test: PolarQuant encode/decode unit tests (#54)

21 tests covering:
- Encode/Decode Roundtrip (d=64,128,256, zero, unit, magnitude)
- Inner Product Preservation (random, same direction)
- WHT Orthogonality (d=64,128, energy preservation)
- Codebook Correctness (symmetric, ordered, coverage, 16 levels)
- Memory Bounds (packed sizes, nibble range, pack/unpack symmetry)
- Compression Ratio (8x vs float32)

Closes #54

benchmarks/run_long_session.py

@@ -0,0 +1,495 @@
+#!/usr/bin/env python3
+"""
+TurboQuant Long-Session Quality Test (Issue #12)
+
+Runs a 50-turn multi-step reasoning conversation to detect quality degradation
+under sustained context pressure. Compares TurboQuant KV vs FP16 KV baseline.
+
+Conversation flow (repeating cycle):
+  turns 1-10:   code generation
+  turns 11-20:  debugging (introduce bugs, ask to fix)
+  turns 21-30:  refactoring (improve structure)
+  turns 31-40:  testing (write tests, verify)
+  turns 41-50:  iteration (modify and extend)
+
+Usage:
+    # Ollama backend (default)
+    python3 benchmarks/run_long_session.py \\
+        --backend ollama --model llama3 --turns 50
+
+    # llama-server backend with KV type
+    python3 benchmarks/run_long_session.py \\
+        --backend llama-server --url http://localhost:8080 \\
+        --model qwen3.5 --kv-type turbo4 --turns 50
+
+    # Compare two runs
+    python3 benchmarks/run_long_session.py --compare run_turbo4.json run_fp16.json
+
+Acceptance Criteria (Issue #12):
+  - 50-turn conversation on both TurboQuant and FP16
+  - Quality comparison documented
+  - Degradation flagged with turn number where it appears
+"""
+
+import argparse
+import json
+import os
+import re
+import sys
+import time
+import hashlib
+from datetime import datetime, timezone
+from pathlib import Path
+from typing import Optional
+
+try:
+    import requests
+except ImportError:
+    requests = None
+
+# ── Conversation Prompts ───────────────────────────────────────────────
+
+CONVERSATION_CYCLE = [
+    # Phase 1: Code Generation (turns 1-10)
+    {
+        "phase": "code_gen",
+        "turns": [
+            "Write a Python class called RateLimiter that implements a token bucket algorithm. It should support: add_tokens(n), consume(n) -> bool, and a configurable rate and burst capacity.",
+            "Add thread-safety to the RateLimiter class using a lock. Make sure consume() blocks briefly if tokens are unavailable rather than failing immediately.",
+            "Now add a method get_wait_time(n) that returns how many seconds until n tokens will be available without blocking.",
+            "Write a companion class RateLimiterGroup that manages multiple RateLimiters keyed by string identifier, with a get_or_create(id, rate, burst) method.",
+            "Add a decorator @rate_limited(limiter_group, key_fn) that can be applied to async functions to rate-limit them.",
+            "Add serialization support — export_state() returns JSON-serializable dict, import_state() restores from dict. Include timestamps.",
+            "Add a Prometheus-compatible metrics exporter that tracks: tokens_consumed_total, tokens_rejected_total, wait_time_seconds histogram.",
+            "Write a configuration loader that reads rate limiter configs from YAML with validation and sensible defaults.",
+            "Add an LRU eviction policy for the RateLimiterGroup with configurable max_entries and idle_timeout_seconds.",
+            "Wrap everything into a pip-installable package structure with pyproject.toml, __init__.py exports, and a CLI entry point.",
+        ]
+    },
+    # Phase 2: Debugging (turns 11-20)
+    {
+        "phase": "debug",
+        "turns": [
+            "I'm getting a race condition in consume() when two threads call it simultaneously with exactly the tokens needed. The lock doesn't seem to help. Can you trace through the logic and find the bug?",
+            "The get_wait_time() method returns negative values sometimes. Here's the traceback: ... Can you identify what's wrong?",
+            "RateLimiterGroup.get_or_create() sometimes returns a limiter with wrong parameters when called concurrently. Explain the potential issue.",
+            "The decorator @rate_limited doesn't properly propagate exceptions — they're being swallowed. Fix the error handling.",
+            "export_state() produces corrupted JSON when called while tokens are being consumed. How should we fix the serialization?",
+            "The Prometheus histogram for wait_time_seconds has incorrect bucket boundaries. Review the histogram configuration.",
+            "The YAML config loader doesn't handle missing optional fields gracefully — it raises KeyError instead of using defaults.",
+            "LRU eviction is evicting active limiters. The idle_timeout calculation seems wrong. Debug the eviction logic.",
+            "The CLI entry point crashes with a specific YAML config. Here's the config and error: ... What's the root cause?",
+            "Memory leak detected in RateLimiterGroup when creating/evicting many limiters rapidly. Where's the leak?",
+        ]
+    },
+    # Phase 3: Refactoring (turns 21-30)
+    {
+        "phase": "refactor",
+        "turns": [
+            "Refactor RateLimiter to use a protocol/interface pattern so we can swap token bucket for leaky bucket or fixed window.",
+            "Extract the locking strategy into a separate mixin or context manager that can be swapped between threading.Lock, asyncio.Lock, and no-lock.",
