superlocalmemory 2.8.5 → 3.0.0

package/src/learning/tests/test_signal_inference.py

@@ -1,397 +0,0 @@
-#!/usr/bin/env python3
-# SPDX-License-Identifier: MIT
-# Copyright (c) 2026 SuperLocalMemory (superlocalmemory.com)
-"""SuperLocalMemory V2 - Signal Inference Engine Tests (v2.7.4)
-
-Tests for the implicit feedback signal inference system.
-"""
-import time
-import threading
-import pytest
-import sys
-from pathlib import Path
-
-# Ensure mcp_server module is importable for RecallBuffer
-sys.path.insert(0, str(Path(__file__).parent.parent.parent.parent))
-
-
-class TestRecallBuffer:
-    """Test the _RecallBuffer class from mcp_server."""
-
-    def _make_buffer(self):
-        """Create a fresh RecallBuffer for testing."""
-        # Import directly to avoid starting MCP server
-        from importlib import import_module
-        import types
-
-        # Create a minimal RecallBuffer instance by duplicating the class logic
-        # without importing the full mcp_server (which requires MCP deps)
-        sys.path.insert(0, str(Path(__file__).parent.parent.parent))
-
-        class RecallBuffer:
-            def __init__(self):
-                self._lock = threading.Lock()
-                self._last_recall = {}
-                self._global_last = None
-                self._signal_timestamps = {}
-                self._recent_result_ids = set()
-                self._recall_count = 0
-                self._positive_threshold = 300.0
-                self._inter_recall_times = []
-
-            def record_recall(self, query, result_ids, agent_id="mcp-client"):
-                now = time.time()
-                signals = []
-                with self._lock:
-                    self._recall_count += 1
-                    result_id_set = set(result_ids)
-                    self._recent_result_ids = result_id_set
-                    current = {
-                        "query": query, "result_ids": result_ids,
-                        "result_id_set": result_id_set, "timestamp": now,
-                        "agent_id": agent_id,
-                    }
-                    prev = self._last_recall.get(agent_id)
-                    if prev:
-                        time_gap = now - prev["timestamp"]
-                        self._inter_recall_times.append(time_gap)
-                        if len(self._inter_recall_times) > 100:
-                            self._inter_recall_times = self._inter_recall_times[-100:]
-                        if len(self._inter_recall_times) >= 10:
-                            sorted_times = sorted(self._inter_recall_times)
-                            median = sorted_times[len(sorted_times) // 2]
-                            self._positive_threshold = max(60.0, min(median * 0.8, 1800.0))
-                        if time_gap < 30.0 and query != prev["query"]:
-                            for mid in prev["result_ids"][:5]:
-                                signals.append({
-                                    "memory_id": mid,
-                                    "signal_type": "implicit_negative_requick",
-                                    "query": prev["query"],
-                                    "rank_position": prev["result_ids"].index(mid) + 1,
-                                })
-                        elif time_gap > self._positive_threshold:
-                            for mid in prev["result_ids"][:3]:
-                                signals.append({
-                                    "memory_id": mid,
-                                    "signal_type": "implicit_positive_timegap",
-                                    "query": prev["query"],
-                                    "rank_position": prev["result_ids"].index(mid) + 1,
-                                })
-                        overlap = result_id_set & prev["result_id_set"]
-                        for mid in overlap:
-                            signals.append({
-                                "memory_id": mid,
-                                "signal_type": "implicit_positive_reaccess",
-                                "query": query,
-                            })
-                    global_prev = self._global_last
-                    if global_prev and global_prev["agent_id"] != agent_id:
-                        cross_overlap = result_id_set & global_prev["result_id_set"]
-                        for mid in cross_overlap:
-                            signals.append({
-                                "memory_id": mid,
-                                "signal_type": "implicit_positive_cross_tool",
-                                "query": query,
-                            })
-                    self._last_recall[agent_id] = current
-                    self._global_last = current
-                return signals
-
-            def check_post_action(self, memory_id, action):
-                with self._lock:
-                    if memory_id not in self._recent_result_ids:
-                        return None
-                    if action == "update":
-                        return {"memory_id": memory_id, "signal_type": "implicit_positive_post_update",
-                                "query": self._global_last["query"] if self._global_last else ""}
-                    elif action == "delete":
-                        return {"memory_id": memory_id, "signal_type": "implicit_negative_post_delete",
-                                "query": self._global_last["query"] if self._global_last else ""}
-                return None
-
-            def check_rate_limit(self, agent_id, max_per_minute=5):
-                now = time.time()
-                with self._lock:
-                    if agent_id not in self._signal_timestamps:
-                        self._signal_timestamps[agent_id] = []
-                    self._signal_timestamps[agent_id] = [
-                        ts for ts in self._signal_timestamps[agent_id] if now - ts < 60.0
-                    ]
-                    if len(self._signal_timestamps[agent_id]) >= max_per_minute:
-                        return False
-                    self._signal_timestamps[agent_id].append(now)
-                    return True
-
-            def get_recall_count(self):
-                with self._lock:
-                    return self._recall_count
-
-            def get_stats(self):
-                with self._lock:
-                    return {
-                        "recall_count": self._recall_count,
-                        "tracked_agents": len(self._last_recall),
-                        "positive_threshold_s": round(self._positive_threshold, 1),
-                        "recent_results_count": len(self._recent_result_ids),
-                    }
-
-        return RecallBuffer()
-
-    def test_first_recall_no_signals(self):
-        """First recall should produce no signals (nothing to compare with)."""
