Add Manim scenes for DB concepts and NN line-fit animations

[coding-parrot]

Motivation

• Provide a set of visual, animated illustrations for common backend concepts and a simple neural-network example to use in presentations and docs.
• Standardize render resolution and frame size across scenes so tall vertical layouts remain visible by setting config.pixel_width, config.pixel_height, and config.frame_height/frame_width.

Description

• Add db_connection_pool_diagram.py implementing a DBConnectionPoolDiagram Manim Scene that animates clients acquiring and releasing pooled DB connections with routing arrows and status text.
• Add hierarchical_wal_replication.py implementing HierarchicalWALReplication which visualizes a primary and cascading read replicas with WAL event dots propagating through arrows and node highlighting.
• Add neural_network_line_fit.py implementing NeuralNetworkLearnsLine that shows a tiny computation graph for ŷ = m x + C, animated forward passes, and epoch-style parameter/ loss value updates.
• Add primary_to_read_replicas.py implementing PrimaryToReadReplicas which lays out a primary and a stacked list of read replicas (including adding a 51st replica) and animates request waves and incoming write impacts on the primary.
• Add request_collapsing_cache.py implementing RequestCollapsingInCache that demonstrates cache request collapsing with an in-cache promise, single DB fetch, and fanout of the resolved value to multiple clients.

Testing

• No automated tests were added or run as part of this change; CI/static checks can validate imports and rendering as needed.

---

Codex Task