fix: resolve model role confusion by using chat completion API, and fix emergency card rendering

fix: switch model to gemma-1.1-2b-it for compat with llama-cpp-python GGUF loader

feat: unified Wilderness AI chatbot & 1s GPS updates with proximity highlights

Commit local modifications: switch fallback ASR to transformers and update requirements

Fix: bundle Trento route cache with POIs to avoid Overpass API calls on HF Space

Fix elevation, POIs, and alerts by bundling pre-compiled track cache and adding Gradio allowed_paths

Implement customizable post-trek storyteller styles (Minimal Technical Gist & Social Media Post) with robust mock parser fallback

Fix voice ASR fallback to transformers on Hugging Face and stop simulation timer at trek completion

Remove Docker support, configure Hugging Face for Gradio SDK, add database.py, stage background map tile downloader

feat: smooth marker animations via postMessage and enhanced POI categorization

feat: Phase 2 - Functional simulation, OSM POI Overpass integration, and first-aid RAG

Initial commit: Set up Trailhead structure, Docker, README, and .gitignore