pal-found-cli 0.1.2__py3-none-any.whl

pal_found_cli/data_health/__init__.py

@@ -0,0 +1,9 @@
+"""Foundry Data Health CLI package."""
+
+from pal_found_cli.data_health.scripts.pal_found_data_health_cli import (
+    build_parser,
+    console_main,
+    main,
+)
+
+__all__ = ["build_parser", "console_main", "main"]

pal_found_cli/data_health/metadata-allow-list.md

@@ -0,0 +1,10 @@
+# DataHealth metadata allow-list
+
+| SDK Path | Status | Rationale |
+| --- | --- | --- |
+| `data_health.check.create` | BLOCKED | Creates a check |
+| `data_health.check.delete` | BLOCKED | Deletes a check |
+| `data_health.check.get` | PERMITTED | Single check metadata |
+| `data_health.check.replace` | BLOCKED | Replaces a check configuration |
+| `data_health.check_report.get` | PERMITTED | Single check report metadata |
+| `data_health.check_report.get_latest` | PERMITTED | Latest check report metadata |

pal_found_cli/data_health/scripts/__init__.py

@@ -0,0 +1 @@
+"""Foundry Data Health CLI scripts package."""

pal_found_cli/data_health/scripts/pal_found_data_health_cli.py

@@ -0,0 +1,458 @@
+#!/usr/bin/env python3
+"""Foundry Data Health CLI - 6 canonical Data Health API v2 operations.
+
+Exposes exactly 6 public ``foundry_sdk.v2.data_health`` operations as CLI
+subcommands across the ``Check`` client and its nested ``CheckReport``
+client:
+
+- check (4): create, delete, get, replace
+- check-report (2): get, get-latest
+
+``check_report`` operations dispatch through the nested
+``Check.CheckReport`` accessor. ``check_report get_latest`` takes an integer
+``--limit`` (default 10, maximum 100) that bounds a single response; it is
+not a cursor and never routes through PaginationHelper.
+
+The write set is ``check.create``, ``check.delete``, and ``check.replace``
+(replace inherits the replace-class write classification). The packaged
+metadata-only policy permits exactly 3 operations (check.get,
+check_report.get, check_report.get_latest) and blocks the remaining 3.
+
+Usage: pal-found-data-health <resource> <operation> [options]
+Output: JSON/TOON on stdout, NDJSON diagnostics on stderr (ADR-004, ADR-005).
+Exit codes per ADR-001 taxonomy.
+Access control: AccessControlGuard before client construction (SRS 4.2,
+ADR-007). Retry: transient conditions only per ADR-002; at-least-once
+behavior is disclosed because retrying create/replace can duplicate checks
+or re-run validation.
+Tracing: SDK-native B3 via invocation_scope; include_attribution=False
+(namespace is outside FR-ATTR-4 scope).
+"""
+
+from __future__ import annotations
+
+import argparse
+import asyncio
+import json
+import logging
+import sys
+from pathlib import Path
+from typing import Any, NoReturn
+
+from pal_found_cli.common.access_control_guard import (
+    AccessControlError,
+    AccessControlGuard,
+)
+from pal_found_cli.common.async_client_factory import AsyncClientFactory
+from pal_found_cli.common.config_loader import ConfigLoader, ConfigurationError
+from pal_found_cli.common.error_serializer import (
+    EXIT_ACCESS_CONTROL,
+    EXIT_AUTH,
+    EXIT_CONFIGURATION,
+    EXIT_NOT_FOUND,
+    EXIT_PERMISSION_DENIED,
+    EXIT_RATE_LIMIT,
+    EXIT_SERVER_ERROR,
+    EXIT_SUCCESS,
+    EXIT_TIMEOUT,
+    EXIT_USER_INPUT,
+    ErrorSerializer,
+)
+from pal_found_cli.common.log_setup import LogSetup
+from pal_found_cli.common.output_formatter import OutputFormatter
+from pal_found_cli.common.retry import RetryHandler
+from pal_found_cli.common.sdk_error_utils import sdk_exception_exit_code, sdk_http_status
+
+logger = logging.getLogger(__name__)
+
+OperationSpec = dict[str, Any]
+
+_METADATA_ALLOWLIST_PATH = (
+    Path(__file__).resolve().parents[1] / "metadata-allow-list.md"
+)
+
+
+class CLIInputError(ValueError):
+    """A locally generated input error whose message never echoes values."""
+
+    exit_code = EXIT_USER_INPUT
+
+
+class _ArgumentParser(argparse.ArgumentParser):
+    """Raise parser failures into the structured error boundary."""
+
+    def error(self, message: str) -> NoReturn:
+        raise CLIInputError("Invalid command input")
+
+
+def _op(
+    resource: str,
+    operation: str,
+    client_path: tuple[str, ...],
+    method: str,
+    positional: tuple[str, ...] = (),