-        buf = self._make_buffer()
-        signals = buf.record_recall("test query", [1, 2, 3])
-        assert signals == []
-        assert buf.get_recall_count() == 1
-
-    def test_quick_requery_negative_signals(self):
-        """Quick re-query with different query should produce negative signals."""
-        buf = self._make_buffer()
-        buf.record_recall("first query", [10, 20, 30])
-        # Simulate quick re-query (no sleep needed — time gap is ~0)
-        signals = buf.record_recall("different query", [40, 50])
-        negative_signals = [s for s in signals if s["signal_type"] == "implicit_negative_requick"]
-        assert len(negative_signals) > 0
-        # Should target previous results (10, 20, 30)
-        neg_ids = {s["memory_id"] for s in negative_signals}
-        assert neg_ids.issubset({10, 20, 30})
-
-    def test_same_query_no_negative(self):
-        """Same exact query repeated quickly should NOT produce negative signals."""
-        buf = self._make_buffer()
-        buf.record_recall("same query", [10, 20])
-        signals = buf.record_recall("same query", [10, 20])
-        negative_signals = [s for s in signals if s["signal_type"] == "implicit_negative_requick"]
-        assert len(negative_signals) == 0
-
-    def test_reaccess_positive_signals(self):
-        """Same memory appearing in consecutive recalls = positive."""
-        buf = self._make_buffer()
-        buf.record_recall("query a", [10, 20, 30])
-        signals = buf.record_recall("query b", [20, 40, 50])
-        reaccess = [s for s in signals if s["signal_type"] == "implicit_positive_reaccess"]
-        assert len(reaccess) == 1
-        assert reaccess[0]["memory_id"] == 20
-
-    def test_cross_tool_positive_signals(self):
-        """Same memory recalled by different agents = cross-tool positive."""
-        buf = self._make_buffer()
-        buf.record_recall("query", [10, 20], agent_id="claude")
-        signals = buf.record_recall("query", [20, 30], agent_id="cursor")
-        cross_tool = [s for s in signals if s["signal_type"] == "implicit_positive_cross_tool"]
-        assert len(cross_tool) == 1
-        assert cross_tool[0]["memory_id"] == 20
-
-    def test_post_action_update_tracked(self):
-        """Update after recall should generate positive signal."""
-        buf = self._make_buffer()
-        buf.record_recall("query", [10, 20, 30])
-        signal = buf.check_post_action(20, "update")
-        assert signal is not None
-        assert signal["signal_type"] == "implicit_positive_post_update"
-        assert signal["memory_id"] == 20
-
-    def test_post_action_delete_tracked(self):
-        """Delete after recall should generate negative signal."""
-        buf = self._make_buffer()
-        buf.record_recall("query", [10, 20, 30])
-        signal = buf.check_post_action(10, "delete")
-        assert signal is not None
-        assert signal["signal_type"] == "implicit_negative_post_delete"
-
-    def test_post_action_unrelated_memory_ignored(self):
-        """Action on memory NOT in recent results should be ignored."""
-        buf = self._make_buffer()
-        buf.record_recall("query", [10, 20, 30])
-        signal = buf.check_post_action(999, "update")
-        assert signal is None
-
-    def test_rate_limiting(self):
-        """Rate limiter should cap signals per agent per minute."""
-        buf = self._make_buffer()
-        agent = "test-agent"
-        # First 5 should be allowed
-        for i in range(5):
-            assert buf.check_rate_limit(agent) is True
-        # 6th should be blocked
-        assert buf.check_rate_limit(agent) is False
-
-    def test_rate_limiting_different_agents(self):
-        """Different agents should have independent rate limits."""
-        buf = self._make_buffer()
-        for i in range(5):
-            assert buf.check_rate_limit("agent1") is True
-        assert buf.check_rate_limit("agent1") is False
-        # Agent2 should still be allowed
-        assert buf.check_rate_limit("agent2") is True
-
-    def test_recall_count_increments(self):
-        """Recall count should increment with each recall."""
