Skip to main content

Python reference implementation of the markstay spec (v1.1): source-level block identity for Markdown

Project description

markstay , Python reference implementation (v1 core)

The Python reference implementation of the markstay spec (v1.1). markstay is a source-level identity primitive for Markdown blocks: an id token that stays bound to its block across edits (marker stay:), so a reference to a block survives the document being rewritten, including by an LLM.

This is the parser-free core: everything string-level and parser-independent (§8 hashing, §3/§4 marker grammar, §5 blank-line segmentation, §7/§11 lint, §9 quote recovery, §9.1 resolution ladder). It mirrors the JavaScript reference (markstay on npm); both are gated by a shared language-neutral conformance corpus, which turns "two implementations agree" from an assertion into a tested fact.

Install

pip install markstay

Zero runtime dependencies (Python standard library only). CommonMark-tree segmentation (§5.2) is an optional extra:

pip install "markstay[commonmark]"   # pulls in markdown-it-py

Requires Python >= 3.9.

Library

import markstay as M

md = "The ingest stage retries three times.\n<!-- stay:a1b2 -->\n"

# parse into content blocks with attached markers (§5)
blocks = M.parse_document(md)

# well-formedness + intra-doc invariants (§7): duplicate/orphan/malformed/drift
_, findings = M.lint_document(md)

# regeneration diff (§11): what an edit did to the ids (dropped/duplicated/moved)
findings = M.lint_diff(before_md, after_md)

# §8 content hash (ASCII-normalized SHA-256)
M.body_hash("some block body")

# §9.1 resolution ladder: re-attach ids after an edit, or report DETACHED
anchors = M.build_anchors(before_md)
resolutions = M.resolve(anchors, after_md)   # id -> marker | hash | quote | detached

Public API (mirrors the JS index.js surface): normalize_body, body_hash, Marker, find_markers, strip_markers, segment_blank_line, segment_commonmark, Block, parse_document, Finding, lint_document, lint_diff, sort_findings, has_errors, Selector, normalize, body_score, context_bonus, best_match, CONTEXT_CHARS, Anchor, Resolution, build_anchors, resolve, DEFAULT_THRESHOLD, DEFAULT_MARGIN.

CLI

markstay FILE [FILE ...]            # well-formedness + intra-doc checks
markstay --before OLD.md NEW.md     # regeneration diff (dropped/duplicated/relocated ids)
markstay --json ...                 # machine-readable findings
markstay --commonmark ...           # §5.2 CommonMark-tree segmentation (needs the extra)

Exit status is non-zero when any error-level finding is reported, so it can gate a commit hook or an agent's post-edit step.

The conformance corpus (the actual deliverable)

The corpus under conformance/ is shared with the JavaScript reference. 276 vectors across two tiers:

  • spec/ , hand-authored from the spec prose, asserting what the words require. These are authority; a spec/ vector the reference fails is a reference bug, not a corpus error.
  • gen/ , emitted from the reference for breadth/regression.

The JS reference runs the same JSON, so the two runners are a cross-impl regression sentinel: any later change to either implementation that breaks agreement fails one of them.

Running the tests

pip install -e ".[commonmark]"
pytest

License

MIT

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

markstay-0.1.0.tar.gz (30.0 kB view details)

Uploaded Source

Built Distribution

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

markstay-0.1.0-py3-none-any.whl (16.5 kB view details)

Uploaded Python 3

File details

Details for the file markstay-0.1.0.tar.gz.

File metadata

  • Download URL: markstay-0.1.0.tar.gz
  • Upload date:
  • Size: 30.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for markstay-0.1.0.tar.gz
Algorithm Hash digest
SHA256 bb2ef933abf7ae5881b96e330810b9f2825e4f8e42d5ea5cbcba7daa29d49cd7
MD5 77040af8c3d59e5895a41cf83b2c6dd9
BLAKE2b-256 13739e441c0400c66fbebf7158b0f5ee24de95dabd636b792af0a22fb72243fb

See more details on using hashes here.

Provenance

The following attestation bundles were made for markstay-0.1.0.tar.gz:

Publisher: publish.yml on markstaymd/markstay-py

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file markstay-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: markstay-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 16.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for markstay-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 3d634b7193a20398f2241b3fda478a62db55c8a845a4d521ead19b0ae56fb6a9
MD5 9031fe2a8c7a3d5fe839e9e23baf02fa
BLAKE2b-256 11ab718de9f81badf8ecd5ccc1f5e87b17b56a8464ff93717910c640da484751

See more details on using hashes here.

Provenance

The following attestation bundles were made for markstay-0.1.0-py3-none-any.whl:

Publisher: publish.yml on markstaymd/markstay-py

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

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