openom-core 0.1.0__tar.gz

openom_core-0.1.0/.gitignore

@@ -0,0 +1,62 @@
+# --- OpenOM repo hygiene (see CLAUDE.md Rule 2) ---
+
+# Marketing mock - NEVER committed
+OpenOM - Make your deal legible to the buy-side's AI.htm
+OpenOM - Make your deal legible to the buy-side's AI_files/
+
+# Planning scratch - git-ignored by rule
+docs/superpowers/
+.planning/
+
+# Eval / tooling scratch
+.deepeval/
+
+# Real OM corpus - private test fixtures, never committed (confidential)
+OMs/
+
+# --- Python (/core, /cli, /mcp) ---
+__pycache__/
+*.py[cod]
+*.egg-info/
+.eggs/
+build/
+dist/
+.venv/
+venv/
+.env
+.mypy_cache/
+.ruff_cache/
+.pytest_cache/
+.coverage
+htmlcov/
+
+# --- Node / TS (/js, /extension) ---
+node_modules/
+*.tsbuildinfo
+.npm/
+extension/src/generated/
+extension/test-results/
+extension/playwright-report/
+
+# --- OS / editor ---
+.DS_Store
+Thumbs.db
+.idea/
+.vscode/
+coverage/
+extension/*.zip
+
+js/test-results/
+
+# marketing mock (Rule 2 - never commit)
+*.htm
+*_files/
+.playwright-mcp/
+om-normative-spec-draft-v0.1.md
+om-standard-handoff-v4-updated.md
+recon.md
+openOM-decision-memo.md
+
+# audit/backlog scratch is temp - lives in .planning/ (see Rule 2)
+/BACKLOG.md
+/*-audit*.md

openom_core-0.1.0/.hypothesis/.gitignore

@@ -0,0 +1,9 @@
+# This .gitignore file was automatically created by Hypothesis. Hypothesis gitignores
+# .hypothesis by default, because we generally recommend that .hypothesis not be checked
+# into version control.
+#
+# If you *would* like to check .hypothesis into version control, you should delete this
+# file. Hypothesis will not re-create this .gitignore unless .hypothesis is deleted (and
+# if it does, that's a bug - please report it!)
+
+*

openom_core-0.1.0/.hypothesis/constants/001fe83bc75f70ff

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\xmp.py
+# hypothesis_version: 6.155.2
+
+[b'\xef\xbb\xbf', ' </rdf:RDF>', '&', '&amp;', '&gt;', '&lt;', '/Metadata', '<', '</rdf:RDF>', '>', 'assertedDate', 'omspec', 'openOM', 'payloadFilename', 'payloadHash', 'replace', 'sourceDocHash', 'specName', 'specVersion', 'supersedes', 'utf-8', '{', '}']

openom_core-0.1.0/.hypothesis/constants/17983501d9df90f5

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\__init__.py
+# hypothesis_version: 6.155.2
+
+['0.0.0+unknown', '0.1', 'DealSummary', 'Finding', 'ImageManifest', 'OMPayload', 'PayloadTooLargeError', 'Profile', 'ReadResult', 'RealEstateListing', 'Report', 'SPEC_VERSION', 'Severity', 'TextResult', 'Tolerances', '__version__', 'canonicalize', 'classify', 'embed', 'extract_images', 'extract_text', 'hash_bytes', 'load_schema', 'openom-core', 'payload_hash', 'read', 'reembed_warnings', 'summarize_deal', 'validate']

openom_core-0.1.0/.hypothesis/constants/1bad5597e3545771

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\images.py
+# hypothesis_version: 6.165.9
+
+[100000000, 'DeviceGray', 'DeviceRGB', 'colorspace', 'contentHash', 'deduped', 'error', 'hasSMask', 'height', 'image/png', 'images', 'mime', 'path', 'pdf', 'sha256:', 'unknown', 'width', 'xref']

openom_core-0.1.0/.hypothesis/constants/1e120eed62cb9e41

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\schema.py
+# hypothesis_version: 6.155.2
+
+['om-0.1.schema.json', 'openom_core', 'spec', 'utf-8']