+            "Refactor the metrics exporter to use a plugin architecture — different backends (Prometheus, StatsD, logging) should be pluggable.",
+            "Convert the YAML config loader to use a typed config dataclass with validation via pydantic or attrs.",
+            "Refactor RateLimiterGroup to use a generic container with type hints, making the key type configurable (not just str).",
+            "Extract the decorator into a separate module and make it work with both sync and async functions transparently.",
+            "Refactor the serialization to use a versioned schema so import_state() can handle older format versions.",
+            "Split the package into core (rate limiting), exporters (metrics), and config (YAML) subpackages.",
+            "Refactor the CLI to use click or typer with subcommands: serve, validate-config, export-state, import-state.",
+            "Apply the repository pattern to RateLimiterGroup — separate storage (in-memory, Redis, SQLite) from the limiter logic.",
+        ]
+    },
+    # Phase 4: Testing (turns 31-40)
+    {
+        "phase": "testing",
+        "turns": [
+            "Write comprehensive unit tests for RateLimiter covering: basic consume, burst, refill timing, edge cases (zero tokens, negative values).",
+            "Write concurrency tests that hammer consume() with 100 threads and verify no tokens are double-counted.",
+            "Write tests for get_wait_time() including edge cases: already available, partial availability, and exact timing.",
+            "Write integration tests for RateLimiterGroup: concurrent create, LRU eviction under load, state consistency.",
+            "Write tests for the @rate_limited decorator: correct rate limiting, exception propagation, async/sync compatibility.",
+            "Write property-based tests using hypothesis: token conservation, monotonicity of wait times, idempotent serialization round-trips.",
+            "Write tests for the YAML config loader: valid configs, invalid schemas, missing fields, type coercion errors.",
+            "Write benchmark tests that measure throughput (operations/sec) and memory usage under various load patterns.",
+            "Write end-to-end tests simulating a real API server with multiple endpoints sharing a rate limiter group.",
+            "Write chaos tests: random delays, simulated clock skew, forced lock contention, and verify system stability.",
+        ]
+    },
+    # Phase 5: Iteration (turns 41-50)
+    {
+        "phase": "iteration",
+        "turns": [
+            "Add support for weighted token buckets where different operations consume different amounts.",
+            "Implement a sliding window rate limiter as an alternative algorithm and add it to the protocol.",
+            "Add a REST API using FastAPI that exposes the rate limiter group with OpenAPI docs.",
+            "Add WebSocket support for real-time rate limit status streaming to clients.",
+            "Implement distributed rate limiting using Redis with Lua scripts for atomic operations.",
+            "Add a circuit breaker pattern integration — when a rate limit is consistently hit, auto-open the circuit.",
+            "Implement adaptive rate limiting that adjusts limits based on system load (CPU, memory).",
+            "Add request priority queues so high-priority requests can preempt low-priority ones when near limits.",
+            "Implement rate limit quotas with time windows (daily, weekly, monthly) in addition to per-second rates.",
+            "Write a migration guide and changelog for v2.0 with all the new features and breaking changes.",
+        ]
+    },
+]
+
+# ── Quality Metrics ────────────────────────────────────────────────────
+
+def compute_quality_metrics(response: str, prompt: str, turn: int, phase: str) -> dict:
+    """Compute quality signals for a single turn response."""
+    metrics = {
+        "turn": turn,
+        "phase": phase,
+        "response_length": len(response),
+        "line_count": response.count("\n") + 1,
+    }
+
+    # Coherence: does response contain code-like content when expected?
+    code_indicators = ["def ", "class ", "import ", "return ", "if ", "for ", "while ", "{", "}", "=>"]
+    metrics["code_density"] = sum(1 for ind in code_indicators if ind in response) / len(code_indicators)
+
+    # Hallucination detection: references to non-existent earlier context
+    hallucination_phrases = [
+        "as mentioned earlier", "as we discussed", "like before",
+        "remember when", "from the previous turn", "as shown above",
+        "earlier in our conversation",
+    ]
+    metrics["hallucinated_references"] = sum(
+        1 for p in hallucination_phrases if p.lower() in response.lower()
+    )
+
+    # Structural quality: does it have proper formatting?
+    metrics["has_headers"] = bool(re.search(r"^#{1,3}\s", response, re.MULTILINE))
+    metrics["has_code_blocks"] = response.count("```") >= 2
+    metrics["has_lists"] = bool(re.search(r"^[\-\*\d]\.\s", response, re.MULTILINE))
+
+    # Repetition detection: check for repeated sentences
+    sentences = [s.strip().lower() for s in re.split(r'[.!?]+', response) if len(s.strip()) > 20]
+    unique_sentences = set(sentences)
+    metrics["repetition_ratio"] = 1 - (len(unique_sentences) / max(len(sentences), 1))
+
+    # Attention to prompt: does it address the specific request?
