Add memba MVP: C++ core, Python SDK, CLI, examples, experiments

C++ core (libmemba.so):
- include/memba/state.h — C API (state_new/free/save/load/get_size)
- src/state.cpp — MEMB file format: magic, version, SHA-256 model_id,
  CRC-32, opaque llama_state_*_data() blob
- src/cli.cpp — minimal demo binary with greedy sampler
- CMakeLists.txt + build.sh with llama.cpp submodule, CUDA auto-detect

Python SDK (memba):
- core.py — file I/O via llama-cpp-python's exposed C functions,
  unwraps _LlamaContext to access raw context pointer (≥0.3.x)
- session.py — high-level Session with auto-save/load, ChatML wrapper
  for instruct models, raw mode for base models
- cli.py — typer-based: chat (REPL), run (one-shot), list, rm, info

Examples:
- 01_basic_save_load.py, 02_chat_session.py

Experiments (throwaway POCs documenting product-direction findings):
- recall_poc.py — git log → state → cross-process query
- mood_poc.py — batch sentiment trajectory, Mamba vs Transformer
- mood_stream_poc.py, mood_batch_poc.py — variants
- diag_saveload.py — minimal save/load isolation test
- README.md documents the headline finding: save/load is byte-identical,
  but Falcon-Mamba-7B-Instruct does not retain facts across conversation
  turns even in-process — limits viable products to single-prompt analysis
  and persona priming.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

build.sh

@@ -0,0 +1,44 @@
+#!/usr/bin/env bash
+# build.sh — build libmemba.so and memba-cli
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+BUILD_DIR="${SCRIPT_DIR}/build"
+
+# ── Submodule ────────────────────────────────────────────────────
+if [[ ! -f "${SCRIPT_DIR}/llama.cpp/CMakeLists.txt" ]]; then
+    echo "==> Initialising llama.cpp submodule…"
+    git -C "${SCRIPT_DIR}" submodule update --init --recursive
+fi
+
+# ── GPU detection ─────────────────────────────────────────────────
+EXTRA_ARGS=()
+if command -v nvcc &>/dev/null; then
+    echo "==> CUDA found ($(nvcc --version | grep release | awk '{print $6}')) — enabling GGML_CUDA"
+    EXTRA_ARGS+=("-DGGML_CUDA=ON")
+elif [[ "$(uname)" == "Darwin" ]]; then
+    echo "==> macOS — enabling Metal"
+    EXTRA_ARGS+=("-DGGML_METAL=ON")
+else
+    echo "==> No GPU backend detected — CPU-only build"
+fi
+
+# Relay any extra cmake args from the command line
+EXTRA_ARGS+=("$@")
+
+# ── Configure & build ─────────────────────────────────────────────
+mkdir -p "${BUILD_DIR}"
+cmake -S "${SCRIPT_DIR}" -B "${BUILD_DIR}" \
+    -DCMAKE_BUILD_TYPE=Release \
+    "${EXTRA_ARGS[@]}"
+
+cmake --build "${BUILD_DIR}" --config Release -j "$(nproc 2>/dev/null || sysctl -n hw.ncpu)"
+
+# ── Report ───────────────────────────────────────────────────────
+echo ""
+echo "==> Build complete."
+echo "    libmemba : ${BUILD_DIR}/libmemba.so"
+echo "    memba-cli: ${BUILD_DIR}/memba-cli"
+echo ""
+echo "Python install (editable, uses libmemba.so from build/):"
+echo "    pip install -e ."