awoooi/apps/api/src/services/javascript_package_inventory.py

"""
JavaScript / pnpm 套件盤點快照。

只讀取已提交的 JSON 快照；不安裝套件、不升級套件、不寫 lockfile、
不呼叫外部 CVE / audit 服務、不改生產路由。
"""

from __future__ import annotations

import json
from pathlib import Path
from typing import Any

from src.services.snapshot_paths import default_evaluations_dir

_DEFAULT_EVALUATIONS_DIR = default_evaluations_dir(Path(__file__))
_SNAPSHOT_PATTERN = "javascript_package_inventory_*.json"
_SCHEMA_VERSION = "javascript_package_inventory_v1"


def load_latest_javascript_package_inventory(
    evaluations_dir: Path | None = None,
) -> dict[str, Any]:
    """載入最新已提交的 JavaScript / pnpm 套件盤點快照。"""
    directory = evaluations_dir or _DEFAULT_EVALUATIONS_DIR
    candidates = sorted(directory.glob(_SNAPSHOT_PATTERN))
    if not candidates:
        raise FileNotFoundError(f"no JavaScript package inventory snapshots found in {directory}")

    latest = candidates[-1]
    with latest.open(encoding="utf-8") as handle:
        payload = json.load(handle)

    if not isinstance(payload, dict):
        raise ValueError(f"{latest}: expected JSON object")
    _require_schema(payload, _SCHEMA_VERSION, str(latest))
    _require_read_only_boundaries(payload, str(latest))
    _require_operation_boundaries(payload, str(latest))
    _require_rollup_consistency(payload, str(latest))
    return payload


def _require_schema(payload: dict[str, Any], expected: str, label: str) -> None:
    actual = payload.get("schema_version")
    if actual != expected:
        raise ValueError(f"{label}: expected schema_version={expected}, got {actual!r}")


def _require_read_only_boundaries(payload: dict[str, Any], label: str) -> None:
    program_status = payload.get("program_status") or {}
    if program_status.get("read_only_mode") is not True:
        raise ValueError(f"{label}: program_status.read_only_mode must be true")

    boundaries = payload.get("approval_boundaries") or {}
    blocked_flags = {
        "sdk_installation_allowed",
        "paid_api_call_allowed",
        "shadow_or_canary_allowed",
        "production_routing_allowed",
        "destructive_operation_allowed",
    }
    allowed = sorted(flag for flag in blocked_flags if boundaries.get(flag) is not False)
    if allowed:
        raise ValueError(f"{label}: approval boundaries must remain false: {allowed}")


def _require_operation_boundaries(payload: dict[str, Any], label: str) -> None:
    boundaries = payload.get("operation_boundaries") or {}
    if boundaries.get("read_only_api_allowed") is not True:
        raise ValueError(f"{label}: read_only_api_allowed must be true")

    blocked_flags = {
        "package_installation_allowed",
        "package_upgrade_allowed",
        "lockfile_write_allowed",
        "external_cve_lookup_allowed",
        "npm_audit_allowed",
        "pnpm_install_allowed",
        "production_routing_allowed",
    }
    allowed = sorted(flag for flag in blocked_flags if boundaries.get(flag) is not False)
    if allowed:
        raise ValueError(f"{label}: operation boundaries must remain false: {allowed}")

    lockfile_summary = payload.get("lockfile_summary") or {}
    if lockfile_summary.get("write_allowed") is not False:
        raise ValueError(f"{label}: lockfile_summary.write_allowed must be false")


def _require_rollup_consistency(payload: dict[str, Any], label: str) -> None:
    workspaces = payload.get("workspaces") or []
    rollups = payload.get("rollups") or {}
    total = rollups.get("total_workspaces")
    if total != len(workspaces):
        raise ValueError(f"{label}: rollups.total_workspaces must equal workspaces length")

    action_required = set(rollups.get("action_required_workspace_ids") or [])
    actual_action_required = {
        workspace.get("workspace_id")
        for workspace in workspaces
        if workspace.get("status") == "action_required"
    }
    if action_required != actual_action_required:
        raise ValueError(
            f"{label}: rollups.action_required_workspace_ids must match action_required workspaces"
        )

    planned_next = set(rollups.get("planned_next_workspace_ids") or [])
    actual_planned_next = {
        workspace.get("workspace_id")
        for workspace in workspaces
        if workspace.get("status") == "planned_next"
    }
    if planned_next != actual_planned_next:
        raise ValueError(
            f"{label}: rollups.planned_next_workspace_ids must match planned_next workspaces"
        )

    total_dependencies = sum(
        (workspace.get("dependency_counts") or {}).get("total", 0)
        for workspace in workspaces
    )
    if rollups.get("total_direct_dependencies") != total_dependencies:
        raise ValueError(
            f"{label}: rollups.total_direct_dependencies must equal workspace dependency totals"
        )

    drift = payload.get("lockfile_drift") or {}
    if rollups.get("manifest_lock_mismatch_count") != len(drift.get("specifier_mismatches") or []):
        raise ValueError(
            f"{label}: rollups.manifest_lock_mismatch_count must equal specifier_mismatches length"
        )
    if rollups.get("missing_in_lockfile_count") != len(drift.get("missing_in_lockfile") or []):
        raise ValueError(
            f"{label}: rollups.missing_in_lockfile_count must equal missing_in_lockfile length"
        )
    if rollups.get("extra_in_lockfile_count") != len(drift.get("extra_in_lockfile") or []):
        raise ValueError(
            f"{label}: rollups.extra_in_lockfile_count must equal extra_in_lockfile length"
        )