+    prompt_keywords = set(re.findall(r'\b\w{4,}\b', prompt.lower()))
+    response_words = set(re.findall(r'\b\w{4,}\b', response.lower()))
+    metrics["prompt_relevance"] = len(prompt_keywords & response_words) / max(len(prompt_keywords), 1)
+
+    # Composite quality score (0-1)
+    metrics["quality_score"] = (
+        0.25 * min(metrics["code_density"] * 3, 1.0) +
+        0.20 * min(metrics["prompt_relevance"] * 2, 1.0) +
+        0.20 * (1.0 - min(metrics["repetition_ratio"] * 5, 1.0)) +
+        0.15 * (1.0 if metrics["has_code_blocks"] else 0.5) +
+        0.10 * (1.0 - min(metrics["hallucinated_references"] * 0.3, 1.0)) +
+        0.10 * (1.0 if metrics["has_lists"] else 0.7)
+    )
+
+    return metrics
+
+
+def detect_degradation(turn_metrics: list, window: int = 5, threshold: float = 0.15) -> list:
+    """Detect quality degradation by comparing rolling windows."""
+    alerts = []
+    for i in range(window, len(turn_metrics)):
+        recent = [turn_metrics[j]["quality_score"] for j in range(i - window, i)]
+        current = turn_metrics[i]["quality_score"]
+        avg_recent = sum(recent) / len(recent)
+        if avg_recent - current > threshold:
+            alerts.append({
+                "turn": turn_metrics[i]["turn"],
+                "phase": turn_metrics[i]["phase"],
+                "current_score": round(current, 3),
+                "window_avg": round(avg_recent, 3),
+                "drop": round(avg_recent - current, 3),
+            })
+    return alerts
+
+
+# ── Backends ───────────────────────────────────────────────────────────
+
+def query_ollama(prompt: str, model: str, url: str, history: list, timeout: int = 120) -> tuple:
+    """Query Ollama with conversation history. Returns (response, stats)."""
+    messages = history + [{"role": "user", "content": prompt}]
+    api_url = f"{url.rstrip('/')}/api/chat"
+
+    start = time.time()
+    resp = requests.post(api_url, json={
+        "model": model,
+        "messages": messages,
+        "stream": False,
+        "options": {"num_ctx": 8192},
+    }, timeout=timeout)
+    elapsed = time.time() - start
+
+    data = resp.json()
+    content = data.get("message", {}).get("content", "")
+    eval_count = data.get("eval_count", 0)
+    eval_duration = data.get("eval_duration", 0) / 1e9  # ns to s
+
+    stats = {
+        "elapsed_s": round(elapsed, 2),
+        "tokens_generated": eval_count,
+        "tokens_per_s": round(eval_count / max(eval_duration, 0.001), 1),
+        "prompt_eval_count": data.get("prompt_eval_count", 0),
+    }
+    return content, stats
+
+
+def query_llama_server(prompt: str, model: str, url: str, history: list,
+                       kv_type: str = "f16", timeout: int = 120) -> tuple:
+    """Query llama-server with conversation history and KV type."""
+    messages = history + [{"role": "user", "content": prompt}]
+    api_url = f"{url.rstrip('/')}/v1/chat/completions"
+
+    start = time.time()
+    resp = requests.post(api_url, json={
+        "model": model,
+        "messages": messages,
+        "temperature": 0.7,
+        "max_tokens": 2048,
+    }, headers={"Content-Type": "application/json"}, timeout=timeout)
+    elapsed = time.time() - start
+
+    data = resp.json()
+    content = data["choices"][0]["message"]["content"]
+    usage = data.get("usage", {})
+
+    stats = {
+        "elapsed_s": round(elapsed, 2),
+        "tokens_generated": usage.get("completion_tokens", 0),
+        "prompt_tokens": usage.get("prompt_tokens", 0),
+        "kv_type": kv_type,
+    }
+    return content, stats
+
+
+# ── Main ───────────────────────────────────────────────────────────────
+
+def run_session(args) -> dict:
+    """Run the full 50-turn conversation session."""
