feat(gcp): migrate AutoML to Vertex AI + implement LlmProvider

[Priyanshu-rgbb]

closes #53

Summary--->

Migrates GCP AI module from deprecated AutoML to Vertex AI and implements the LlmProvider trait with Gemini model support.

Problem--->

Google is sunsetting standalone AutoML in favor of Vertex AI. Our gcp_automl.rs module uses deprecated endpoints that will stop working.

Solution

• Created VertexAI struct implementing full LlmProvider trait
• Migrated all dataset/model management methods to Vertex AI endpoints
• Added Gemini support: text generation, streaming, embeddings, function calling
• Marked old AutoML struct as #[deprecated] (non-breaking)

Changes--->

New files:

• gcp_vertex_ai.rs - Core implementation with LlmProvider trait
• gcp_vertex_ai_types.rs - 19 serde types for Vertex AI/Gemini API
• gcp_vertex_ai_operations.rs - 16 test cases

Modified:

• Cargo.toml - Added reqwest json/stream features
• mod.rs files - Module registration
• gcp_automl.rs - Added deprecation warning

Checklist--->

• Implements LlmProvider trait for GCP
  -Migrates AutoML methods to Vertex AI
  -Includes test coverage
• Non-breaking (old code compiles with warning)
  -Passes cargo fmt