prints the parsed model. This is the equivalent of python-ort's ort-validate (src/tools/ort_validate.py).

Interactive TUI

modelo tui [FILE] (or modelo with no arguments) opens a ratatui interface: pick the model kind, type a file path, see the validation result. ↑/↓ (or k/j) move and scroll, Enter confirms, b goes back from the result screen, q/Esc quits.

Using it from Python

pip install modelo

or, from a checkout:

pip install maturin
maturin develop     # into the active virtualenv
# or: pip install .                   # maturin is the build backend

The modelo.ort submodule exposes the three top-level ORT models, each parsed from YAML and navigated with plain attribute access, the way python-ort's pydantic models are:

from pprint import pprint

from modelo.ort import LicenseClassifications, OrtResult, RepositoryConfiguration

result = OrtResult.from_yaml_file("tests/data/evaluation-result.yml")
config = RepositoryConfiguration.from_yaml_str(open(".ort.yml").read())

pprint(result.analyzer)                                 # AnalyzerRun(start_time=..., ...)
result.analyzer.environment.ort_version
result.analyzer.result.projects[0].id
result.repository.vcs.url

Invalid input raises ValueError (with the failing field path, as pydantic's ValidationError does); an unreadable path raises OSError.

Every nested model is an instance of a class named after it (type(result.analyzer).__name__ == "AnalyzerRun"), all of them subclasses of modelo.ort.Object. Besides attribute access they support keys(), values(), items(), obj["field"], "field" in obj, len(obj), == and to_dict():

import json

result.to_dict()                                        # nested plain dicts and lists
json.loads(result.to_json()) == result.to_dict()        # True — one source of truth

Fields that ORT itself serializes as scalars stay scalars: Identifier is a string ("PIP::requirements.txt:1.0"), enums are their member names, and free-form maps such as labels are dicts.

Type checking

The wheel is a PEP 561 typed package: modelo/py.typed ships alongside modelo/ort.pyi, so mypy, pyright and ty resolve import modelo.ort and the model classes without falling back to Any. The compiled extension lives at modelo._modelo; modelo.ort is a thin Python re-export of it, which is what makes the module statically resolvable.

Examples

Runnable counterparts of python-ort's examples/, using only the standard library:

python examples/license_classifications.py tests/data/license-classifications.yml
python examples/repo_config.py             tests/data/repo_config/curations.yml
python examples/ort_result.py              tests/data/evaluation-result.yml --analyzer

Checking parity against python-ort

examples/parity_check.py parses every fixture in tests/data/ with both libraries and diffs the serialized result, ignoring set-backed field ordering (arbitrary in both):

pip install . python-ort
python examples/parity_check.py

All 9 fixtures agree; analyzer-result.yml is rejected by both libraries for the same unknown field. The behaviours this covers:

Int-backed enums (Severity, *Reason, …) accept the numeric value or the case-sensitive member name, and serialize as the member name — python-ort's ValidatedIntEnum.
Unknown fields are rejected wherever python-ort sets extra="forbid", and accepted where it sets extra="allow"/"ignore".
Loose scalars are coerced as python-ort's mode="before" validators do (an integer start_lines, non-string options values).

Continuous integration

.github/workflows/ci.yml runs on every push and pull request: format (rustfmt), lint (clippy, warnings denied), test (Linux/macOS/Windows), python (bindings built and smoke-tested on 3.10 and 3.13) and parity (the diff above against python-ort).

Every third-party action is pinned to a full commit SHA with the released tag in a trailing comment (uses: actions/checkout@3d3c42e... # v7.0.1), so a moving tag cannot change what runs. To bump one, resolve the new tag's SHA (gh api repos/<owner>/<repo>/commits/<tag> --jq .sha) and update both the SHA and the comment. actionlint .github/workflows/*.yml validates the files.

Deployment

Both registries publish from a v* tag (git tag -s v0.1.0 && git push origin v0.1.0) or a manual Run workflow dispatch. Bump the version in Cargo.toml and pyproject.toml first — they are separate files and must agree.

PyPI — `.github/workflows/release-pypi.yml`

Builds wheels with maturin-action for Linux (x86_64, aarch64), macOS (x86_64, aarch64) and Windows (x64), plus an sdist, then uploads everything in one publish job via pypa/gh-action-pypi-publish (maturin upload only authenticates with an API token, so it cannot do the OIDC exchange trusted publishing needs).