+    required: tuple[str, ...] = (),
+    optional: tuple[str, ...] = (),
+    *,
+    json_args: frozenset[str] = frozenset(),
+    list_json_args: frozenset[str] = frozenset(),
+) -> OperationSpec:
+    """Build one operation specification entry."""
+    return {
+        "resource": resource,
+        "operation": operation,
+        "client_path": client_path,
+        "method": method,
+        "positional": positional,
+        "required": required,
+        "optional": optional,
+        "json_args": json_args,
+        "list_json_args": list_json_args,
+    }
+
+
+OP_SPECS: tuple[OperationSpec, ...] = (
+    _op(
+        "check",
+        "create",
+        ("Check",),
+        "create",
+        required=("config",),
+        optional=("intent",),
+        json_args=frozenset({"config"}),
+    ),
+    _op(
+        "check",
+        "delete",
+        ("Check",),
+        "delete",
+        ("check_rid",),
+    ),
+    _op(
+        "check",
+        "get",
+        ("Check",),
+        "get",
+        ("check_rid",),
+    ),
+    _op(
+        "check",
+        "replace",
+        ("Check",),
+        "replace",
+        ("check_rid",),
+        required=("config",),
+        optional=("intent",),
+        json_args=frozenset({"config"}),
+    ),
+    _op(
+        "check_report",
+        "get",
+        ("Check", "CheckReport"),
+        "get",
+        ("check_rid", "check_report_rid"),
+    ),
+    _op(
+        "check_report",
+        "get_latest",
+        ("Check", "CheckReport"),
+        "get_latest",
+        ("check_rid",),
+        optional=("limit",),
+    ),
+)
+
+OPERATION_BY_RESOURCE: dict[str, dict[str, OperationSpec]] = {}
+for _spec in OP_SPECS:
+    OPERATION_BY_RESOURCE.setdefault(_spec["resource"], {})[_spec["operation"]] = _spec
+
+
+# Structured JSON arguments across the catalog use the ``-json`` flag suffix.
+_JSON_ARG_NAMES: frozenset[str] = frozenset(
+    name
+    for spec in OP_SPECS
+    for name in (*spec["json_args"], *spec["list_json_args"])
+)
+
+# Integer-typed SDK options.
+_INT_ARG_NAMES: frozenset[str] = frozenset({"limit"})
+
+
+def _common_parser() -> _ArgumentParser:
+    """Build operation-level common options."""
+    parser = _ArgumentParser(add_help=False)
+    parser.add_argument("--timeout", type=int, default=None)
+    parser.add_argument("--format", choices=("json", "toon", "auto"), default="auto")
+    parser.add_argument("--pretty", action="store_true")
+    return parser
+
+
+def _kebab(name: str) -> str:
+    """Convert snake_case names to CLI kebab-case."""
+    return name.replace("_", "-")
+
+
+def _add_kwarg(
+    parser: argparse.ArgumentParser, arg_name: str, *, required: bool
+) -> None:
+    """Add one named operation argument."""
+    if arg_name in _JSON_ARG_NAMES:
+        flag = "--" + _kebab(arg_name) + "-json"
+    else:
+        flag = "--" + _kebab(arg_name)
+    kwargs: dict[str, Any] = {"required": required, "default": None, "dest": arg_name}
+    if arg_name in _INT_ARG_NAMES:
+        kwargs["type"] = int
+    parser.add_argument(flag, **kwargs)
+
+
+def build_parser() -> argparse.ArgumentParser:
+    """Build argparse parser with all 6 Data Health operations."""
+    parser = _ArgumentParser(
+        prog="pal-found-data-health",
+        description="Foundry Data Health CLI - 6 Data Health API v2 operations",
+    )
+    subparsers = parser.add_subparsers(dest="resource", help="Resource type")
+
+    for resource in sorted(OPERATION_BY_RESOURCE):
+        res_parser = subparsers.add_parser(_kebab(resource))
+        op_sub = res_parser.add_subparsers(dest="operation")
+        for operation, spec in sorted(OPERATION_BY_RESOURCE[resource].items()):
+            op_parser = op_sub.add_parser(
+                _kebab(operation), parents=[_common_parser()]
+            )
+            for arg_name in spec["positional"]:
+                op_parser.add_argument(arg_name)
+            for arg_name in spec["required"]:
+                _add_kwarg(op_parser, arg_name, required=True)
+            for arg_name in spec["optional"]:
+                _add_kwarg(op_parser, arg_name, required=False)
+
+    return parser
+
+
+def _spec_for(resource: str, operation: str) -> OperationSpec:
