Add a reading-stats endpoint

Aggregate numbers are the entire reason someone keeps a tracker running for months instead of abandoning it after a week -- "how many books this year," "what's my average rating" are the kind of small satisfying answers that make the habit stick. Forcing every client to fetch every row and compute this itself means: it gets recomputed slightly differently by every client that bothers to build it, it doesn't scale as someone's library grows, and most clients will just... not bother, so the feature effectively doesn't exist even though "the data is all there."

There's no way to answer "how many books have I read, and what's my average rating?" without fetching every single book and computing it client-side. A stub action already exists at GET /api/books/stats/ (books/views.py), but calling it unconditionally raises NotImplementedError.

curl -s http://127.0.0.1:8000/api/books/stats/

-> 500 Internal Server Error

books/views.py's BookViewSet.stats unconditionally raises NotImplementedError

This is a detail=False DRF @action -- it operates on the whole collection, not one book, so unlike the "start reading" action there's no get_object() involved. The interesting part is the aggregation: Django's Avg() aggregate over an empty/all-null queryset returns None, not 0 -- the implementation needs to pass that None straight through rather than coercing it, and needs to make sure the Avg() is computed only over read books with a non-null rating, not over every book.

Look at BookViewSet.stats in books/views.py, and at django.db.models.Count/Avg used with Book.objects -- .filter() + .aggregate(), or .values("status").annotate(...), are both reasonable starting points depending on how you want to shape the counts.