One-time setup: on PyPI, add a trusted publisher for the modelo project pointing at this repository and the workflow file release-pypi.yml with environment pypi; then create a GitHub environment named pypi. No API token is stored — the job authenticates with the OIDC token from id-token: write.

crates.io — `.github/workflows/release-crates.yml`

Runs the test suite and cargo package first, then cargo publish.

One-time setup: create a crates.io API token scoped to publish-update for this crate, and store it as the secret CARGO_REGISTRY_TOKEN in a GitHub environment named crates-io.

Local dry runs

cargo publish --dry-run
maturin build --release   # wheel lands in target/wheels/

License

Download files

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

Source Distribution

python_modelo-0.9.0.tar.gz (151.4 kB view details)

Uploaded Aug 14, 2026 Source

Built Distributions

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

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

python_modelo-0.9.0-pp311-pypy311_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 14, 2026 PyPymanylinux: glibc 2.17+ x86-64

python_modelo-0.9.0-pp311-pypy311_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 PyPymanylinux: glibc 2.17+ ARM64

python_modelo-0.9.0-cp315-cp315t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.15tmanylinux: glibc 2.17+ x86-64

python_modelo-0.9.0-cp315-cp315-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.15manylinux: glibc 2.17+ x86-64

python_modelo-0.9.0-cp314-cp314t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.14tmanylinux: glibc 2.17+ x86-64

python_modelo-0.9.0-cp314-cp314t-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.14tmanylinux: glibc 2.17+ ARM64

python_modelo-0.9.0-cp314-cp314-win_amd64.whl (849.7 kB view details)

Uploaded Aug 14, 2026 CPython 3.14Windows x86-64

python_modelo-0.9.0-cp314-cp314-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.14manylinux: glibc 2.17+ x86-64

python_modelo-0.9.0-cp314-cp314-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.14manylinux: glibc 2.17+ ARM64

python_modelo-0.9.0-cp314-cp314-macosx_11_0_arm64.whl (936.4 kB view details)

Uploaded Aug 14, 2026 CPython 3.14macOS 11.0+ ARM64

python_modelo-0.9.0-cp313-cp313-win_amd64.whl (849.8 kB view details)

Uploaded Aug 14, 2026 CPython 3.13Windows x86-64

python_modelo-0.9.0-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.13manylinux: glibc 2.17+ x86-64

python_modelo-0.9.0-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.13manylinux: glibc 2.17+ ARM64

python_modelo-0.9.0-cp313-cp313-macosx_11_0_arm64.whl (936.1 kB view details)

Uploaded Aug 14, 2026 CPython 3.13macOS 11.0+ ARM64

python_modelo-0.9.0-cp312-cp312-win_amd64.whl (849.4 kB view details)

Uploaded Aug 14, 2026 CPython 3.12Windows x86-64

python_modelo-0.9.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.12manylinux: glibc 2.17+ x86-64

python_modelo-0.9.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.12manylinux: glibc 2.17+ ARM64

python_modelo-0.9.0-cp312-cp312-macosx_11_0_arm64.whl (935.3 kB view details)

Uploaded Aug 14, 2026 CPython 3.12macOS 11.0+ ARM64

python_modelo-0.9.0-cp311-cp311-win_amd64.whl (852.3 kB view details)

Uploaded Aug 14, 2026 CPython 3.11Windows x86-64

python_modelo-0.9.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.11manylinux: glibc 2.17+ x86-64

python_modelo-0.9.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 14, 2026 CPython 3.11manylinux: glibc 2.17+ ARM64

python_modelo-0.9.0-cp311-cp311-macosx_11_0_arm64.whl (946.7 kB view details)

Uploaded Aug 14, 2026 CPython 3.11macOS 11.0+ ARM64

File details

Details for the file python_modelo-0.9.0.tar.gz.

