feat: Add GPT OSS 20B and 120B

[jcabrero]

This PR adds docker compose files for GPT OSS 20B and 120B. Additionally it adds small fixes to two small problems.

[jcabrero]

This change is the most relevant. If the MODEL_CONCURRENT_RATE_LIMIT doesn't exist for such given model, it switches to "default" which should work for any model and otherwise 50. This prevents a failure state in most cases.

[Copilot]

Pull Request Overview

This PR adds support for GPT OSS 20B and 120B models by creating their docker compose configurations, while also implementing defensive programming fixes for model validation and rate limiting.

• Adds docker compose files for GPT OSS 20B and 120B model deployments
• Implements null/empty string validation for model IDs in the state management
• Replaces exception-based rate limiting with default fallback logic

Reviewed Changes

Copilot reviewed 8 out of 9 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
nilai-api/src/nilai_api/state.py	Adds null/empty validation for model_id parameter
nilai-api/src/nilai_api/routers/private.py	Replaces KeyError exception with default fallback for rate limits
nilai-api/src/nilai_api/config/config.yaml	Adds rate limit configuration for new GPT OSS 20B model and default
docker/vllm.Dockerfile	Updates base image to custom jcabrero/vllm version
docker/compose/docker-compose.gpt-20b-gpu.yml	New docker compose configuration for GPT OSS 20B
docker/compose/docker-compose.gpt-120b-gpu.yml	New docker compose configuration for GPT OSS 120B
.env.sample	Adds BRAVE_SEARCH_API environment variable
.env.ci	Adds BRAVE_SEARCH_API environment variable for CI

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}