openom_core-0.1.0/.hypothesis/constants/2599456d76b7b639

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\text.py
+# hypothesis_version: 6.165.9
+
+[100000, ',', '-', ':', 'ascii', 'h', 'malformed cursor', 'nextCursor', 'off', 'page', 'pageRange', 'pdf', 'rows', 'tables', 'text', 'truncated']

openom_core-0.1.0/.hypothesis/constants/2599a2740a02ac70

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\canonical.py
+# hypothesis_version: 6.165.9
+
+[55296, 57343, 'NFC', 'meta', 'sha256:', 'signature']

openom_core-0.1.0/.hypothesis/constants/2bbc47f0d389ac49

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\images.py
+# hypothesis_version: 6.155.2
+
+[1.0, 2.0, 100000000, 'DeviceGray', 'DeviceRGB', 'colorspace', 'contentHash', 'deduped', 'error', 'hasSMask', 'height', 'image/png', 'images', 'mime', 'page', 'path', 'pdf', 'rendered-page', 'sha256:', 'source', 'unknown', 'width', 'xref']

openom_core-0.1.0/.hypothesis/constants/2dc28f4fae47c5df

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\inspect.py
+# hypothesis_version: 6.155.2
+
+[0.0, 0.2, 0.5, 0.6, 0.8, 0.85, 1.0, 3.0, 100, 'Profile', 'bbox', 'chars', 'class', 'classConfidence', 'colorspaces', 'count', 'hasSMask', 'hashValid', 'hybrid', 'images', 'native', 'ocrOverlay', 'pages', 'payload', 'pdf', 'present', 'scanned', 'specVersion', 'text', 'textCoverage', 'type', 'unknown']

openom_core-0.1.0/.hypothesis/constants/391bb7a7db122863

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\xmp.py
+# hypothesis_version: 6.165.9
+
+[b'\xef\xbb\xbf', ' </rdf:RDF>', '&', '&amp;', '&gt;', '&lt;', '/Metadata', '<', '</rdf:RDF>', '>', 'assertedDate', 'omspec', 'openOM', 'payloadFilename', 'payloadHash', 'replace', 'specName', 'specVersion', 'supersedes', 'utf-8', '{', '}']

openom_core-0.1.0/.hypothesis/constants/3b99d738d4809a52

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\summary.py
+# hypothesis_version: 6.155.2
+
+[', ', 'USD', 'address', 'addressLocality', 'addressRegion', 'askingPrice', 'assertedBy', 'assertedDate', 'broker', 'brokerage', 'buildingSF', 'capRate', 'commencement', 'currency', 'deal', 'expiration', 'lease', 'leaseTypeAsserted', 'license', 'noi', 'noiAsOfDate', 'noiType', 'postalCode', 'pricePerSF', 'pricePerUnit', 'property', 'propertyType', 'streetAddress', 'tenantEntity', 'termMonths', 'units']

openom_core-0.1.0/.hypothesis/constants/3bee2e16f7a8fc8e

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\canonical.py
+# hypothesis_version: 6.155.2
+
+[55296, 57343, 'NFC', 'meta', 'sha256:', 'signature']

openom_core-0.1.0/.hypothesis/constants/517da07006529f54

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\embed.py
+# hypothesis_version: 6.165.9
+
+[5000000, '/AF', '/DP', '/DecodeParms', '/EF', '/F', '/Filter', '/FlateDecode', '/UF', '/application/ld+json', '/assertedDate', '0.1', 'OMW-W051', 'application/ld+json', 'assertedDate', 'om.json', 'payloadHash', 'specVersion', 'supersedes', 'warning']