File metadata

Download URL: python_modelo-0.9.0.tar.gz
Upload date: Aug 14, 2026
Size: 151.4 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: uv/0.12.4 {"installer":{"name":"uv","version":"0.12.4","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for python_modelo-0.9.0.tar.gz
Algorithm	Hash digest
SHA256	`7af8d7b8faef3fadfbc5d5c56d52d744fee11e503a392a1a84c3e2dbb8561183`
MD5	`baea66800e0760baee7a5940f01ef534`
BLAKE2b-256	`7f3ebb2e8494b0816fdf4917f734fbe17a60118a0a2cdcfbacacd2881b1ff04e`

Algorithm	Hash digest
SHA256	`7909a248e7ecd9b7b829d9801d52b852db1924ff6be4f223d2e267ffc16e8f2b`
MD5	`26d39a264160005a7a6471ed0c4fe98a`
BLAKE2b-256	`982b20a7ad1e81b95aa2711b02e4162f260233f0e1e5fe23ac643c4db3d10e3a`

Algorithm	Hash digest
SHA256	`9808c27a62310b4a0fac063ecd4f3db7887f8e3021cde98bdb520ba8ac10eceb`
MD5	`a7cde6bd6da786557ba6ca6552ad2bb1`
BLAKE2b-256	`bcd71382369bb813019fe1672460d99075c3d8455305e94177b0ac2064c776ac`

Algorithm	Hash digest
SHA256	`f0f9058c6272175b0ec440b4d27c527064e8e8beb5478e7316aaca7ae9b19135`
MD5	`22d2fd62c6af759e33d3112e0c8800cf`
BLAKE2b-256	`1f46e88cb1b3b0c0c7e4ba53fa2850eaf30264dd97eff6f27d079d1497494008`

Algorithm	Hash digest
SHA256	`feabb4f4b0cf6d131f2e56a397659b1efedf151161c18b2c3a4b1225be58d727`
MD5	`fc867ecf57ff6f3576dd95aa4aeecfcb`
BLAKE2b-256	`8170660fcee9335b6768be1854e5aa6cad0d4990d32b6c7312d6a45eeb8ad577`

Algorithm	Hash digest
SHA256	`847c5bf858b9b085d5eb50de65ee86dce75f2d9cba8e703ee83469aba8ac4f0a`
MD5	`78a397a38262bd1ce4aac67c4fe2f137`
BLAKE2b-256	`371e90956b830143871da94fb020a895e884537af84de5463a752bb91657c905`

Algorithm	Hash digest
SHA256	`acaeb4b3e78c49262b765c3654e5896a5aeeb0ee5a1aea12ad4b7f3ed15ce7e1`
MD5	`bfeae1a35ab0c045854166dda7310071`
BLAKE2b-256	`86488f3530070fec55464c44cf1ec62c1cbd3affb9511325a56313970cd014e9`

Algorithm	Hash digest
SHA256	`134535c8c9210d75c9be7c281df6980d4115a91b4b5aed20528fc028778f3fa5`
MD5	`bfe8d578eb9e0fa21b744f986b289320`
BLAKE2b-256	`514b368c6f6f19eb57fef4aa4546282cfcbbb374337682a83b05942c0d446f0e`

Algorithm	Hash digest
SHA256	`50879b8e69ba2ac4071bf9fd5d570e975cdaae31794ddf4869008c8414c1b3fb`
MD5	`5771b89ff45dcbac344653b9e6042ac1`
BLAKE2b-256	`879aeb19f66c6c18bb1af6625da9f16749cb299e049f867b39c661d62bd14663`

Algorithm	Hash digest
SHA256	`1d82beca8b78be8c53f9d1caaed0e1a875e4ec94e891c5ef1c3fc90609851263`
MD5	`e54c1d1f5ee45234b3e2c1f7d6416485`
BLAKE2b-256	`6d65edc5e1bde0e35a20d52a4145f99907520751ef9e721fdf4bfbd92f020a36`

Algorithm	Hash digest
SHA256	`bcfa67202145a412a9de2a3dd6d98c7d45e0de8fa5f074dd654cf016d47beed9`
MD5	`6fc9163e0223deecf6913b2226836151`
BLAKE2b-256	`1af1b59e56b7251e4cfbe36ae5b901f1f46f12d13130179eb24b88834e695153`

Algorithm	Hash digest
SHA256	`7fd236af289c68701022a9d127c5d15d126c18c5068cf83e78ca8955290b7d7e`
MD5	`1e70b3e7152d8e6381e8ebce27bce912`
BLAKE2b-256	`64c26b5e657436d02e8d036b6518e542d1ee1974148c5c5e4102c0d23dfcd6cc`

Algorithm	Hash digest
SHA256	`a451afd2afab95abbc5119642878048d5fa7dc901863ac349a639cd789bc5915`
MD5	`1cc172dac07f1431dbe65b44a49de3c4`
BLAKE2b-256	`02c8e2946522e7c2cdab94323088e1d949e0bef10f4ca6294eb5593cf0933da4`

