Implement "ai/" as the default namespace for model references

[Copilot]

This PR implements the feature requested in docker#126 to make "ai/" the default namespace for model references, allowing users to run commands like docker model run gpt-oss without explicitly specifying the "ai/" prefix.

Changes Made

Added a smart namespace defaulting system that automatically prepends "ai/" to simple model names while preserving existing behavior for qualified references:

Core Implementation

• Added addDefaultNamespace() function with intelligent logic to determine when to apply the default namespace
• Updated 4 client methods to use the default namespace:
  • PullModel() - for pulling models from registry
  • GetModel() - for retrieving stored models
  • IsModelInStore() - for checking model existence
  • DeleteModel() - for removing models

Smart Logic

The namespace defaulting only applies to simple model names, preserving existing behavior for:

• Models with existing namespaces: my-namespace/model → unchanged
• Repository:tag references: some-repo:v1.0 → unchanged
• Registry references: registry.com/namespace/model:tag → unchanged
• SHA256 digests: sha256:abcd1234... → unchanged

Examples

# Simple model names get ai/ prefix automatically
docker model pull gpt-oss          # → ai/gpt-oss
docker model get llama2             # → ai/llama2

# Qualified references remain unchanged  
docker model pull ai/gpt-oss        # → ai/gpt-oss
docker model pull custom/model:v1   # → custom/model:v1
docker model pull repo.com/ns/model # → repo.com/ns/model

Testing

Added comprehensive test coverage including:

• Unit tests for the namespace function logic with all edge cases
• Integration tests demonstrating end-to-end functionality
• Verification that existing behavior is preserved for all qualified reference types

This implementation follows the minimal-change principle and maintains backward compatibility while providing the convenience feature similar to how Ollama handles the "library/" namespace.

Closes docker#126

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.