openom_core-0.1.0/.hypothesis/constants/5273145f2f12a42a

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\validate.py
+# hypothesis_version: 6.165.9
+
+[0.005, 0.01, 0.02, 0.2, 30.4375, 31.0, "'", '-', '/', '/assertedDate', '/currency', '/deal', '/deal/askingPrice', '/deal/capRate', '/deal/noi', '/deal/noiAsOfDate', '/deal/pricePerSF', '/lease/expiration', '/lease/rentSchedule', '/lease/termMonths', '/meta/signature', '/meta/supersedes', '/property/buildingSF', 'E', 'I', 'NN', 'NNN', 'OM-CONS-010', 'OM-CONS-011', 'OM-CONS-012', 'OM-CONS-013', 'OM-CONS-014', 'OM-CONS-020', 'OM-CONS-021', 'OM-CONS-022', 'OM-CONS-023', 'OM-CONS-024', 'OM-CONS-025', 'OM-CONS-026', 'OM-CONS-030', 'OM-CONS-031', 'OM-CONS-032', 'OM-CONS-033', 'OM-CONS-034', 'OM-CONS-040', 'OM-CONS-041', 'OM-CONS-050', 'OM-CONS-060', 'OM-DD', 'OM-DD-001', 'OM-DD-002', 'OM-DD-004', 'OM-ERR-013', 'OM-ERR-014', 'OM-ERR-090', 'OMI-I001', 'OMI-I002', 'OMI-I003', 'OMV-E001', 'OMV-E002', 'OMV-E003', 'OMV-E010', 'OMW-W010', 'OMW-W011', 'OMW-W012', 'OMW-W013', 'OMW-W014', 'OMW-W020', 'OMW-W021', 'OMW-W022', 'OMW-W023', 'OMW-W024', 'OMW-W025', 'OMW-W026', 'OMW-W030', 'OMW-W031', 'OMW-W032', 'OMW-W033', 'OMW-W034', 'OMW-W040', 'OMW-W041', 'OMW-W050', 'OMW-W060', 'W', 'absolute-net', 'annualRent', 'askingPrice', 'assertedDate', 'buildingSF', 'cam', 'capRate', 'commencement', 'currency', 'deal', 'error', 'escalationFromPrior', 'expiration', 'gross', 'hvac', 'in-place', 'inf', 'info', 'insurance', 'lease', 'leaseTypeAsserted', 'meta', 'modified-gross', 'monthlyRent', 'noi', 'noiAsOfDate', 'noiType', 'parking', 'periodEnd', 'periodStart', 'pricePerSF', 'pro-forma', 'property', 'remainingTermMonths', 'rentPSF', 'rentSchedule', 'required', 'roof', 'signature', 'source', 'structure', 'supersedes', 'taxes', 'termMonths', 'verified', 'warning']

openom_core-0.1.0/.hypothesis/constants/8a59fd8a0a527e9a

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\inspect.py
+# hypothesis_version: 6.165.9
+
+[0.2, 0.5, 0.8, 0.85, 1.0, 3.0, 100, 'Profile', 'bbox', 'class', 'classConfidence', 'colorspaces', 'count', 'hasSMask', 'hashValid', 'hybrid', 'images', 'native', 'pages', 'payload', 'pdf', 'present', 'scanned', 'specVersion', 'text', 'textCoverage', 'unknown']

openom_core-0.1.0/.hypothesis/constants/b82be5d38e8f5622

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\xmp.py
+# hypothesis_version: 6.165.9
+
+[b'\xef\xbb\xbf', ' </rdf:RDF>', '&', '&amp;', '&gt;', '&lt;', '/Metadata', '<', '</rdf:RDF>', '>', 'assertedDate', 'omspec', 'openOM', 'payloadFilename', 'payloadHash', 'replace', 'specName', 'specVersion', 'supersedes', 'utf-8', '{', '}']

openom_core-0.1.0/.hypothesis/constants/ccfdabc45d5fe4c6

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\types.py
+# hypothesis_version: 6.155.2
+
+['@context', '@type', 'OMPayload', 'assertedBy', 'assertedDate', 'currency', 'deal', 'ext', 'lease', 'meta', 'property', 'specVersion']

openom_core-0.1.0/.hypothesis/constants/d7dad3c83a335ac4

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\__init__.py
+# hypothesis_version: 6.165.9
+
+['0.1.0.dev0']

