feat(deploy): add Coolify-ready compose file and deployment guide

[Bumpfi]

Summary

Deploying Odysseus on Coolify currently fails in three independent ways: Coolify's compose parser mangles the ${APP_DATA_DIR:-./data}/ssh-style bind-mount sources in docker-compose.yml into empty strings (invalid spec: :/app/.ssh:z: empty section between colons); the image is built from source on the server, which takes 10+ minutes and can OOM small VPSes even though CI already publishes a multi-arch image to GHCR; and the 127.0.0.1 host-port bindings are unreachable behind Coolify's Traefik proxy while SearXNG's host port 8080 collides with other stacks. This PR adds a separate, opt-in docker-compose.coolify.yml that fixes all three — pre-built ghcr.io/pewdiepie-archdaemon/odysseus:latest image, named volumes only, no host port bindings, and Coolify magic variables for domain routing (SERVICE_FQDN_ODYSSEUS_7000), a generated admin password, and the CORS origin — plus a step-by-step guide in docs/deploy/coolify.md. docker-compose.yml has a zero-line diff, so nothing changes for existing users.

Related: #3835 addresses some of the same symptoms but does so by changing the existing compose defaults (bind mounts → named volumes, SearXNG port remap), which would affect current deployments. This PR is purely additive and does not conflict with it.

Target branch

• This PR targets dev, not main.

Linked Issue

Fixes #5384

Type of Change

• Bug fix (non-breaking — fixes a confirmed issue)
• New feature (non-breaking — adds new behaviour)
• Breaking change (changes or removes existing behaviour)
• Refactor / cleanup (behaviour unchanged)
• Documentation only
• CI / tooling / configuration

Checklist

• I searched open issues and open PRs — this is not a duplicate (closest is fix(docker): resolve searxng crash and coolify data persistence #3835, see Summary for how this differs).
• This PR targets dev
• My changes are limited to the scope described above — no unrelated refactors or whitespace changes mixed in.
• I actually ran the app and verified the change works end-to-end (see How to Test — full stack booted from the new compose file with the published GHCR image).

How to Test

1. Validate both compose files parse: docker compose -f docker-compose.yml config -q and docker compose -f docker-compose.coolify.yml config -q (the second prints warnings about unset SERVICE_* variables — those are Coolify magic variables, filled in by Coolify at deploy time).
2. Confirm existing users are unaffected: git diff dev -- docker-compose.yml is empty.
3. Boot the Coolify stack locally, simulating the two magic variables:
   SERVICE_PASSWORD_ADMIN=testpass123 SERVICE_URL_ODYSSEUS=http://localhost docker compose -f docker-compose.coolify.yml -p coolifytest up -d
4. Wait for docker compose -p coolifytest ps to show odysseus and searxng as healthy, then verify from inside the network (no host ports are published by design):
   • docker exec coolifytest-odysseus-1 curl -s -o /dev/null -w "%{http_code}" http://localhost:7000/ → 302 (redirect to login)
   • docker exec coolifytest-odysseus-1 curl -s -o /dev/null -w "%{http_code}" "http://searxng:8080/search?q=test&format=json" → 200
   • login as admin / testpass123 via POST /api/auth/login → 200
5. Clean up: docker compose -p coolifytest down -v
6. Optional, on a real Coolify v4 instance: New → Public Repository → this repo, Build Pack "Docker Compose", Compose Location /docker-compose.coolify.yml, deploy — follows docs/deploy/coolify.md.

Visual / UI changes — REQUIRED if you touched anything that renders

N/A — no UI or rendering changes; compose file and docs only.