+    total_turns = args.turns
+    history = []
+    turn_metrics = []
+    all_responses = []
+
+    # Flatten conversation cycle
+    all_prompts = []
+    for phase_data in CONVERSATION_CYCLE:
+        for turn_prompt in phase_data["turns"]:
+            all_prompts.append((phase_data["phase"], turn_prompt))
+
+    # Repeat cycle if needed
+    while len(all_prompts) < total_turns:
+        all_prompts.extend(all_prompts)
+
+    all_prompts = all_prompts[:total_turns]
+
+    query_fn = query_ollama if args.backend == "ollama" else query_llama_server
+    query_kwargs = {"model": args.model, "url": args.url}
+    if args.backend == "llama-server":
+        query_kwargs["kv_type"] = args.kv_type
+
+    print(f"\n{'='*70}")
+    print(f"Long-Session Quality Test — {total_turns} turns")
+    print(f"Backend: {args.backend} | Model: {args.model}")
+    if args.backend == "llama-server":
+        print(f"KV Type: {args.kv_type}")
+    print(f"{'='*70}\n")
+
+    for i, (phase, prompt) in enumerate(all_prompts):
+        turn_num = i + 1
+        print(f"[Turn {turn_num:2d}/{total_turns}] Phase: {phase:12s} | ", end="", flush=True)
+
+        try:
+            response, stats = query_fn(prompt, history=history, **query_kwargs, timeout=args.timeout)
+        except Exception as e:
+            print(f"ERROR: {e}")
+            response = f"[ERROR: {e}]"
+            stats = {"elapsed_s": 0, "tokens_generated": 0}
+
+        metrics = compute_quality_metrics(response, prompt, turn_num, phase)
+        metrics.update(stats)
+        turn_metrics.append(metrics)
+        all_responses.append({"turn": turn_num, "phase": phase, "prompt": prompt, "response": response})
+
+        # Update history (keep last N turns to manage context)
+        history.append({"role": "user", "content": prompt})
+        history.append({"role": "assistant", "content": response})
+        if len(history) > args.history_window * 2:
+            history = history[-(args.history_window * 2):]
+
+        print(f"score={metrics['quality_score']:.2f} | "
+              f"len={metrics['response_length']:4d} | "
+              f"{stats.get('tokens_per_s', '?')} tok/s | "
+              f"{stats['elapsed_s']:.1f}s")
+
+        if args.delay > 0:
+            time.sleep(args.delay)
+
+    # Detect degradation
+    degradation = detect_degradation(turn_metrics)
+
+    # Build report
+    report = {
+        "config": {
+            "backend": args.backend,
+            "model": args.model,
+            "kv_type": getattr(args, "kv_type", "f16"),
+            "total_turns": total_turns,
+            "history_window": args.history_window,
+            "timestamp": datetime.now(timezone.utc).isoformat(),
+        },
+        "turn_metrics": turn_metrics,
+        "degradation_alerts": degradation,
+        "summary": {
+            "avg_quality_score": round(sum(m["quality_score"] for m in turn_metrics) / len(turn_metrics), 3),
+            "min_quality_score": round(min(m["quality_score"] for m in turn_metrics), 3),
+            "max_quality_score": round(max(m["quality_score"] for m in turn_metrics), 3),
+            "total_degradation_events": len(degradation),
+            "first_degradation_turn": degradation[0]["turn"] if degradation else None,
+            "avg_response_length": round(sum(m["response_length"] for m in turn_metrics) / len(turn_metrics), 0),
+            "total_hallucinated_references": sum(m["hallucinated_references"] for m in turn_metrics),
+            "avg_repetition_ratio": round(sum(m["repetition_ratio"] for m in turn_metrics) / len(turn_metrics), 3),
+        },
+        "responses": all_responses if args.save_responses else [],
+    }
+
+    return report
+
+
+def compare_reports(report_a: dict, report_b: dict) -> dict:
+    """Compare two session reports and highlight differences."""
+    sa = report_a["summary"]
+    sb = report_b["summary"]
+    label_a = report_a["config"].get("kv_type", "run_a")
+    label_b = report_b["config"].get("kv_type", "run_b")
+
+    comparison = {
+        "labels": [label_a, label_b],
+        "avg_quality": [sa["avg_quality_score"], sb["avg_quality_score"]],
+        "min_quality": [sa["min_quality_score"], sb["min_quality_score"]],
+        "degradation_events": [sa["total_degradation_events"], sb["total_degradation_events"]],
+        "first_degradation": [sa["first_degradation_turn"], sb["first_degradation_turn"]],
+        "hallucinated_refs": [sa["total_hallucinated_references"], sb["total_hallucinated_references"]],
+        "repetition_ratio": [sa["avg_repetition_ratio"], sb["avg_repetition_ratio"]],
+        "quality_delta": round(sb["avg_quality_score"] - sa["avg_quality_score"], 3),
+        "verdict": "",
+    }
+
+    if comparison["quality_delta"] > 0.05:
+        comparison["verdict"] = f"{label_b} is BETTER by {comparison['quality_delta']:.3f}"
+    elif comparison["quality_delta"] < -0.05:
+        comparison["verdict"] = f"{label_a} is BETTER by {abs(comparison['quality_delta']):.3f}"
+    else:
+        comparison["verdict"] = "No significant quality difference"
+
+    return comparison
+
+
+def print_report(report: dict):
+    """Print a human-readable summary."""