+    """Return one catalog operation specification."""
+    try:
+        return OPERATION_BY_RESOURCE[resource][operation]
+    except KeyError as exc:
+        raise CLIInputError(f"Unknown operation: {resource}.{operation}") from exc
+
+
+def _get_client(root_client: Any, client_path: tuple[str, ...]) -> Any:
+    """Resolve the exact public nested Data Health SDK resource."""
+    client = root_client.data_health
+    for attribute in client_path:
+        client = getattr(client, attribute)
+    return client
+
+
+def _required_text(value: Any, *, field: str) -> str:
+    """Validate a non-empty string without echoing its value."""
+    if not isinstance(value, str) or not value.strip():
+        raise CLIInputError(f"{field} must not be empty")
+    return value
+
+
+def _decode_json(value: str, *, field: str) -> Any:
+    try:
+        return json.loads(value)
+    except json.JSONDecodeError as exc:
+        raise CLIInputError(f"{field} must contain valid JSON") from exc
+
+
+def _parse_json_object(value: str, *, field: str) -> dict[str, Any]:
+    """Decode one JSON object."""
+    result = _decode_json(value, field=field)
+    if not isinstance(result, dict):
+        raise CLIInputError(f"{field} must be a JSON object")
+    return result
+
+
+def _validate_timeout(value: int) -> int:
+    """Validate the effective per-attempt timeout from ADR-002."""
+    if isinstance(value, bool) or not isinstance(value, int) or not 1 <= value <= 3600:
+        raise CLIInputError("timeout must be between 1 and 3600 seconds")
+    return value
+
+
+def _validate_limit(value: int) -> int:
+    """Validate the ``check_report get_latest`` response bound (1..100).
+
+    The installed SDK exposes ``CheckReportLimit`` as a plain int alias
+    without bounds; the server would otherwise reject out-of-range values
+    with a 400. Validating locally gives an earlier, privacy-safe error
+    (ADR-001 exit 1) before any client is created.
+    """
+    if isinstance(value, bool) or not isinstance(value, int) or not 1 <= value <= 100:
+        raise CLIInputError("limit must be between 1 and 100")
+    return value
+
+
+def _validate_inputs(spec: OperationSpec, args: argparse.Namespace) -> None:
+    """Validate scalars and decode JSON arguments before client creation."""
+    for name in (*spec["positional"], *spec["required"]):
+        value = getattr(args, name, None)
+        if isinstance(value, str):
+            _required_text(value, field=name)
+    for name in spec["json_args"]:
+        value = getattr(args, name, None)
+        if value is not None:
+            setattr(args, name, _parse_json_object(value, field=name))
+    for name in spec["optional"]:
+        if name not in _INT_ARG_NAMES or name != "limit":
+            continue
+        value = getattr(args, name, None)
+        if value is not None:
+            setattr(args, name, _validate_limit(value))
+
+
+def _model_to_dict(value: Any) -> Any:
+    """Convert SDK models to JSON-compatible values."""
+    if value is None:
+        return None
+    if isinstance(value, list):
+        return [_model_to_dict(item) for item in value]
+    if isinstance(value, dict):
+        return {key: _model_to_dict(item) for key, item in value.items()}
+    for method_name in ("to_dict", "model_dump", "dict"):
+        method = getattr(value, method_name, None)
+        if callable(method):
+            result = method()
+            if isinstance(result, dict):
+                return {key: _model_to_dict(item) for key, item in result.items()}
+    return value
+
+
+def _build_kwargs(
+    spec: OperationSpec, args: argparse.Namespace, timeout: int
+) -> dict[str, Any]:
+    """Build SDK keyword arguments, omitting absent optional values."""
+    kwargs: dict[str, Any] = {}
+    for name in (*spec["required"], *spec["optional"]):
+        value = getattr(args, name, None)
+        if value is not None:
+            kwargs[name] = value
+    kwargs["request_timeout"] = timeout
+    return kwargs
+
+
+async def _invoke(
+    spec: OperationSpec,
+    client: Any,
+    args: argparse.Namespace,
+    timeout: int,
+) -> Any:
+    """Invoke one SDK operation."""
+    positional = [getattr(args, name) for name in spec["positional"]]