openom_core-0.1.0/.hypothesis/constants/de803f019b0b542d

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\validate.py
+# hypothesis_version: 6.155.2
+
+[0.005, 0.01, 0.02, 0.2, 30.4375, 31.0, "'", '-', '/', '/assertedDate', '/currency', '/deal', '/deal/askingPrice', '/deal/capRate', '/deal/noi', '/deal/noiAsOfDate', '/deal/pricePerSF', '/lease/expiration', '/lease/rentSchedule', '/lease/termMonths', '/meta/signature', '/meta/supersedes', '/property/buildingSF', 'E', 'I', 'NN', 'NNN', 'OM-CONS-010', 'OM-CONS-011', 'OM-CONS-012', 'OM-CONS-013', 'OM-CONS-014', 'OM-CONS-020', 'OM-CONS-021', 'OM-CONS-022', 'OM-CONS-023', 'OM-CONS-024', 'OM-CONS-025', 'OM-CONS-026', 'OM-CONS-030', 'OM-CONS-031', 'OM-CONS-032', 'OM-CONS-033', 'OM-CONS-034', 'OM-CONS-040', 'OM-CONS-041', 'OM-CONS-050', 'OM-CONS-060', 'OM-DD-001', 'OM-DD-002', 'OM-DD-003', 'OM-DD-004', 'OM-ERR-013', 'OM-ERR-014', 'OM-ERR-090', 'OM-TRUST-009', 'OM-TRUST-010', 'OMI-I001', 'OMI-I002', 'OMI-I003', 'OMV-E001', 'OMV-E002', 'OMV-E003', 'OMV-E010', 'OMW-W010', 'OMW-W011', 'OMW-W012', 'OMW-W013', 'OMW-W014', 'OMW-W020', 'OMW-W021', 'OMW-W022', 'OMW-W023', 'OMW-W024', 'OMW-W025', 'OMW-W026', 'OMW-W030', 'OMW-W031', 'OMW-W032', 'OMW-W033', 'OMW-W034', 'OMW-W040', 'OMW-W041', 'OMW-W050', 'OMW-W051', 'OMW-W052', 'OMW-W060', 'OMW-W061', 'US', 'W', 'absolute-net', 'address', 'addressCountry', 'annualRent', 'askingPrice', 'assertedDate', 'buildingSF', 'cam', 'capRate', 'commencement', 'currency', 'deal', 'error', 'escalationFromPrior', 'expiration', 'gross', 'hvac', 'in-place', 'inf', 'info', 'insurance', 'lease', 'leaseTypeAsserted', 'meta', 'modified-gross', 'monthlyRent', 'noi', 'noiAsOfDate', 'noiType', 'parking', 'periodEnd', 'periodStart', 'pricePerSF', 'pro-forma', 'property', 'remainingTermMonths', 'rentPSF', 'rentSchedule', 'required', 'roof', 'source', 'structure', 'supersedes', 'taxes', 'termMonths', 'verified', 'warning']

openom_core-0.1.0/.hypothesis/constants/e1e4fe9940bd1f03

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\errors.py
+# hypothesis_version: 6.165.9
+
+['OM-IO-BADUTF8', 'OM-IO-BOMB', 'OM-IO-DUPKEY', 'OM-IO-NUMRANGE', 'OM-IO-STRUCTURE', 'error', 'info', 'warning']

openom_core-0.1.0/.hypothesis/constants/ec0635288f17334e

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\embed.py
+# hypothesis_version: 6.155.2
+
+[5000000, '/AF', '/DP', '/DecodeParms', '/EF', '/F', '/Filter', '/FlateDecode', '/UF', '/application/ld+json', '/assertedDate', '0.1', 'OMW-W051', 'application/ld+json', 'assertedDate', 'om.json', 'payloadHash', 'sourceDocHash', 'specVersion', 'supersedes', 'warning']

openom_core-0.1.0/.hypothesis/constants/fa756589e5af3e25

@@ -0,0 +1,4 @@
+# file: E:\OpenOM\core\src\openom_core\text.py
+# hypothesis_version: 6.155.2
+
+[100000, ',', '-', ':', 'ascii', 'h', 'malformed cursor', 'nextCursor', 'off', 'page', 'pageRange', 'pdf', 'rows', 'tables', 'text', 'truncated']