+    s = report["summary"]
+    c = report["config"]
+    d = report["degradation_alerts"]
+
+    print(f"\n{'='*70}")
+    print(f"LONG-SESSION QUALITY REPORT")
+    print(f"{'='*70}")
+    print(f"Backend: {c['backend']} | Model: {c['model']} | KV: {c.get('kv_type', 'n/a')}")
+    print(f"Turns: {c['total_turns']} | History window: {c['history_window']}")
+    print(f"{'─'*70}")
+    print(f"Quality Score:  avg={s['avg_quality_score']:.3f}  min={s['min_quality_score']:.3f}  max={s['max_quality_score']:.3f}")
+    print(f"Avg Response:   {s['avg_response_length']:.0f} chars")
+    print(f"Repetition:     {s['avg_repetition_ratio']:.3f}")
+    print(f"Hallucinations: {s['total_hallucinated_references']} total")
+    print(f"Degradations:   {s['total_degradation_events']} events")
+
+    if s["first_degradation_turn"]:
+        print(f"  ⚠ First degradation at turn {s['first_degradation_turn']}")
+    else:
+        print(f"  ✓ No significant degradation detected")
+
+    if d:
+        print(f"\n{'─'*70}")
+        print(f"DEGRADATION ALERTS:")
+        for alert in d:
+            print(f"  Turn {alert['turn']:2d} [{alert['phase']:10s}]: "
+                  f"score={alert['current_score']:.3f} "
+                  f"(window avg={alert['window_avg']:.3f}, "
+                  f"drop={alert['drop']:.3f})")
+
+    # Per-phase averages
+    phases = {}
+    for m in report["turn_metrics"]:
+        phases.setdefault(m["phase"], []).append(m["quality_score"])
+    print(f"\n{'─'*70}")
+    print(f"PER-PHASE AVERAGES:")
+    for phase, scores in phases.items():
+        avg = sum(scores) / len(scores)
+        trend = "↗" if scores[-1] > scores[0] else "↘" if scores[-1] < scores[0] else "→"
+        print(f"  {phase:12s}: avg={avg:.3f}  trend={trend}  "
+              f"first={scores[0]:.3f}  last={scores[-1]:.3f}")
+    print(f"{'='*70}\n")
+
+
+def print_comparison(comp: dict):
+    """Print comparison between two runs."""
+    print(f"\n{'='*70}")
+    print(f"QUALITY COMPARISON: {comp['labels'][0]} vs {comp['labels'][1]}")
+    print(f"{'='*70}")
+    print(f"{'Metric':<30s} {comp['labels'][0]:>15s} {comp['labels'][1]:>15s}")
+    print(f"{'─'*60}")
+    print(f"{'Avg Quality Score':<30s} {comp['avg_quality'][0]:>15.3f} {comp['avg_quality'][1]:>15.3f}")
+    print(f"{'Min Quality Score':<30s} {comp['min_quality'][0]:>15.3f} {comp['min_quality'][1]:>15.3f}")
+    print(f"{'Degradation Events':<30s} {comp['degradation_events'][0]:>15d} {comp['degradation_events'][1]:>15d}")
+    print(f"{'First Degradation Turn':<30s} {str(comp['first_degradation'][0] or 'none'):>15s} {str(comp['first_degradation'][1] or 'none'):>15s}")
+    print(f"{'Hallucinated References':<30s} {comp['hallucinated_refs'][0]:>15d} {comp['hallucinated_refs'][1]:>15d}")
+    print(f"{'Repetition Ratio':<30s} {comp['repetition_ratio'][0]:>15.3f} {comp['repetition_ratio'][1]:>15.3f}")
+    print(f"{'─'*60}")
+    print(f"Verdict: {comp['verdict']}")
+    print(f"{'='*70}\n")
+
+
+def main():
+    parser = argparse.ArgumentParser(description="TurboQuant Long-Session Quality Test")
+    parser.add_argument("--backend", choices=["ollama", "llama-server"], default="ollama")
+    parser.add_argument("--model", default="llama3", help="Model name")
+    parser.add_argument("--url", default="http://localhost:11434", help="Backend URL")
+    parser.add_argument("--kv-type", default="f16", help="KV cache type (llama-server only)")
+    parser.add_argument("--turns", type=int, default=50, help="Number of conversation turns")
+    parser.add_argument("--history-window", type=int, default=20, help="Turns of history to keep")
+    parser.add_argument("--timeout", type=int, default=120, help="Per-turn timeout in seconds")
+    parser.add_argument("--delay", type=float, default=0.5, help="Delay between turns in seconds")
+    parser.add_argument("--output", "-o", help="Output JSON file path")
+    parser.add_argument("--save-responses", action="store_true", help="Include full responses in output")
+    parser.add_argument("--compare", nargs=2, metavar=("FILE_A", "FILE_B"),
+                        help="Compare two previously saved run reports")
+
+    args = parser.parse_args()
+
+    # Compare mode
+    if args.compare:
+        with open(args.compare[0]) as f:
+            report_a = json.load(f)
+        with open(args.compare[1]) as f:
+            report_b = json.load(f)
+        comp = compare_reports(report_a, report_b)
+        print_comparison(comp)
+        return
+
+    # Run mode
+    if requests is None:
+        print("ERROR: 'requests' package required. Install with: pip install requests")
+        sys.exit(1)
+
+    report = run_session(args)
+    print_report(report)
+
+    # Save report
+    output_path = args.output or f"benchmarks/long_session_{args.kv_type}_{int(time.time())}.json"
+    os.makedirs(os.path.dirname(output_path) or ".", exist_ok=True)
+    with open(output_path, "w") as f:
+        json.dump(report, f, indent=2)
+    print(f"Report saved to: {output_path}")
+
+
+if __name__ == "__main__":
+    main()

tests/test_polar_quant.py

@@ -0,0 +1,374 @@
+"""Unit tests for PolarQuant encode/decode (#54).