+    return await getattr(client, spec["method"])(
+        *positional, **_build_kwargs(spec, args, timeout)
+    )
+
+
+def _serialize_error(exception: BaseException) -> int:
+    """Write one privacy-safe ADR-001 error envelope to stdout."""
+    serializer = ErrorSerializer()
+    status = sdk_http_status(exception)
+    declared = getattr(exception, "exit_code", None)
+    sdk_exit = sdk_exception_exit_code(exception)
+    if isinstance(declared, int) and 1 <= declared <= 9:
+        exit_code = declared
+    elif sdk_exit is not None:
+        exit_code = sdk_exit
+    elif status == 401:
+        exit_code = EXIT_AUTH
+    elif status == 403:
+        exit_code = EXIT_PERMISSION_DENIED
+    elif status == 404:
+        exit_code = EXIT_NOT_FOUND
+    elif status == 429:
+        exit_code = EXIT_RATE_LIMIT
+    elif isinstance(exception, ConfigurationError):
+        exit_code = EXIT_CONFIGURATION
+    elif isinstance(exception, (TypeError, ValueError)):
+        exit_code = EXIT_USER_INPUT
+    elif isinstance(exception, TimeoutError):
+        exit_code = EXIT_TIMEOUT
+    else:
+        exit_code = EXIT_SERVER_ERROR
+
+    safe = isinstance(
+        exception,
+        (CLIInputError, ConfigurationError, TimeoutError),
+    ) or declared in {EXIT_ACCESS_CONTROL, EXIT_CONFIGURATION}
+    message = str(exception) if safe else "DataHealth operation failed"
+    envelope = serializer.create_error_envelope(
+        exit_code, message, type(exception).__name__, serializer.call_id
+    )
+    if isinstance(status, int):
+        envelope["http_status"] = status
+    sys.stdout.write(json.dumps(envelope, default=str) + "\n")
+    sys.stdout.flush()
+    logger.error(
+        "DataHealth operation failed",
+        extra={
+            "call_id": serializer.call_id,
+            "exception_type": type(exception).__name__,
+            "http_status": status,
+        },
+    )
+    return exit_code
+
+
+async def main() -> int:
+    """Run one stateless Foundry Data Health invocation."""
+    parser = build_parser()
+    try:
+        args = parser.parse_args()
+        if not args.resource or not getattr(args, "operation", None):
+            raise CLIInputError("a DataHealth operation is required")
+    except CLIInputError as exc:
+        return _serialize_error(exc)
+
+    try:
+        cfg = ConfigLoader()
+        cfg.load()
+        LogSetup.configure(log_level=cfg.log_level)
+
+        resource = args.resource.replace("-", "_")
+        operation = args.operation.replace("-", "_")
+        spec = _spec_for(resource, operation)
+        _validate_inputs(spec, args)
+        timeout = _validate_timeout(
+            args.timeout if args.timeout is not None else cfg.timeout_s
+        )
+        AccessControlGuard(
+            cfg,
+            "DATA_HEALTH",
+            metadata_allowlist_path=str(_METADATA_ALLOWLIST_PATH),
+        ).check(resource, operation)
+
+        factory = AsyncClientFactory()
+        with factory.invocation_scope(cfg, include_attribution=False):
+            root_client = factory.create(cfg, include_attribution=False)
+            client = _get_client(root_client, spec["client_path"])
+            retry_handler = RetryHandler(timeout_s=timeout)
+            result = await retry_handler.execute(
+                _invoke, spec, client, args, timeout
+            )
+
+        formatter = OutputFormatter(format_setting=args.format, pretty=args.pretty)
+        print(formatter.format(_model_to_dict(result)))
+        return EXIT_SUCCESS
+    except AccessControlError as exc:
+        ErrorSerializer().serialize(exc)
+        return EXIT_ACCESS_CONTROL
+    except asyncio.CancelledError:
+        return _serialize_error(TimeoutError("Operation cancelled"))
+    except Exception as exc:
+        return _serialize_error(exc)
+
+
+def console_main() -> int:
+    """Run the async CLI through one event-loop boundary."""
+    return asyncio.run(main())
+
+
+if __name__ == "__main__":
+    raise SystemExit(console_main())

pal_found_cli/datasets/__init__.py

@@ -0,0 +1 @@
+"""Foundry Datasets CLI package."""

pal_found_cli/datasets/scripts/__init__.py

@@ -0,0 +1 @@
+"""Console entry point modules for Foundry Datasets."""