openom_core-0.1.0/PKG-INFO

@@ -0,0 +1,55 @@
+Metadata-Version: 2.5
+Name: openom-core
+Version: 0.1.0
+Summary: openOM deterministic core - embed/read/inspect/validate CRE offering-memorandum payloads. Zero inference, ever.
+Project-URL: Homepage, https://openom.app
+Project-URL: Documentation, https://openom.app/docs/
+Project-URL: Repository, https://github.com/Vervelio-Labs/OpenOM
+License-Expression: MIT
+Keywords: commercial-real-estate,cre,data-standard,factur-x,json-ld,json-schema,offering-memorandum,open-standard,openom,pdf,pdf-a-3,proptech,real-estate
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Topic :: Office/Business :: Financial
+Classifier: Topic :: Text Processing :: Markup
+Classifier: Typing :: Typed
+Requires-Python: >=3.11
+Requires-Dist: jsonschema>=4.20
+Requires-Dist: pikepdf>=9
+Requires-Dist: rfc8785>=0.1.4
+Provides-Extra: dev
+Requires-Dist: hypothesis>=6.100; extra == 'dev'
+Requires-Dist: mypy>=1.10; extra == 'dev'
+Requires-Dist: numpy>=1.26; extra == 'dev'
+Requires-Dist: pillow>=10; extra == 'dev'
+Requires-Dist: pymupdf>=1.24; extra == 'dev'
+Requires-Dist: pypdfium2>=4; extra == 'dev'
+Requires-Dist: pytest-cov>=5; extra == 'dev'
+Requires-Dist: pytest>=8; extra == 'dev'
+Requires-Dist: ruff>=0.5; extra == 'dev'
+Requires-Dist: types-jsonschema>=4.20; extra == 'dev'
+Provides-Extra: render
+Requires-Dist: pymupdf>=1.24; extra == 'render'
+Description-Content-Type: text/markdown
+
+# openom-core
+
+The deterministic Python core of the [openOM](../README.md) standard: the PDF/data verbs
+(`embed`, `read`, `inspect`, `extract`, `validate`) and RFC 8785 canonicalization + SHA-256
+integrity. **Zero inference, ever** (CI's `boundary` job enforces it). Byte-parity with `js/`.
+
+```sh
+# From a clone (openom-core is not yet on PyPI):
+pip install -e core                 # add the [dev] extra only to contribute
+# Once published:  pip install openom-core
+```
+
+```python
+from openom_core.embed import embed, read
+out = embed(pdf_bytes, payload, asserted_date="2026-08-16")   # non-destructive
+r = read(out)                                                 # r.present / r.hash_valid / r.payload
+```
+
+Entrypoints: `openom_core.{canonical,embed,validate,inspect,images,text,schema,xmp}`.
+Tests: `pytest core -q --cov=openom_core --cov-fail-under=90`. Schema + vectors live in [`/spec`](../spec).

openom_core-0.1.0/README.md

@@ -0,0 +1,20 @@
+# openom-core
+
+The deterministic Python core of the [openOM](../README.md) standard: the PDF/data verbs
+(`embed`, `read`, `inspect`, `extract`, `validate`) and RFC 8785 canonicalization + SHA-256
+integrity. **Zero inference, ever** (CI's `boundary` job enforces it). Byte-parity with `js/`.
+
+```sh
+# From a clone (openom-core is not yet on PyPI):
+pip install -e core                 # add the [dev] extra only to contribute
+# Once published:  pip install openom-core
+```
+
+```python
+from openom_core.embed import embed, read
+out = embed(pdf_bytes, payload, asserted_date="2026-08-16")   # non-destructive
+r = read(out)                                                 # r.present / r.hash_valid / r.payload
+```
+
+Entrypoints: `openom_core.{canonical,embed,validate,inspect,images,text,schema,xmp}`.
+Tests: `pytest core -q --cov=openom_core --cov-fail-under=90`. Schema + vectors live in [`/spec`](../spec).