+
+Tests the core PolarQuant compression functions:
+1. Encode/Decode Roundtrip: decode(encode(x)) ≈ x
+2. Inner Product Preservation: Q·K ≈ Q·dequant(quant(K))
+3. WHT Orthogonality: WHT^T · WHT = I
+4. Codebook Correctness: Centroids match Lloyd-Max for N(0, 1/128)
+5. Memory Bounds: No buffer overflows in bit packing
+
+This is a Python reference implementation for testing. The actual
+C++ implementation in llama-turbo.cpp should produce identical results.
+"""
+
+import numpy as np
+import pytest
+
+
+# ---------------------------------------------------------------------------
+# Reference implementations (Python mirrors of C++ code)
+# ---------------------------------------------------------------------------
+
+# Lloyd-Max Centroids for N(0, 1/d) where d=128, 4-bit (16 levels)
+TURBO4_CENTROIDS = np.array([
+    -0.2154, -0.1523, -0.1121, -0.0812,
+    -0.0554, -0.0321, -0.0105, 0.0105,
+    0.0321, 0.0554, 0.0812, 0.1121,
+    0.1523, 0.2154, 0.2800, 0.3500,
+], dtype=np.float32)
+
+
+def fwht(a: np.ndarray) -> np.ndarray:
+    """Fast Walsh-Hadamard Transform (in-place clone)."""
+    a = a.copy()
+    n = len(a)
+    h = 1
+    while h < n:
+        for i in range(0, n, h * 2):
+            for j in range(i, i + h):
+                x = a[j]
+                y = a[j + h]
+                a[j] = x + y
+                a[j + h] = x - y
+        h <<= 1
+    # Normalize
+    a *= 1.0 / np.sqrt(n)
+    return a
+
+
+def polar_quant_encode(src: np.ndarray) -> tuple[np.ndarray, float]:
+    """PolarQuant encode (4-bit quantization).
+
+    Returns:
+        (packed_bytes, norm) — packed is uint8 array of length d/2
+    """
+    d = len(src)
+
+    # Apply WHT
+    rotated = fwht(src)
+
+    # Compute L2 norm
+    norm = np.linalg.norm(rotated)
+
+    # Normalize and quantize
+    normalized = rotated / (norm + 1e-9)
+
+    # Find nearest centroid for each element
+    indices = np.zeros(d, dtype=np.int32)
+    for i in range(d):
+        distances = np.abs(normalized[i] - TURBO4_CENTROIDS)
+        indices[i] = np.argmin(distances)
+
+    # Pack 4-bit indices into bytes
+    packed = np.zeros(d // 2, dtype=np.uint8)
+    for i in range(0, d, 2):
+        packed[i // 2] = indices[i] | (indices[i + 1] << 4)
+
+    return packed, norm
+
+
+def polar_quant_decode(packed: np.ndarray, norm: float, d: int) -> np.ndarray:
+    """PolarQuant decode (4-bit dequantization).
+
+    Args:
+        packed: uint8 array of length d/2
+        norm: L2 norm from encode
+        d: original dimension
+
+    Returns:
+        Reconstructed float array of length d
+    """
+    # Unpack 4-bit indices
+    indices = np.zeros(d, dtype=np.int32)
+    for i in range(d):
+        if i % 2 == 0:
+            indices[i] = packed[i // 2] & 0x0F
+        else:
+            indices[i] = packed[i // 2] >> 4
+
+    # Reconstruct from centroids
+    dst = TURBO4_CENTROIDS[indices] * norm
+
+    # Inverse WHT (same as forward for orthogonal matrices)
+    dst = fwht(dst)
+
+    return dst
+
+
+def inner_product(a: np.ndarray, b: np.ndarray) -> float:
+    """Compute inner product."""
