Add all 15-language indexing script, update Colab notebook, and commit full 148k passage indexes

Add unignored en/hi/mr corpus, Colab GPU indexing script and notebook, and pipeline updates

feat: extend Indic RAG pipeline to all 14 Indic languages + English with FAISS indexes, centroids, and multi-language tests

fix(pipeline): critical network killswitch, FAISS self-healing startup, post-retrieval confidence calibration, and semantic answer cache

feat(core): complete voice-rag-indic pipeline with STT, FAISS HNSW, BM25 rerank, guardrails, FastAPI, benchmarks, and demo

feat(data): add build_corpus.py and augment_longdocs.py with dynamic language iteration, broad passage deduplication, and long doc generation