openom_core-0.1.0/pyproject.toml

@@ -0,0 +1,111 @@
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project]
+name = "openom-core"
+version = "0.1.0"
+description = "openOM deterministic core - embed/read/inspect/validate CRE offering-memorandum payloads. Zero inference, ever."
+readme = "README.md"
+requires-python = ">=3.11"
+license = "MIT"
+keywords = [
+  "openom", "offering-memorandum", "commercial-real-estate", "cre", "real-estate", "proptech",
+  "pdf", "pdf-a-3", "factur-x", "json-ld", "json-schema", "open-standard", "data-standard",
+]
+classifiers = [
+  "Development Status :: 4 - Beta",
+  "Intended Audience :: Developers",
+  "License :: OSI Approved :: MIT License",
+  "Programming Language :: Python :: 3.11",
+  "Topic :: Office/Business :: Financial",
+  "Topic :: Text Processing :: Markup",
+  "Typing :: Typed",
+]
+# The embed/read/validate/canonical core is pikepdf-only, so `openom-core` stays MIT-clean. PyMuPDF
+# (AGPL-3.0-or-commercial) powers only extract/inspect/render + the signed-OM incremental save, so it
+# is an OPTIONAL `[render]` extra - a portal / commercial integrator bundling the core does not inherit
+# AGPL unless they opt into rendering.
+dependencies = [
+  "pikepdf>=9",
+  "rfc8785>=0.1.4",
+  "jsonschema>=4.20",
+]
+
+[project.optional-dependencies]
+render = [
+  "PyMuPDF>=1.24",  # AGPL-3.0-or-commercial - powers extract/inspect/render + signed-incremental embed
+]
+dev = [
+  "openom-core[render]",  # tests exercise extract/inspect/render, so dev pulls the render extra
+  "pytest>=8",
+  "pytest-cov>=5",
+  "mypy>=1.10",
+  "ruff>=0.5",
+  "numpy>=1.26",
+  "pillow>=10",
+  "hypothesis>=6.100",
+  "types-jsonschema>=4.20",
+  "pypdfium2>=4",  # independent 2nd PDF renderer - the differential image cross-check (#167)
+]
+
+[project.urls]
+Homepage = "https://openom.app"
+Documentation = "https://openom.app/docs/"
+Repository = "https://github.com/Vervelio-Labs/OpenOM"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/openom_core"]
+
+# Ship the JSON Schema INSIDE the package (#149) so a pip-installed openom-core/-mcp can load it via
+# importlib.resources - no reliance on an adjacent /spec dir. It is committed as ordinary package data
+# at src/openom_core/om-0.1.schema.json (NOT a force-include from ../spec, which cannot survive into a
+# self-contained sdist -> `python -m build` would fail building the wheel from the sdist). /spec stays
+# the single source of truth; the in-package copy is drift-locked to it by tests/test_schema_packaged.py.
+
+[tool.ruff]
+line-length = 100
+src = ["src", "tests"]
+
+[tool.ruff.lint]
+select = ["E", "F", "I", "B", "UP", "W"]
+
+[tool.mypy]
+python_version = "3.11"
+strict = true
+files = ["src"]
+
+# PyMuPDF ships incomplete type info; treat it as untyped so strict no-untyped-call
+# doesn't fire on its API (we still type our own wrappers around it).
+[[tool.mypy.overrides]]
+module = ["pymupdf", "pymupdf.*"]
+follow_imports = "skip"
+disallow_untyped_calls = false
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+addopts = "-ra"
+markers = [
+  "cross_impl: cross-implementation round-trip ([OM-VEC-002]); needs Node + OPENOM_XIMPL_DIR",
+]
+# Coverage is enforced in CI (pytest --cov=openom_core --cov-fail-under=90), not in the
+# default addopts, so single-file dev runs stay fast and don't trip the floor.
+
+[tool.coverage.run]
+source = ["openom_core"]
+branch = true
+
+[tool.coverage.report]
+show_missing = true
+exclude_lines = ["pragma: no cover", "raise NotImplementedError", "if TYPE_CHECKING:"]
+
+# Mutation testing (#133) - assertion-STRENGTH, beyond line/branch coverage. Targets the correctness-
+# critical modules where a flipped comparison or wrong rounding would produce a divergent-but-passing
+# standard. Runs on a schedule (not per-push); surviving mutants are tracked as coverage debt.
+[tool.mutmut]
+paths_to_mutate = [
+  "src/openom_core/canonical.py",
+  "src/openom_core/validate.py",
+  "src/openom_core/embed.py",
+]
+tests_dir = ["tests/"]

