open-data-products 0.1.0__tar.gz

open_data_products-0.1.0/.flake8

@@ -0,0 +1,16 @@
+[flake8]
+max-line-length = 88
+extend-ignore =
+    E203,
+    E501,
+    W503
+exclude =
+    .git,
+    .mypy_cache,
+    .pytest_cache,
+    __pycache__,
+    build,
+    dist,
+    open_data_products.egg-info
+per-file-ignores =
+    open_data_products/odps/__init__.py:F401,F403,E402

open_data_products-0.1.0/.gitignore

@@ -0,0 +1,7 @@
+__pycache__/
+*.py[cod]
+build/
+dist/
+*.egg-info/
+.worktrees/
+.DS_Store

open_data_products-0.1.0/.mcp.json

@@ -0,0 +1,8 @@
+{
+  "mcpServers": {
+    "open_data_products": {
+      "command": "open-data-products",
+      "args": ["serve"]
+    }
+  }
+}

open_data_products-0.1.0/AGENTS.md

@@ -0,0 +1,102 @@
+# AGENTS.md
+
+Instruction file for AI coding agents (Claude Code, Codex, Cursor, Gemini)
+working in this repository. Authoritative project conventions; precedence over
+generic defaults.
+
+## Project
+
+Python SDK + MCP surface for the OpenDataProducts.org standards family
+(ODPS, ODPC, ODPG, ODPV). Single package: `open_data_products`.
+
+## Stack
+
+- Python ≥ 3.8 (tested through 3.12; CI runs on 3.14)
+- Runtime deps: `PyYAML`, `certifi`, `jsonschema`, `pycountry`, `phonenumbers`
+- Dev deps: `pytest`, `pytest-cov`, `black`, `flake8`, `mypy`, `build`, `twine`
+- No web framework, no DB, no async runtime.
+
+## Style
+
+- 4-space indent. Black with `line-length = 88`. No `any` in type hints.
+- Public API surface lives in `open_data_products/__init__.py` — list new
+  exports there explicitly; do not rely on barrel re-exports.
+- Per-spec helpers live under `open_data_products/<spec>/`. Cross-spec
+  facades live at the package root (`agent.py`, `cli.py`, `summary.py`,
+  `pricing.py`, `mcp/`).
+- Internal-only modules use a leading `_` prefix; do not export them.
+- Docstrings: one-line summary, optional short body. No paragraph essays.
+
+## Structure
+
+```
+open_data_products/
+  __init__.py          # public API surface
+  agent.py             # cross-spec load/detect/validate/explain/refs
+  cli.py               # `open-data-products` console script
+  summary.py           # load_summary — artifact references (no body)
+  pricing.py           # ODPS PricingPlans → HTTP 402 envelope
+  resources.py         # bundled resource registry
+  results.py           # ValidationResult, Reference, Resource dataclasses
+  mcp/
+    tools.py           # MCP tool registry (data + handlers)
+    manifest.py        # ARWS agent manifest
+    server.py          # stdio JSON-RPC 2.0 MCP server
+  odps/                # spec-specific package
+  odpc/  odpg/  odpv/
+skills/                # SKILL.md bundles for agent hosts
+tests/
+```
+
+New code goes in the smallest existing module that fits its spec namespace.
+Do not introduce parallel `validation.py` / `validators.py` style splits;
+that confusion already exists and should be merged, not extended.
+
+## Testing
+
+- `pytest -q` from repo root. Target: 100% green.
+- New behaviour requires a test in `tests/test_<area>.py`.
+- Conformance to agenticpatterns.veso.ai prescriptions is enforced by
+  `tests/test_agentic_patterns.py`. Treat that file as the spec — when
+  changing the agent surface, update the test first, watch it fail, then
+  implement.
+- Use `tmp_path` fixtures, not real filesystem state.
+
+## Git
+
+- Branch: `feat/<short-description>`, `fix/<short-description>`, etc.
+- Commit messages: imperative, one logical change per commit.
+- One PR per concern. Don't bundle a refactor with a feature.
+- Never `--amend` after pushing. Never `git push --force` to `main`.
+
+## Security
+
+- No hardcoded secrets in source or fixtures. The SDK has no credential
+  surface today; if you add one, read from env vars and document it here.
+- The MCP server is `safe`-class only today (read-only). When you add a
+  `state-changing` or `destructive` tool, set the `class` field in
+  `mcp/tools.py` accordingly and update `tests/test_agentic_patterns.py`.
+- Bundled JSON Schemas are loaded with `jsonschema`'s default checker.
+  Do not load arbitrary remote schemas at validate time.
+
+## Pre-Commit Checklist
+
+Before marking work complete, run all four:
+
+1. `pytest -q` — all tests pass
+2. `python -c "import open_data_products"` — package imports cleanly
+3. `python -m open_data_products.cli manifest --json | python -m json.tool`
+   — manifest renders and parses
+4. No new files in `docs/superpowers/` (that path is reserved for legacy
+   AI-planning artifacts already deleted; do not recreate)
+
+## Anti-patterns to avoid
+
+- Returning full document bodies from MCP handlers — use `load_summary`
+  for references; only load full docs when validation/explanation needs them
+- Adding console_script entry points beyond `open-data-products` — the
+  unified subcommand CLI is the contract
+- Importing from `open_data_products.odps.codecs` in user-facing code —
+  that module is internal glue between models and YAML/JSON
+- Storing absolute filesystem paths in tool responses — return logical IDs
+  and let the resource registry resolve them

