Add books languages endpoint

[aviralgarg05]

Summary

This PR adds a new public endpoint to expose the list of ISO639-3 language codes currently used by books managed in the CMS.

The new endpoint is:

GET /v1/books/languages

It returns a sorted list of distinct language codes collected from production books only.

Example response:

{
  "languages": ["deu", "eng", "fra", "spa"]
}

What changed

• added GET /v1/books/languages under the existing books API
• added a backend helper to collect language codes from Book.zim_metadata["Language"]
• restricted the result set to books with location_kind = "prod"
• split comma-separated language values
• trimmed surrounding whitespace
• ignored empty values
• removed duplicates
• returned the final list in sorted order
• added API and DB-level tests for the new behavior

Why

The Language metadata can contain multiple comma-separated ISO639-3 codes. Exposing the normalized set of codes used by production books makes it easier for downstream consumers to know which values they may encounter and validate their support accordingly.

Notes

• only production books are considered
• the endpoint is public, in line with the intended use of this metadata
• the route is placed under /v1/books because the source of truth is Book.zim_metadata

Testing

I verified the change with:

• hatch run lint:black
• hatch run lint:ruff
• hatch run check:pyright
• targeted DB test for language extraction
• targeted API test for the new endpoint
• full backend test suite with CI-equivalent environment settings
• backend Docker image builds for API, mill, and shuttle
• container smoke checks for mill and shuttle entrypoints
• API container healthcheck after applying Alembic migrations

Full backend result:

• 267 passed

Related

Closes #102