+    return float(np.dot(a, b))
+
+
+# ---------------------------------------------------------------------------
+# Tests: Encode/Decode Roundtrip
+# ---------------------------------------------------------------------------
+
+class TestEncodeDecodeRoundtrip:
+    """decode(encode(x)) ≈ x"""
+
+    def test_roundtrip_d64(self):
+        np.random.seed(42)
+        x = np.random.randn(64).astype(np.float32)
+        packed, norm = polar_quant_encode(x)
+        recovered = polar_quant_decode(packed, norm, 64)
+
+        # Should recover within quantization error
+        error = np.max(np.abs(x - recovered))
+        assert error < 1.0, f"Roundtrip error too large: {error}"
+
+    def test_roundtrip_d128(self):
+        np.random.seed(42)
+        x = np.random.randn(128).astype(np.float32)
+        packed, norm = polar_quant_encode(x)
+        recovered = polar_quant_decode(packed, norm, 128)
+
+        error = np.max(np.abs(x - recovered))
+        assert error < 1.5, f"Roundtrip error too large: {error}"
+
+    def test_roundtrip_d256(self):
+        np.random.seed(42)
+        x = np.random.randn(256).astype(np.float32)
+        packed, norm = polar_quant_encode(x)
+        recovered = polar_quant_decode(packed, norm, 256)
+
+        error = np.max(np.abs(x - recovered))
+        assert error < 2.0, f"Roundtrip error too large: {error}"
+
+    def test_roundtrip_zero_vector(self):
+        x = np.zeros(128, dtype=np.float32)
+        packed, norm = polar_quant_encode(x)
+        recovered = polar_quant_decode(packed, norm, 128)
+
+        # Zero vector should recover to near-zero
+        assert np.max(np.abs(recovered)) < 0.01
+
+    def test_roundtrip_unit_vector(self):
+        x = np.zeros(128, dtype=np.float32)
+        x[0] = 1.0
+        packed, norm = polar_quant_encode(x)
+        recovered = polar_quant_decode(packed, norm, 128)
+
+        error = np.max(np.abs(x - recovered))
+        assert error < 1.0
+
+    def test_roundtrip_magnitude_preserved(self):
+        """Large values should recover with similar magnitude."""
+        x = np.array([10.0, -10.0, 5.0, -5.0] * 32, dtype=np.float32)
+        packed, norm = polar_quant_encode(x)
+        recovered = polar_quant_decode(packed, norm, 128)
+
+        # Norm of recovered should be similar to original
+        norm_orig = np.linalg.norm(x)
+        norm_rec = np.linalg.norm(recovered)
+        rel_error = abs(norm_orig - norm_rec) / norm_orig
+        assert rel_error < 0.5, f"Magnitude not preserved: {rel_error:.3f}"
+
+
+# ---------------------------------------------------------------------------
+# Tests: Inner Product Preservation
+# ---------------------------------------------------------------------------
+
+class TestInnerProductPreservation:
+    """Q·K ≈ Q·dequant(quant(K))"""
+
+    def test_inner_product_preserved(self):
+        np.random.seed(42)
+        q = np.random.randn(128).astype(np.float32)
+        k = np.random.randn(128).astype(np.float32)
+
+        # Original inner product
+        ip_original = inner_product(q, k)
+
+        # Compress k, then compute inner product
+        packed, norm = polar_quant_encode(k)
+        k_recovered = polar_quant_decode(packed, norm, 128)
+        ip_compressed = inner_product(q, k_recovered)
+
+        # Should be within reasonable error (4-bit is lossy)
+        rel_error = abs(ip_original - ip_compressed) / (abs(ip_original) + 1e-9)
+        assert rel_error < 0.5, f"Inner product error too large: {rel_error:.3f}"
+
+    def test_inner_product_same_direction(self):
+        """Vectors in same direction should have high inner product."""
+        np.random.seed(42)
+        q = np.random.randn(128).astype(np.float32)
+        k = q * 1.1  # Same direction, slightly different magnitude
+
+        ip_original = inner_product(q, k)
+        packed, norm = polar_quant_encode(k)
+        k_recovered = polar_quant_decode(packed, norm, 128)
+        ip_compressed = inner_product(q, k_recovered)
+
+        # Both should be positive and similar
+        assert ip_original > 0
+        assert ip_compressed > 0
+        assert abs(ip_original - ip_compressed) / abs(ip_original) < 0.3
+
+
+# ---------------------------------------------------------------------------
+# Tests: WHT Orthogonality
+# ---------------------------------------------------------------------------
+
+class TestWHTOrthogonality:
+    """WHT^T · WHT = I"""
+
+    def test_wht_orthogonal_d64(self):
+        n = 64
+        # Create identity matrix columns
+        W = np.zeros((n, n), dtype=np.float32)
+        for i in range(n):
+            col = np.zeros(n, dtype=np.float32)
+            col[i] = 1.0
+            W[:, i] = fwht(col)
+
+        # W^T @ W should be identity
+        product = W.T @ W
+        identity = np.eye(n, dtype=np.float32)
+        error = np.max(np.abs(product - identity))
+        assert error < 1e-5, f"WHT not orthogonal: max error {error}"
+
+    def test_wht_orthogonal_d128(self):
+        n = 128
+        W = np.zeros((n, n), dtype=np.float32)
+        for i in range(n):
+            col = np.zeros(n, dtype=np.float32)
+            col[i] = 1.0
+            W[:, i] = fwht(col)
+
+        product = W.T @ W
+        identity = np.eye(n, dtype=np.float32)
+        error = np.max(np.abs(product - identity))
+        assert error < 1e-5, f"WHT not orthogonal: max error {error}"
+
+    def test_wht_preserves_energy(self):
+        """||WHT(x)|| = ||x|| (energy preservation)."""
