rosterops 0.1.0__tar.gz

rosterops-0.1.0/.github/workflows/ci.yml

@@ -0,0 +1,31 @@
+name: CI
+
+on:
+  push:
+    branches: ["master", "main"]
+  pull_request:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.10", "3.11", "3.12", "3.13"]
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+      - name: Install package with dev + optional extras
+        run: pip install -e ".[dev,excel,pandas]"
+      - name: Lint
+        run: ruff check .
+      - name: Format check
+        run: ruff format --check .
+      - name: Type check
+        run: mypy src
+      - name: Tests with coverage
+        run: pytest --cov=rosterops --cov-report=term-missing
+      - name: Build distributions
+        run: python -m build

rosterops-0.1.0/.gitignore

@@ -0,0 +1,24 @@
+# Python
+__pycache__/
+*.py[cod]
+*.egg-info/
+.eggs/
+dist/
+build/
+.venv/
+venv/
+env/
+
+# Tooling
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.coverage
+coverage.xml
+htmlcov/
+.hypothesis/
+
+# Local runs / sensitive data
+output/
+*.local.yaml
+*.local.yml

rosterops-0.1.0/CHANGELOG.md

@@ -0,0 +1,60 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+
+### Added
+
+- Planning baseline: `REQUIREMENTS.md` (86-rule catalogue, modes A/B, domain
+  model, effective-date semantics) and `IMPLEMENTATION_PHASES.md` tracker.
+- Package skeleton (src layout, hatchling build, CLI entry point, py.typed).
+- Strict versioned configuration layer (`RosterConfig`) with policy, grain,
+  input-mapping, and rule-override schemas.
+- Canonical domain schemas and enumerations (employee, territory, assignment,
+  event, coverage; grain and severity definitions).
+- Half-open interval primitives and effective-date helpers (termination-date
+  semantics, reconstruction-window filtering).
+- Rule registry: 86 stable rule specifications with default severities, mode
+  applicability, and parameter declarations; deterministic finding model.
+- `rosterops rules list` / `rosterops rules describe` CLI commands.
+- Tooling: ruff, mypy (strict), pytest + coverage, hypothesis, CI workflow
+  matrix for Python 3.10–3.13.
+- Input ingestion layer: `read_raw`/`detect_format` for CSV (always),
+  Parquet (always), and Excel (optional `[excel]` extra via `fastexcel`);
+  graceful `ING-001/002/003` handling with actionable install hints.
+- Source→canonical mapping (`mapping/canonical.py`): explicit mappings,
+  aliases, configurable null tokens, per-field date formats, categorical value
+  maps/validation, list parsing, trim policy, and per-row lineage
+  (`source_file`/`source_sheet`/`source_row`). Implements `SCH-001..005` and
+  `ID-005`; no row is ever silently dropped.
+- Leading-zero identifier preservation (string keys end-to-end).
+- Canonical event ledger (`events/`): all eight event types normalized into a
+  single model with namespaced identity plus content fingerprints, cross-file
+  duplicate detection (`ID-007`), baseline-window cutoff (`DAT-006`), same-date
+  conflict detection with an explicit policy hook (`DAT-007`), and auditable
+  termination-date-semantics conversion (`DAT-008`). Excluded events are never
+  dropped — they retain a row with a status and finding. Duplicate movements
+  surface as `REC-006` (and `PRO-003` for promotions); conflicting transfers as
+  `TRF-006`.
+- `Lineage` value object shared by loaded and derived records.
+- Rule engine (`engine/`): grain enforcement (`GRN-001..003`), baseline state
+  reconstruction with half-open intervals, snapshot (Mode B) and process
+  (Mode A) rule scoping, movement metrics, vacancy derivation, event
+  reconciliation (`REC-001..004`), and a deterministic finding sort with stable
+  issue keys/ids. Post-reconstruction lifecycle rules cover transfers
+  (`TRF-005/006/007/008`), promotions (`PRO-002/003/004/005`), leave
+  (`LOA-006/007/009/010`), coverage (`COV-003/004/005`), events after
+  termination (`DAT-005`), vacancy intervals (`TER-007`), and the `LOA-008`
+  guard.
+- Result writers (`outputs/`): the full Mode A output set and a Mode B subset
+  that omits unavailable artefacts with reasons; safety checks refuse to write
+  into input folders unless `overwrite` is set; Excel via `openpyxl`, plus a
+  console report.
+- Public API (`RosterEngine.process` / `.validate`) and a wired CLI
+  (`rosterops process`, `rosterops validate`) with documented exit codes.
+- Optional pandas interop (`[pandas]` extra): `to_polars`/`to_pandas` without
+  mutating caller frames (decision D11).

rosterops-0.1.0/CONTRIBUTING.md

@@ -0,0 +1,32 @@
+# Contributing to rosterops
+
+## Development setup
+
+```bash
+python -m venv .venv
+.venv\Scripts\activate        # Windows
+pip install -e ".[dev]"
+```
+
+## Quality gates (run before submitting changes)
+
+```bash
+ruff check .
+ruff format --check .
+mypy src
+pytest                        # unit + rules + property + integration
+```
+
+CI runs the same gates on Python 3.10–3.13.
+
+## Guidelines
+
+- Requirements live in `REQUIREMENTS.md`; progress in `IMPLEMENTATION_PHASES.md`.
+  A behavior change updates both in the same change.
+- Every rule needs a stable ID, at least one positive and one negative test,
+  and documentation that matches the code. Never advertise an unimplemented rule.
+- Domain decisions (date semantics, grain, policy defaults) are configuration,
+  not hard-coded facts. When ambiguity appears, pick a configurable default,
+  document it, and log it in the phases tracker backlog.
+- Use synthetic data only in tests and examples. Never commit real roster data.
+- Findings must carry source lineage (file, sheet, row, column) wherever known.