feat: add MiniMax as first-class LLM provider

[octo-patch]

Summary

Adds MiniMax as a fully integrated LLM provider in Pezzo, on par with OpenAI. MiniMax offers an OpenAI-compatible API with powerful models (M2.7 with 1M token context, M2.5 with 204K context).

Changes

• Type System (libs/types): Add MiniMax to Provider enum and MiniMaxChatCompletion to PromptService enum with provider details and prompt mappings
• Console UI (apps/console):
  • Add MiniMax to provider selector (fully available, not "Coming Soon")
  • Add MiniMax chat completion settings definition (model selector, temperature 0.01-1.0, max tokens, top_p)
  • Add MiniMax to API key management providers list
  • Add MiniMax logo asset
• Proxy Server (apps/proxy): Add MiniMaxV1Handler and /minimax/v1/* router for proxying requests to https://api.minimax.io/v1 with observability and caching support
• Server (apps/server): Add MiniMax cost calculation in report builder with per-model pricing (M2.7, M2.7-highspeed, M2.5, M2.5-highspeed)
• Client SDK (libs/client): Add PezzoMiniMax class wrapping OpenAI SDK with MiniMax base URL, temperature clamping, and full observability integration
• README: Add supported LLM providers table listing MiniMax alongside OpenAI

MiniMax Models Supported

Model	Context Window	Use Case
MiniMax-M2.7	1M tokens	Latest flagship model
MiniMax-M2.7-highspeed	1M tokens	Fast inference variant
MiniMax-M2.5	204K tokens	Balanced performance
MiniMax-M2.5-highspeed	204K tokens	Fast inference variant

Key Implementation Details

• MiniMax uses an OpenAI-compatible API (https://api.minimax.io/v1), so the PezzoMiniMax client wraps the OpenAI SDK with a custom base URL
• Temperature is clamped to (0.01, 1.0] range as required by MiniMax API
• Cost tracking uses per-model pricing tiers
• Full proxy support with caching and observability reporting

Test Plan

• 13 unit + integration tests for provider types (enum values, provider details, mappings)
• 9 unit + integration tests for server report builder (cost calculation, model variants, fallback pricing)
• 6 unit tests for proxy handler (API proxying, temperature clamping, cache hits, error handling, properties parsing)
• All 28 tests passing via npx nx test types && npx nx test server && npx nx test proxy

19 files changed, 1218 additions