Feature: Security levels

[axelander]

Summary

This PR introduces the concept of security levels, enhancing the control over which AI models can be utilized within different Spaces based on their security requirements. Security levels are defined by a name, description, and a numeric value that dictates access permissions. This feature allows administrators to assign security levels to both models and Spaces, ensuring that only models meeting the required security criteria are accessible within a given Space.

Notes

• Changing security level on a space disables models that will no longer be available. The existing validation of model availability on assistants remains the same.
• Running pnpm install updates frontend/pnpm-lock.yaml by using ' instead of ", also some packages are removed (widgets for example). This PR does not add/change any dependencies so the updated file is not included in this PR.
• The frontend/packages/intric-js/src/types/schema.d.ts does not seem to be up-to-date. We get a bigger diff than expected that contains unrelated changes. Not sure how this is supposed to be managed, but we have manually changed the frontend/packages/intric-js/update.js to use the local backend when regenerating.

Screenshots

List security levels

Create/edit security level

Space security level admin

Change space security level message