Add batch mode API for efficient mass policy operations

[Copilot]

Mass policy operations (thousands of removePolicy()/removeGroupingPolicy() calls) cause severe performance degradation because each operation triggers database write, role link rebuild, and watcher notification. Users were also unnecessarily calling loadPolicy() after each modification, compounding the problem.

Changes

New Batch Mode API

• startBatch(): Suspends autoSave, autoBuildRoleLinks, and autoNotifyWatcher
• endBatch(): Restores settings, performs single save and role link rebuild
• Handles adapters that don't implement savePolicy() gracefully

Documentation

• Added batch mode usage section to README
• Clarified that loadPolicy() is unnecessary after modifications when autoSave is enabled

Usage

// Instead of 1000 DB writes + 1000 role rebuilds
enforcer.startBatch();
for (const policy of policies) {
  await enforcer.removePolicy(...policy);
}
await enforcer.endBatch();  // 1 save + 1 rebuild

Performance improvement: O(N) → O(1) for N policy operations.

100% backward compatible - existing code unaffected.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Without enforcer.loadPolicy() Casbin behave unpredictably on mass updates.</issue_title>
<issue_description>In our ephemeral environments, we reconstruct permissions by frequently invoking methods like enforcer.removePolicy() and enforcer.removeGroupingPolicy() thousands of times. However, each change necessitates a call to enforcer.loadPolicy(). This is increasingly problematic as the service gets progressively burdened with the growing number of policies to reload with each update, ultimately leading to service downtime.

enforcer is setup as follows:

export const casbinProviders = [
  {
    provide: PROVIDER_CASBIN_ENFORCER,
    useFactory: async (dataSource: DataSource) => {
      const adapter = await TypeORMAdapter.newAdapter({ connection: dataSource });
      const enforcer = await newEnforcer(newModelFromString(PERMISSIONS_MODEL), adapter);
      await enforcer.addFunction('isMatch', casbinIsMatch);
      await enforcer.addFunction('includes', casbinIncludes);
      await enforcer.loadPolicy();
      await enforcer.enableAutoSave(true);
      return enforcer;
    },
    inject: [getDataSourceToken()]
  }
];

In our production environment, the substantial number of policies we manage also poses a significant challenge. The frequent use of methods like enforcer.removePolicy() and enforcer.removeGroupingPolicy(), followed by enforcer.loadPolicy() after each modification, is proving to be problematic. As the number of policies grows, these operations increasingly strain the server during normal usage, leading to severe performance issues and potential service disruptions.</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes Without enforcer.loadPolicy() Casbin behave unpredictably on mass updates. #463

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}