helakit 0.2.0

A toolkit for validating and working with Sri Lankan data.

Helakit

A toolkit for validating and working with Sri Lankan data.

Helakit is a small, dependency-free Python library for validating and parsing Sri Lankan identifiers — NIC numbers, phone numbers, postal codes, and more. Each identifier lives in its own self-contained subpackage and shares a single result type, so adding a new validator is a matter of dropping in a folder.

For usage and API reference, see the docs. This page is a tour of how the project is laid out — useful if you want to contribute, audit, or fork.

Design principles

1. Modular monolith. Every validation domain (nic, phone, postal, …) is a self-contained subpackage. Adding a new domain is adding a new folder; no plugin system or registry to wire up.
2. Zero runtime dependencies. The library itself uses only the Python standard library. Dev and docs tools are kept in extras.
3. Data as Python dicts. Lookup tables (provinces, districts, mobile prefixes, NIC encoding rules) live as module-level dict constants in .py files, not JSON / Parquet / SQLite. Lookups are O(1) and cost nothing after import.
4. One result shape, two entry points. Every validator exposes both a rich validate_X(value) -> ValidationResult and a boolean is_valid_X(value) -> bool. ValidationResult is truthy when valid, so the rich form drops into if statements naturally.
5. src/ layout. Tests run against the installed package, not the source tree, so packaging mistakes can't masquerade as passing tests.
6. Strict typing end to end. A py.typed marker ships with the wheel and CI runs mypy --strict. Type hints use modern syntax (X | Y, list[X], dict[K, V]).

Repository layout

src/helakit/
├── __init__.py        # public API surface + __version__
├── py.typed           # PEP 561 marker
├── _core/             # shared primitives, no domain logic
│   ├── result.py      # ValidationResult + ValidationError dataclasses
│   ├── base.py        # Validator Protocol
│   └── exceptions.py  # HelakitError hierarchy
├── _data/             # cross-domain lookup tables (provinces, districts, …)
└── <domain>/          # one folder per identifier type
    ├── __init__.py    # re-exports validate_X / is_valid_X
    ├── validator.py   # the actual rules
    ├── exceptions.py  # domain-specific exceptions
    └── _data.py       # domain-specific lookup tables (optional)

Domains currently in the tree: nic/, phone/, postal/. The underscore-prefixed packages (_core/, _data/) are private — nothing outside the package should import from them.

Around the source:

tests/                 # mirror of src/helakit/, one folder per domain
docs/                  # MkDocs site (Material theme, mike for versioning)
.github/workflows/     # ci.yml, docs.yml, release.yml
pyproject.toml         # hatchling build, ruff, mypy, pytest, coverage config

How a domain is structured

Each domain is independent. The contract:

• validate_X(value: str) -> ValidationResult — full validation, returns parsed fields in result.data and structured errors in result.errors.
• is_valid_X(value: str) -> bool — boolean shorthand.
• A domain-specific XError exception class inheriting from HelakitError, for unrecoverable misuse (not validation failures — those go through ValidationResult).
• A _data.py for any tables only that domain cares about.

The __init__.py re-exports the two functions and the exception, and the top-level helakit/__init__.py re-exports them again so users can write from helakit import validate_nic without thinking about layout.

Validators

Validator	Function	Status
NIC	validate_nic	Available
Phone	validate_phone	Planned
Postal	validate_postal	Planned
Passport	(planned)	Planned

More (driving licence numbers, BR numbers, …) will follow. Stubs for unimplemented validators raise NotImplementedError so call-sites written against the planned API don't silently no-op.

Adding a new validator

1. Create src/helakit/<domain>/ mirroring the layout above.
2. Add a tests/test_<domain>/ folder with at least one test for the happy path and one for each error class.
3. Add a docs/validators/<domain>.md page.
4. Re-export the public functions from src/helakit/__init__.py.
5. Add a ## [Unreleased] entry to CHANGELOG.md.

See CONTRIBUTING.md for the dev-environment setup and the PR checklist.

Data contributions

A lot of helakit's value lives in its lookup tables. Adding or correcting entries (postal codes, mobile prefixes, district names) is the easiest way to contribute and doesn't require deep Python — each table is a plain dict in src/helakit/_data/ or src/helakit/<domain>/_data.py. Cite a source in the PR description so we can verify it.

Documentation

Docs are versioned with mike and published in three flavours:

• /dev/ — built from every push to main. Tracks in-progress work.
• /<MAJOR>.<MINOR>/ — built from each release tag. Frozen.
• /latest/ — alias for the most recent release.

The bare site URL (Aswikinz.github.io/Helakit) redirects to /latest/ once a release exists, and to /dev/ until then.

Links: docs · dev · changelog · contributing.

Install

pip install helakit

Requires Python 3.10+.

Status

Helakit is alpha software. The public API may change before 1.0; pin versions accordingly.

License

Apache License 2.0 — see LICENSE.