openom_core-0.1.0/scripts/characterize_corpus.py

@@ -0,0 +1,71 @@
+#!/usr/bin/env python
+# SPDX-License-Identifier: MIT
+"""Local corpus characterization / classifier-tuning tool (Task 8).
+
+Runs ``inspect`` over every PDF under ``OMs/`` (the confidential, git-ignored corpus) and
+prints class, confidence, textCoverage, and page count. Given a labels file
+(JSON ``{relpath: "native"|"hybrid"|"scanned"}``) it also prints a confusion matrix so
+thresholds can be tuned against real ground truth.
+
+    python core/scripts/characterize_corpus.py [--labels labels.json]
+
+Not a CI test: the corpus is not committed, and ground-truth labels require human judgment (Q3).
+"""
+
+from __future__ import annotations
+
+import argparse
+import json
+from collections import Counter
+from pathlib import Path
+
+from openom_core.inspect import inspect
+
+ROOT = Path(__file__).resolve().parents[2]
+OMS = ROOT / "OMs"
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--labels", type=Path, default=None, help="JSON {relpath: class}")
+    args = parser.parse_args()
+
+    if not OMS.exists():
+        print(f"corpus not present: {OMS}")
+        return
+
+    labels: dict[str, str] = (
+        json.loads(args.labels.read_text(encoding="utf-8")) if args.labels else {}
+    )
+    dist: Counter[str] = Counter()
+    confusion: Counter[tuple[str, str]] = Counter()
+    pdfs = sorted(OMS.rglob("*.pdf"))
+
+    for pdf in pdfs:
+        rel = pdf.relative_to(OMS).as_posix()
+        try:
+            prof = inspect(pdf.read_bytes())
+        except Exception as exc:  # noqa: BLE001 - tool: report and keep going
+            print(f"ERROR {rel}: {exc}")
+            continue
+        cls = prof["class"]
+        dist[cls] += 1
+        print(
+            f"{cls:8} conf={prof['classConfidence']:.2f} tc={prof['textCoverage']:.2f} "
+            f"pages={prof['pages']:>3}  {rel}"
+        )
+        if rel in labels:
+            confusion[(labels[rel], cls)] += 1
+
+    print(f"\nDISTRIBUTION over {len(pdfs)} files: {dict(dist)}")
+    if confusion:
+        correct = sum(n for (t, p), n in confusion.items() if t == p)
+        total = sum(confusion.values())
+        print(f"ACCURACY: {correct}/{total}")
+        print("CONFUSION (true -> predicted):")
+        for (t, p), n in sorted(confusion.items()):
+            print(f"  {t:8} -> {p:8}: {n}{'' if t == p else '   <-- MISS'}")
+
+
+if __name__ == "__main__":
+    main()

