Add book-domain recommender demo

[darshan-Jahagirdar]

Refs aolabsai/ao_pyth#9

This adds a distinct book-domain recommender demo for the paid bounty. It keeps the implementation self-contained and reviewable:

• local 10-book catalog with genre, pace, length, tone, format, and reader-context labels
• deterministic binary input encoding for book attributes plus current reader mood
• deterministic fallback ranking so reviewers can run it without paid APIs or private AO packages
• optional AO Agent integration through arch__BookRecommender.py when ao_core and ao_arch are installed
• Streamlit demo in book_recommender.py
• README usage and validation notes
• unit tests for encoding shape and recommendation behavior

Validation run locally:

• python -m unittest discover -s tests
• python -m py_compile book_domain.py book_recommender.py arch__BookRecommender.py tests/test_book_domain.py
• git diff --cached --check before commit

This uses only a local sample catalog; no paid APIs, accounts, or cloud services are required.

[darshan-Jahagirdar]

Pushed e704cf1 with two small reviewability/compatibility cleanups:

• deterministic fallback ranking now sorts by score descending and title ascending for stable tie handling
• removed st.container(border=True) so the demo remains compatible with older Streamlit versions

Revalidated locally:

• python -m unittest discover -s tests
• python -m py_compile book_domain.py book_recommender.py arch__BookRecommender.py tests/test_book_domain.py
• git diff --check

[darshan-Jahagirdar]

Validation update for this PR:

python -B -m unittest tests.test_book_domain
git diff --check

Result: 3 tests passed and the diff whitespace check passed locally.