Skip to main content

Architectural Decision Record framework with bi-temporal validity, AST checkers, and FastMCP integration

Project description

iil-adrfw

ADR Framework for the IIL Platform — schema validation, loader normalization, constitution graph, audit tooling.

PyPI Python 3.12+ License: MIT

What it does

  • Schema v3 — strict JSON Schema for ADR frontmatter with additionalProperties: false
  • Phase 1 Normalizer — 15+ field aliases, status normalization, type coercion, reference extraction
  • Phase 2 Validator — jsonschema validation against adr_frontmatter.schema.json
  • Phase 3 Domain — typed ADR dataclass with Status enum, temporal fields, relations
  • Constitution Graph — cross-ADR dependency/supersession graph with cycle detection
  • Audit — staleness checks, implementation evidence verification, drift detection
  • CLI + MCP Serveriil-adrfw CLI and iil-adrfw-mcp FastMCP server

Real-world validation

Tested against 156 real platform ADRs (as of v0.3.1):

Mode Result
Schema validation 156/156 (100%)
Staleness (>6mo) 0 stale
Broken references 0
Dependency edges 31

Installation

pip install iil-adrfw

# From source with dev dependencies:
pip install -e ".[dev]"

Usage

Python API

from pathlib import Path
from iil_adrfw.persistence import load_adr, load_adrs

# Load single ADR
adr = load_adr(Path("docs/adr/ADR-099.md"), Path("schemas/"))

# Load all ADRs in directory
adrs = load_adrs(Path("docs/adr/"), Path("schemas/"))

# Diagnosis mode (skip normalization)
raw_adr = load_adr(path, schemas, raw=True)

CLI

# Validate all ADR frontmatters against schema v3
iil-adrfw validate docs/adr/

# Check staleness (>6 months), broken references, missing reviews
iil-adrfw staleness docs/adr/ --months 6

# Generate dependency graph (text, DOT, or JSON)
iil-adrfw graph docs/adr/ --dot > graph.dot

# Export Outline-compatible markdown registry
iil-adrfw export docs/adr/ -o adr-registry.md

# Audit constitution health
iil-adrfw audit docs/adr/

# Query ADRs by question/domain
iil-adrfw query docs/adr/ "Which ADR governs deployment?"

MCP Server (11 tools)

iil-adrfw-mcp  # stdio transport, register in Windsurf mcp_config.json

Tools: adr_validate, adr_staleness, adr_impact, adr_check, adr_explain, adr_query, adr_audit, adr_validate_cross_repo, adr_propose, adr_diff, adr_narrate

Schema v3 highlights

  • 5 new fields: updated, version, review_status, owner, implementation_done_when
  • 3 removed fields: glossary, review_cadence, next_review_date
  • Status enum: {draft, proposed, accepted, deprecated, superseded, rejected, experimental}
  • Implementation status: {none, planned, in_progress, partial, implemented, complete, verified, rolled_back}

Loader normalizations (Phase 1)

Step What
C.1 12 field aliases (datedecision_date, authorowner, etc.)
C.2 Status normalization (case, suffixes)
C.3 Scalar-to-list auto-wrapping
C.4 Reference field normalization (ADR-NNN extraction)
C.5 ID inference, title inference, domains default, deciders default
C.6 Amended-format normalization
C.7 implemented field → implementation_status mapping
C.8 Strip unknown properties

See SCHEMA_V3_SPEC.md for full specification.

Running tests

python examples/test_e2e_schema_v3.py   # Schema v3 tests (22 cases)
python examples/test_e2e.py              # Core tests
python examples/test_e2e_v11.py          # v1.1 format tests

Project structure

iil-adrfw/
├── schemas/                    # JSON Schema files
│   ├── adr_frontmatter.schema.json
│   ├── adr_rules.schema.json
│   └── constitution.schema.json
├── src/iil_adrfw/
│   ├── persistence/            # Loader (normalize + validate + construct)
│   ├── domain/                 # ADR dataclass, Status enum
│   ├── graph/                  # Constitution graph
│   ├── audit/                  # Staleness, evidence checks
│   ├── checkers/               # AST-based code checkers
│   ├── cli.py                  # CLI entry point
│   └── server.py               # FastMCP server
├── examples/                   # Example ADRs + test suites
├── SCHEMA_V3_SPEC.md          # Full v3 specification
└── SCHEMA_V3_CHANGELOG.md     # Changelog for downstream consumers

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

iil_adrfw-0.5.0.tar.gz (139.6 kB view details)

Uploaded Source

Built Distribution

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

iil_adrfw-0.5.0-py3-none-any.whl (78.8 kB view details)

Uploaded Python 3

File details

Details for the file iil_adrfw-0.5.0.tar.gz.

File metadata

  • Download URL: iil_adrfw-0.5.0.tar.gz
  • Upload date:
  • Size: 139.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.13

File hashes

Hashes for iil_adrfw-0.5.0.tar.gz
Algorithm Hash digest
SHA256 c595456869bc034459c0037ef79a4d24e00c1dc9bf4c094d5c1cbed73b816255
MD5 9452fea7fe897377c3fdc5b10ad4f163
BLAKE2b-256 e242262fb94e0b695747cc44687c276e384bbd3777edee364a6bb3377f672b5e

See more details on using hashes here.

File details

Details for the file iil_adrfw-0.5.0-py3-none-any.whl.

File metadata

  • Download URL: iil_adrfw-0.5.0-py3-none-any.whl
  • Upload date:
  • Size: 78.8 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.13

File hashes

Hashes for iil_adrfw-0.5.0-py3-none-any.whl
Algorithm Hash digest
SHA256 004099deb2230c7b10cac36abf00ecf1bcf544c5bb69ac8a7f28e3b944248f9a
MD5 f83c59051a312c10644d3f2079550b25
BLAKE2b-256 647d8b92442c0a4c776afa0206650e49c51bf720bb2f9c4905ea5f30b4278523

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