feat(content): add Ollama cheat sheet and icon - #798

Summary of Changes

Adds a comprehensive quick reference cheat sheet and SVG icon for Ollama (local LLM runtime):

• Cheat sheet (source/_posts/ollama.md):
  • Getting Started: Multi-platform installation (macOS, Linux, Windows, Docker) and first run commands.
  • CLI Commands: Complete subcommands reference and flags (--verbose, --format json, --keepalive).
  • Interactive Chat: Session controls (/set system, /set parameter, /save, /load, /show info, /clear, multiline """ prompts, piping).
  • Popular Models Matrix: Parameter sizes, context windows, and recommended use cases (Llama 3.3, Llama 3.2, DeepSeek-R1, Qwen 2.5-Coder, Mistral, Gemma 2, Phi-4, LLaVA, Nomic-Embed-Text).
  • Modelfile: Full specification table, runtime parameters (temperature, num_ctx, repeat_penalty, stop, seed), and runnable example.
  • REST API: Complete endpoints overview (/api/generate, /api/chat, /api/embed, JSON mode).
  • OpenAI Compatibility: Built-in /v1 endpoint integration for LangChain, LlamaIndex, Cursor, and Continue.
  • SDKs: Python and JavaScript/TypeScript code examples with streaming and embeddings.
  • Server Configuration: Environment variables (OLLAMA_HOST, OLLAMA_MODELS, OLLAMA_KEEP_ALIVE, OLLAMA_FLASH_ATTENTION, etc.).
  • Deployment: Docker with CPU and GPU acceleration (NVIDIA --gpus=all, AMD ROCm) + Open WebUI setup.
  • Troubleshooting & Prompting: Common errors resolution and practical prompt patterns.
• Icon (source/assets/icon/ollama.svg): Official vector SVG logo with currentColor filling for light/dark theme adaptation.
• Index (README.md): Registered Ollama under the Toolkit section.