open_data_products-0.1.0/BUILD.md

@@ -0,0 +1,207 @@
+# Building and Distributing Open Data Products Python SDK
+
+This guide explains how to package the Open Data Products Python SDK as a wheel file for distribution.
+
+## Quick Reference
+
+```bash
+# Install build tools
+pip install build
+
+# Build the package with minimal output
+python -m build --quiet
+
+# Install locally for testing
+pip install dist/open_data_products-0.2.0-py3-none-any.whl
+
+# Test functionality
+python -c "import open_data_products; print(open_data_products.__version__)"
+```
+
+## Prerequisites
+
+Install build tools:
+```bash
+pip install build twine
+```
+
+## Method 1: Using `build` (Recommended - PEP 517 compliant)
+
+The modern way using `pyproject.toml`:
+
+```bash
+# Install build dependencies
+pip install build
+
+# Build the package with minimal output (creates both wheel and source distribution)
+python -m build --quiet
+
+# This creates files in dist/:
+# - open_data_products-0.2.0-py3-none-any.whl (wheel file)
+# - open_data_products-0.2.0.tar.gz (source distribution)
+```
+
+## Method 2: Build with specific options
+
+```bash
+# Build with clean environment
+python -m build --wheel --outdir dist/
+
+# Build source distribution only
+python -m build --sdist
+
+# Build with verbose output
+python -m build --wheel --verbose
+```
+
+
+## Installing the Built Wheel
+
+After building, install the wheel locally for testing:
+
+```bash
+# Install the built wheel
+pip install dist/open_data_products-0.2.0-py3-none-any.whl
+
+# Or install in development mode with extras
+pip install -e ".[dev]"
+```
+
+## Verifying the Package
+
+Test the installed package:
+
+```python
+import open_data_products
+print(open_data_products.__version__)  # Should print 0.2.0
+
+from open_data_products.odps import OpenDataProduct, ODPSValidator
+from open_data_products.odps.models import ProductDetails
+
+# Test basic functionality
+product = ProductDetails(
+    name="Test", 
+    product_id="test-001",
+    visibility="public",
+    status="production", 
+    type="dataset"
+)
+odp = OpenDataProduct(product)
+odp.validate()
+print("✓ Package working correctly")
+```
+
+## Publishing to PyPI
+
+### Test PyPI (Recommended for testing)
+
+```bash
+# Upload to Test PyPI first
+python -m twine upload --repository testpypi dist/*
+
+# Install from Test PyPI to verify
+pip install --index-url https://test.pypi.org/simple/ odps-python
+```
+
+### Production PyPI
+
+```bash
+# Upload to production PyPI
+python -m twine upload dist/*
+
+# Configure credentials first:
+# Create ~/.pypirc with your API tokens
+```
+
+## Package Structure
+
+The built wheel contains:
+```
+odps_python-0.2.0-py3-none-any.whl
+├── odps/
+│   ├── __init__.py          # Package entry point with version
+│   ├── core.py              # Main OpenDataProduct class
+│   ├── models.py            # Data models for ODPS components  
+│   └── validators.py        # International standards validation
+├── odps_python-0.2.0.dist-info/
+│   ├── METADATA             # Package metadata
+│   ├── WHEEL                # Wheel metadata
+│   ├── RECORD               # File checksums
+│   └── top_level.txt        # Top-level packages
+```
+
+## Build Configuration
+
+### pyproject.toml (Modern)
+- Defines build system requirements
+- Specifies metadata and dependencies
+- Configures development tools (black, mypy)
+- PEP 517/518 compliant
+
+
+## Troubleshooting
+
+### Common Issues
+
+1. **Import errors during build**:
+   ```bash
+   # Ensure all dependencies are installed
+   pip install -r requirements.txt
+   ```
+
+2. **Version conflicts**:
+   ```bash
+   # Clean previous builds
+   rm -rf build/ dist/ *.egg-info/
+   python -m build --quiet
+   ```
+
+3. **Missing files in wheel**:
+   ```bash
+   # Check MANIFEST.in or pyproject.toml package discovery
+   python -m build --verbose
+   ```
+
+4. **Upload failures**:
+   ```bash
+   # Check credentials and package name availability
+   python -m twine check dist/*
+   ```
+
+## Distribution Checklist
+
+- [ ] Version number updated in `__init__.py`
+- [ ] CHANGELOG.md updated with release notes
+- [ ] README.md reflects current features
+- [ ] All tests passing
+- [ ] Dependencies properly specified
+- [ ] Built wheel installs and imports correctly
+- [ ] Uploaded to Test PyPI successfully
+- [ ] Documentation is current
+
+## Automated Building
+
+For CI/CD, add to your workflow:
+
+```yaml
+# .github/workflows/build.yml
+name: Build and Test
+on: [push, pull_request]
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v3
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: 3.9
+    - name: Install dependencies
+      run: |
+        pip install build twine
+        pip install -e ".[dev]"
+    - name: Build package
+      run: python -m build --quiet
+    - name: Check package
+      run: python -m twine check dist/*
+```