openom_core-0.1.0/scripts/gen_fuzz_corpus.py

@@ -0,0 +1,87 @@
+#!/usr/bin/env python
+# SPDX-License-Identifier: MIT
+"""Generate the cross-language JCS differential-fuzz corpus (#129).
+
+The RFC 8785 canonicalization is the anti-fork keystone: if the Python and TS reference impls ever
+serialize the same value to different bytes, a PDF verifies in one client and fails in another.
+This produces a DETERMINISTIC, edge-weighted corpus of accepted payloads + their canonical hashes.
+Both cores canonicalize the SAME corpus and MUST match `expected` byte-for-byte (test_fuzz.py /
+fuzz.test.ts). The generator is seeded (no Date/urandom) so re-running is a no-op - a diff means a
+hash moved. Run after any canonicalization change:  python core/scripts/gen_fuzz_corpus.py
+"""
+
+from __future__ import annotations
+
+import json
+import random
+from pathlib import Path
+
+from openom_core.canonical import CanonicalizationError, canonicalize, hash_bytes
+
+OUT = Path(__file__).resolve().parents[2] / "spec" / "vectors" / "fuzz"
+SEED = 20260818  # fixed seed -> reproducible corpus (no Date/urandom)
+COUNT = 600
+
+# Values that historically fork JCS number/string serialization - the fuzz weights toward these.
+_NUM_EDGES = [
+    0, -0.0, 1, -1, 2**53 - 1, -(2**53 - 1), 0.1, -0.1, 0.5, 1.5,
+    1e-7, 1.0000001, 9.999999e-7, 0.0000001, 123456.789, 1e-6, 1e-8,
+    3.141592653589793, 2.718281828459045, 100.0, 0.0001, 1234567890, 42,
+]
+# Non-ASCII given as escapes so the source file stays pure ASCII (avoids encoding surprises).
+_STR_EDGES = [
+    "", "a", "A", "z", "é", "café",  # e-acute (NFC precomposed)
+    "\U0001f600", "\U0001f4a9 tail",  # astral / surrogate-pair strings
+    "line1\nline2", "tab\there", 'quote"q', "back\\slash", "slash/fwd",
+    "中文", "key sort z", "key sort a",
+]
+_KEY_EDGES = ["a", "b", "A", "z", "é", "\U0001f600key", "10", "2", "中", "Z", "aa"]
+
+
+def _rand_value(rng: random.Random, depth: int) -> object:
+    if depth >= 4:
+        kind = rng.choice(["num", "str", "bool", "null"])
+    else:
+        kind = rng.choices(
+            ["num", "str", "bool", "null", "arr", "obj"], weights=[30, 30, 8, 8, 12, 12]
+        )[0]
+    if kind == "num":
+        v = rng.choice(_NUM_EDGES)
+        return v if rng.random() < 0.7 else rng.uniform(-1e6, 1e6)
+    if kind == "str":
+        if rng.random() < 0.8:
+            return rng.choice(_STR_EDGES)
+        return "".join(rng.choice("abcXYZ é中") for _ in range(rng.randint(0, 8)))
+    if kind == "bool":
+        return rng.random() < 0.5
+    if kind == "null":
+        return None
+    if kind == "arr":
+        return [_rand_value(rng, depth + 1) for _ in range(rng.randint(0, 4))]
+    keys = rng.sample(_KEY_EDGES, k=rng.randint(1, min(5, len(_KEY_EDGES))))
+    return {k: _rand_value(rng, depth + 1) for k in keys}
+
+
+def main() -> None:
+    OUT.mkdir(parents=True, exist_ok=True)
+    rng = random.Random(SEED)
+    corpus: list[dict] = []
+    while len(corpus) < COUNT:
+        keys = rng.sample(_KEY_EDGES, k=rng.randint(1, 5))
+        obj = {k: _rand_value(rng, 1) for k in keys}
+        try:
+            canonicalize(obj)  # keep only inputs BOTH impls accept (skip range/NFC-dup rejections)
+        except CanonicalizationError:
+            continue
+        corpus.append(obj)
+
+    dumped = (json.dumps(o, ensure_ascii=False, sort_keys=False) for o in corpus)
+    corpus_txt = "\n".join(dumped) + "\n"
+    expected_txt = "\n".join(hash_bytes(canonicalize(o)) for o in corpus) + "\n"
+    (OUT / "corpus.jsonl").write_text(corpus_txt, encoding="utf-8", newline="\n")
+    (OUT / "expected.jsonl").write_text(expected_txt, encoding="utf-8", newline="\n")
+    print(f"wrote {len(corpus)} fuzz vectors + expected hashes -> {OUT}")
+
+
+if __name__ == "__main__":
+    main()