focus-data-toolkit 0.11.0__py3-none-any.whl

focus_data_toolkit/__init__.py

@@ -0,0 +1,69 @@
+"""focus-data-toolkit — FOCUS sample-data generation, 1.x -> 1.4 conversion, validation.
+
+Public API:
+
+- :func:`focus_data_toolkit.convert.convert_to_focus_1_4` — convert FOCUS 1.2/1.3
+  rows into the four FOCUS 1.4 datasets (small-volume, in-memory).
+- :func:`focus_data_toolkit.convert.convert_files` — the same conversion, streamed from a
+  Cost and Usage file with **bounded memory** (large client data), to CSV or Parquet.
+- :func:`focus_data_toolkit.schema.detect_focus_schema` — identify the FOCUS dataset and
+  version of a header row, with a confidence assessment.
+- :func:`validate_dataset_bundle` (alias :func:`validate_bundle`) — cross-dataset
+  (referential / reconciliation / split-allocation / lifecycle) validation of a bundle of
+  datasets. Distinct from the per-dataset linter below.
+- :func:`focus_data_toolkit.model.validator.lint_focus_1_4_structure` — structurally
+  + semantically lint rows against the committed FOCUS 1.4 data model. (A linter, not a full
+  FOCUS conformance validator; ``validate_focus_1_4`` is a deprecated alias.)
+- :mod:`focus_data_toolkit.generators` — deterministic, provider-realistic FOCUS 1.2/1.3
+  source generators for AWS, Azure and GCP.
+"""
+
+from focus_data_toolkit._version import __version__
+from focus_data_toolkit.convert import (
+    ConversionCancelled,
+    ConversionResult,
+    convert_files,
+    convert_to_focus_1_4,
+)
+from focus_data_toolkit.lifecycle import (
+    DatasetInstance,
+    check_dataset_instances,
+    check_instance_chains,
+    check_status_transitions,
+)
+from focus_data_toolkit.model.validator import (
+    LintReport,
+    ValidationReport,
+    Violation,
+    lint_focus_1_4_structure,
+    validate_focus_1_4,
+)
+from focus_data_toolkit.progress import ProgressEvent
+from focus_data_toolkit.schema import SchemaDetectionResult, detect_focus_schema
+from focus_data_toolkit.validate import BundleReport, validate_dataset_bundle
+
+# Alias matching the API name used in the P1 plan / docs.
+validate_bundle = validate_dataset_bundle
+
+__all__ = [
+    "BundleReport",
+    "ConversionCancelled",
+    "ConversionResult",
+    "DatasetInstance",
+    "LintReport",
+    "ProgressEvent",
+    "SchemaDetectionResult",
+    "ValidationReport",
+    "Violation",
+    "__version__",
+    "check_dataset_instances",
+    "check_instance_chains",
+    "check_status_transitions",
+    "convert_files",
+    "convert_to_focus_1_4",
+    "detect_focus_schema",
+    "lint_focus_1_4_structure",
+    "validate_bundle",
+    "validate_dataset_bundle",
+    "validate_focus_1_4",
+]

focus_data_toolkit/__main__.py

@@ -0,0 +1,6 @@
+"""``python -m focus_data_toolkit`` — same entry point as the ``focus-toolkit`` script."""
+
+from focus_data_toolkit.cli import main
+
+if __name__ == "__main__":
+    raise SystemExit(main())

focus_data_toolkit/_version.py

@@ -0,0 +1,8 @@
+"""Single source of truth for the package version.
+
+Kept deliberately import-free so the build backend can read
+``[tool.setuptools.dynamic] version = {attr = "focus_data_toolkit._version.__version__"}``
+without importing the package (which would pull in optional/runtime modules at build time).
+"""
+
+__version__ = "0.11.0"