Reusable MoonBit library and CLI for electronic evidence manifests, integrity verification, audit trails, and compliance reporting

notary

manifest

evidence

verification

digest

audit

pipeline

Download zip

Author

hcjbat

Version

0.1.0

License

Apache-2.0

Repository

github.com/hcjbat/moonbit-notary

Last updated

2 months ago

Downloads

No dependencies

#MoonBit Notary

moonbit-notary is a reusable MoonBit library and CLI for assembling, verifying, reviewing, and reporting electronic evidence manifests. It is designed for evidence hand-off, archive intake, integrity checks, and compliance-oriented workflows where deterministic outputs and auditability matter.

#What it provides

Manifest and evidence data models with a fluent builder and query helpers.
Deterministic FNV-1a, CRC32, Murmur3, Merkle-root, and digest utilities.
Directory-style scanning and category classification for common evidence formats.
Rule-based validation for paths, digests, sources, categories, sealing, and duplicate material.
A pipeline layer for batch policies, canonical snapshots, risk summaries, quality assessment, retention decisions, review queues, reconciliation, and operational diagnostics.
Hash-chained audit records, version diffs, and Markdown/JSON/CSV/HTML reports.
A command-line entrypoint and a reproducible benchmark runner.

The digest helpers in the current release are deterministic integrity fingerprints implemented in MoonBit. They are not a claim of production cryptographic security or a replacement for a reviewed cryptography library.

#Package layout

moonbit-notary/
├── moon.mod
├── LICENSE
├── README.md
├── .github/workflows/test.yml
├── cmd/main/                 # CLI executable
├── cmd/benchmark/            # measured pipeline benchmark
├── examples/                 # runnable evidence scenarios
└── src/
    ├── core/                 # manifests, items, digests, and builders
    ├── digest/               # deterministic hashes and Merkle roots
    ├── scanner/              # path scanning and classification
    ├── validator/            # compliance rules and reports
    ├── pipeline/             # batch processing and operational decisions
    ├── audit/                # hash-chained audit trail
    ├── report/               # manifest reconciliation and diffs
    ├── formatter/            # Markdown, JSON, CSV, and HTML output
    ├── crypto/               # signature and timestamp data structures
    ├── config/               # policy presets
    └── cli/                  # command parsing and dispatch

#Requirements

Install the current stable MoonBit toolchain. Verify the local environment:

moon version --all
moon check --deny-warn
moon test --deny-warn

The repository uses moon.mod and moon.pkg files, so the commands above are run from the repository root.

#CLI quick start

moon run cmd/main -- help
moon run cmd/main -- version
moon run cmd/main -- init MAN-2026-001
moon run cmd/main -- scan ./evidence
moon run cmd/main -- verify manifest.json
moon run cmd/main -- export manifest.json markdown
moon run cmd/main -- seal manifest.json
moon run cmd/main -- audit MAN-2026-001

The CLI examples use deterministic sample inputs for demonstration. Integrate the library APIs below when an application supplies its own file metadata or storage layer.

#Library example

The example assumes these imports in the consuming package's moon.pkg:

import {
  "hcjbat/moonbit-notary/src/core",
  "hcjbat/moonbit-notary/src/digest",
  "hcjbat/moonbit-notary/src/pipeline",
}

let manifest = @core.ManifestBuilder::new()
  .set_id("MAN-2026-001")
  .set_title("Evidence package")
  .set_org("Example archive")
  .set_applicant("Operator")
  .add_evidence_simple(
    1,
    "Contract scan",
    "evidence/contract.pdf",
    @core.Category::PdfDocument,
    "sha256-fingerprint",
    "md5-fingerprint",
    520000L,
    "operator-1",
  )
  .build()

@digest.seal_manifest(manifest)
let batch = @pipeline.EvidenceBatch::from_manifest(manifest)
let assessment = @pipeline.assess_batch(
  batch,
  @pipeline.BatchPolicy::strict(),
)
let report = assessment.to_markdown()

src/pipeline is additive to the existing packages. Applications can start with EvidenceBatch::from_manifest, call verify_batch or assess_batch, and then pass the structured result to their own UI, storage, or reporting layer.

#Benchmarks

Run the real benchmark locally:

moon run cmd/benchmark --

The runner measures 100, 1,000, and 5,000 deterministic evidence records and prints elapsed milliseconds, throughput, canonical snapshot size, archive readiness, and digest prefixes. The recorded observation and reproduction details are in docs/benchmarks.md. Timings are machine- dependent; digest and readiness invariants are the stable checks.

#Development and verification

moon fmt
moon fmt --check
moon info
moon check --deny-warn
moon check --target all
moon test --deny-warn
moon test --target all

moon info regenerates the package interface summaries. Review generated .mbti changes as public API changes. The GitHub Actions workflow runs strict check/test and formatting gates on Linux, macOS, and Windows.

The test suite covers core data models, digest behavior, scanning, validation, formatting, audit trails, reconciliation, the pipeline's policy and boundary behavior, and the benchmark report format.

#License

Licensed under the Apache License 2.0.

#Contributing

Please include focused tests for behavior changes, run the verification commands above, keep public APIs small, and document any third-party source, generated code, fixture, or redistribution constraint introduced by a change.

Site source Report issue Packages Build queue Skills Statistics