Add production-ready LocalEGA deployment guide

[Aberdur]

PR Description

This PR adds a comprehensive installation and operations guide for deploying LocalEGA with rootless Podman in a production environment.

It also aligns the Docker deployment configuration with the documented production workflow.

Changes

• Add LEAME.md, covering:

  • rootless Podman requirements and subordinate ID mappings;
  • separation between NFS data and local persistent storage;
  • SELinux and filesystem considerations;
  • Crypt4GH service and master key generation;
  • environment and PostgreSQL configuration;
  • container image builds and Vault DB initialisation;
  • staged service startup;
  • Inbox retention cleanup;
  • DNS, RabbitMQ, PostgreSQL and permission checks;
  • functional validation, operations, backups and troubleshooting.

• Clean .env.example:

  • remove deployment-specific values;
  • remove unused TLS configuration variables;
  • remove Fake CEGA-only variables;
  • add an explicit affiliate name;
  • use generic CEGA endpoint and AMQPS placeholders;
  • use a generic configurable SSH banner;
  • default application logging to info.

• Keep the production Compose deployment limited to:

  • inbox;
  • mq;
  • handler;
  • inbox-cleaner;
  • vault-db;
  • distribution;
  • nss-sync.

• Remove Fake CEGA services from the production Compose file.

• Remove the obsolete docker-compose.yml.sample configuration.

• Clarify that the contents of deploy/docker/cega are development-only fixtures.

• Update the existing deployment documentation to use .env.example and lega.ini.sample.