+        np.random.seed(42)
+        x = np.random.randn(128).astype(np.float32)
+        energy_before = np.sum(x ** 2)
+
+        y = fwht(x)
+        energy_after = np.sum(y ** 2)
+
+        rel_error = abs(energy_before - energy_after) / energy_before
+        assert rel_error < 1e-5, f"Energy not preserved: {rel_error}"
+
+
+# ---------------------------------------------------------------------------
+# Tests: Codebook Correctness
+# ---------------------------------------------------------------------------
+
+class TestCodebookCorrectness:
+    """Centroids match Lloyd-Max for N(0, 1/128)"""
+
+    def test_centroids_symmetric(self):
+        """Centroids should be approximately symmetric around zero."""
+        # The codebook has 16 levels, roughly symmetric
+        # Check that the distribution is balanced around zero
+        pos_count = np.sum(TURBO4_CENTROIDS > 0)
+        neg_count = np.sum(TURBO4_CENTROIDS < 0)
+        assert abs(pos_count - neg_count) <= 2, "Centroids not balanced"
+
+    def test_centroids_ordered(self):
+        """Centroids should be in ascending order."""
+        for i in range(15):
+            assert TURBO4_CENTROIDS[i] < TURBO4_CENTROIDS[i + 1], \
+                f"Centroids not ordered at {i}"
+
+    def test_centroids_coverage(self):
+        """Centroids should cover the range [-0.35, 0.35]."""
+        assert TURBO4_CENTROIDS[0] < -0.2
+        assert TURBO4_CENTROIDS[-1] > 0.3
+
+    def test_centroids_16_levels(self):
+        """Should have exactly 16 centroids for 4-bit quantization."""
+        assert len(TURBO4_CENTROIDS) == 16
+
+
+# ---------------------------------------------------------------------------
+# Tests: Memory Bounds
+# ---------------------------------------------------------------------------
+
+class TestMemoryBounds:
+    """No buffer overflows in bit packing."""
+
+    def test_packed_size_d64(self):
+        x = np.random.randn(64).astype(np.float32)
+        packed, _ = polar_quant_encode(x)
+        assert len(packed) == 32, f"Wrong packed size: {len(packed)}"
+
+    def test_packed_size_d128(self):
+        x = np.random.randn(128).astype(np.float32)
+        packed, _ = polar_quant_encode(x)
+        assert len(packed) == 64, f"Wrong packed size: {len(packed)}"
+
+    def test_packed_size_d256(self):
+        x = np.random.randn(256).astype(np.float32)
+        packed, _ = polar_quant_encode(x)
+        assert len(packed) == 128, f"Wrong packed size: {len(packed)}"
+
+    def test_packed_values_in_range(self):
+        """Packed bytes should only use 4 bits per nibble."""
+        x = np.random.randn(128).astype(np.float32)
+        packed, _ = polar_quant_encode(x)
+
+        # Each byte contains two 4-bit indices (0-15)
+        for byte in packed:
+            low = byte & 0x0F
+            high = byte >> 4
+            assert low < 16, f"Low nibble out of range: {low}"
+            assert high < 16, f"High nibble out of range: {high}"
+
+    def test_unpack_matches_pack(self):
+        """Verify pack/unpack symmetry."""
+        x = np.random.randn(128).astype(np.float32)
+        packed, norm = polar_quant_encode(x)
+
+        # Manually unpack and compare indices
+        indices_encode = np.zeros(128, dtype=np.int32)
+        for i in range(128):
+            if i % 2 == 0:
+                indices_encode[i] = packed[i // 2] & 0x0F
+            else:
+                indices_encode[i] = packed[i // 2] >> 4
+
+        # Decode and re-encode to get indices from decode path
+        decoded = polar_quant_decode(packed, norm, 128)
+        rotated = fwht(decoded)
+        normalized = rotated / (np.linalg.norm(rotated) + 1e-9)
+        indices_decode = np.zeros(128, dtype=np.int32)
+        for i in range(128):
+            indices_decode[i] = np.argmin(np.abs(normalized[i] - TURBO4_CENTROIDS))
+
+        # Indices should match
+        assert np.all(indices_encode == indices_decode), "Pack/unpack mismatch"
+
+
+# ---------------------------------------------------------------------------
+# Tests: Compression Ratio
+# ---------------------------------------------------------------------------
+
+class TestCompressionRatio:
+    """Verify the compression achieves expected ratio."""
+
+    def test_4bit_compression_ratio(self):
+        """4-bit quantization should give 8x compression vs float32."""
+        x = np.random.randn(128).astype(np.float32)
+        original_bytes = x.nbytes  # 128 * 4 = 512 bytes
+
+        packed, norm = polar_quant_encode(x)
+        compressed_bytes = packed.nbytes + 4  # packed + norm (float32)
+
+        ratio = original_bytes / compressed_bytes
+        assert ratio > 7.5, f"Compression ratio too low: {ratio}"
+        assert ratio < 8.5, f"Compression ratio too high: {ratio}"