Algorithm	Hash digest
SHA256	`9f6d3fd703dd5a52b0ff31bc435451450cc97de336addeda5716942f3096df03`
MD5	`50b25469b17bd9024ce915197aa65f71`
BLAKE2b-256	`3d41fafb0024dee36fbfe69fe9e48bf76719b41e44cde4f5facfe4d3de3fb8ed`

Algorithm	Hash digest
SHA256	`1bfe50da971671abb2a82158c4b1bcba6accf1acc662b674336d8eb6409df82c`
MD5	`50a6e46d55b8ef6f852dcd072cd97dd9`
BLAKE2b-256	`100f61dfb54cdfedab3c3d3ab13fe09231a7cab9db1f5648930f4f2e04af34b3`

Algorithm	Hash digest
SHA256	`9acbdc550bce6b7ce9e475e62eab0228cec52062871c0c21523b6159f82873e0`
MD5	`fc338bb03e0a8b7686105bdfdfd9409a`
BLAKE2b-256	`0d37236ff0b87796fb1f64cc8c4ba2a75163a861861d9d90ca862ef9d8a340fc`

Algorithm	Hash digest
SHA256	`153f278cf09b5fc8e029c05f3ebf5e760b8767074742ba66c070a908a5703275`
MD5	`ab4ea928acbbe9bd08e4f90413d50e77`
BLAKE2b-256	`47448f5634f194e91e8ab86f804b4d0ab87a8e64d3dc7d1c744fe7ef840b2bbf`

Algorithm	Hash digest
SHA256	`ce61a9183a4176b88b8e18147ebf37b75408cec0ef7e69c5aad0da0d199eb3f1`
MD5	`249029356c55aeb12af3fc275f726fff`
BLAKE2b-256	`8301ae68b2dfbd82b6f6b5825064a8507d9918e249b3f722af6e2029b9897f7e`

Algorithm	Hash digest
SHA256	`0b60523d41c074564c96227801d0c0d38be77e254748656dbfaa1538e05354c1`
MD5	`1efb3d7d1475a81abc9a0c6ca140245a`
BLAKE2b-256	`fc129a565a10267812e5e084ddace36fe78d0de9a0c8fab4799e0833fc63b508`

Algorithm	Hash digest
SHA256	`48b2f313c7f9d8728ba62feb15afabc516f4bd4d92bf67e5aa155feaf698f067`
MD5	`819748e61e18771e19160dfb38a9cfe4`
BLAKE2b-256	`cad53dc939d5decc5a864864e0db468a4dc9cd41b7f9e8fea715340a8f1a5527`

Algorithm	Hash digest
SHA256	`b1193233a8b02ab75c5866c2b6f6a279488c3558ce99f87e2e307fb3ebab6416`
MD5	`8c8fe8800b270e6f8704bd9f77856148`
BLAKE2b-256	`4cef9c8979580df28540ab8bf4599a6f7afc68cad268525e11761a033228b07b`

python-modelo 0.9.0

modelo

Namespaces

Workspace layout

Build

Validating files: the modelo CLI

Interactive TUI

Using it from Python

Type checking

Examples

Checking parity against python-ort

Continuous integration

Deployment

PyPI — .github/workflows/release-pypi.yml

crates.io — .github/workflows/release-crates.yml

Local dry runs

License

Project links

Key dates

1 maintainer

Credits

GitHub Statistics

License expression

Requires

Classifiers

Download files

Source Distribution

Built Distributions

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

Validating files: the `modelo` CLI

PyPI — `.github/workflows/release-pypi.yml`

crates.io — `.github/workflows/release-crates.yml`

Algorithm	Hash digest
SHA256	`8e1a4d21a7987bc0dcf01f5f4cf61a641a6d34cca33110abc51fdba0df1c8f14`
MD5	`d49e9bf4db8ef67c3810a33bc828daaa`
BLAKE2b-256	`b5de22e1a2c534836fff27d71ee54ea7a036bc23ff6b5f9c210d54b3367fbb64`

Algorithm	Hash digest
SHA256	`eaa6b1d818a64a66701e241dd6ee29967727172d828a7b2ad2c2c56008c225c0`
MD5	`bf46d728d3a905784319a63b2983c8c9`
BLAKE2b-256	`23b467ab9e48719262cd9271869166a5ca3408e96512dcad73cd892c73ba68fa`