-        buf = self._make_buffer()
-        assert buf.get_recall_count() == 0
-        buf.record_recall("q1", [1])
-        assert buf.get_recall_count() == 1
-        buf.record_recall("q2", [2])
-        assert buf.get_recall_count() == 2
-
-    def test_negative_signals_cap_at_5(self):
-        """Negative signals should only target top 5 results."""
-        buf = self._make_buffer()
-        buf.record_recall("first", list(range(1, 11)))  # 10 results
-        signals = buf.record_recall("different", [99])
-        negative = [s for s in signals if s["signal_type"] == "implicit_negative_requick"]
-        assert len(negative) <= 5
-
-    def test_stats_output(self):
-        """Stats should report correct values."""
-        buf = self._make_buffer()
-        buf.record_recall("q", [1, 2, 3])
-        stats = buf.get_stats()
-        assert stats["recall_count"] == 1
-        assert stats["tracked_agents"] == 1
-        assert stats["recent_results_count"] == 3
-        assert stats["positive_threshold_s"] == 300.0
-
-    def test_thread_safety(self):
-        """Buffer should handle concurrent access without errors."""
-        buf = self._make_buffer()
-        errors = []
-
-        def worker(agent_id):
-            try:
-                for i in range(20):
-                    buf.record_recall(f"query_{i}", [i, i+1], agent_id=agent_id)
-                    buf.check_rate_limit(agent_id)
-            except Exception as e:
-                errors.append(e)
-
-        threads = [threading.Thread(target=worker, args=(f"agent_{i}",)) for i in range(5)]
-        for t in threads:
-            t.start()
-        for t in threads:
-            t.join(timeout=10)
-
-        assert errors == [], f"Thread safety errors: {errors}"
-        assert buf.get_recall_count() == 100  # 5 agents x 20 recalls
-
-
-class TestFeedbackCollectorImplicit:
-    """Test the new implicit signal methods in FeedbackCollector."""
-
-    def test_record_implicit_signal_valid(self):
-        """Valid implicit signal should be stored."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feedback_collector import FeedbackCollector
-        fc = FeedbackCollector()
-        result = fc.record_implicit_signal(
-            memory_id=42,
-            query="test query",
-            signal_type="implicit_positive_timegap",
-        )
-        # Should return row ID or None (depends on DB availability)
-        # Just verify it doesn't crash
-        assert result is None or isinstance(result, int)
-
-    def test_record_implicit_signal_invalid_type(self):
-        """Invalid signal type should return None."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feedback_collector import FeedbackCollector
-        fc = FeedbackCollector()
-        result = fc.record_implicit_signal(
-            memory_id=42,
-            query="test",
-            signal_type="totally_fake_type",
-        )
-        assert result is None
-
-    def test_record_dashboard_feedback_valid(self):
-        """Valid dashboard feedback should be stored."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feedback_collector import FeedbackCollector
-        fc = FeedbackCollector()
-        result = fc.record_dashboard_feedback(
-            memory_id=42,
-            query="test query",
-            feedback_type="thumbs_up",
-        )
-        assert result is None or isinstance(result, int)
-
-    def test_record_dashboard_feedback_invalid_type(self):
-        """Invalid feedback type should return None."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feedback_collector import FeedbackCollector
-        fc = FeedbackCollector()
-        result = fc.record_dashboard_feedback(
-            memory_id=42,
-            query="test",
-            feedback_type="invalid_type",
-        )
-        assert result is None
-
-    def test_signal_values_complete(self):
-        """All declared signal types should have numeric values."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feedback_collector import FeedbackCollector
-        fc = FeedbackCollector()
-        for signal_type, value in fc.SIGNAL_VALUES.items():
-            assert isinstance(value, (int, float)), f"{signal_type} has non-numeric value"
-            assert 0.0 <= value <= 1.0, f"{signal_type} value {value} out of range"
-
-
-class TestFeatureExpansion:
-    """Test the 10→12→20 feature vector expansion."""
-
-    def test_feature_count_is_20(self):
-        """Feature vector should have 20 dimensions (v2.8)."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feature_extractor import FEATURE_NAMES, NUM_FEATURES
-        assert NUM_FEATURES == 20
-        assert len(FEATURE_NAMES) == 20
-
-    def test_new_features_present(self):
-        """signal_count and avg_signal_value should be in feature names."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feature_extractor import FEATURE_NAMES
-        assert 'signal_count' in FEATURE_NAMES
-        assert 'avg_signal_value' in FEATURE_NAMES
-        assert FEATURE_NAMES.index('signal_count') == 10
-        assert FEATURE_NAMES.index('avg_signal_value') == 11
-
-    def test_extract_features_returns_20(self):
-        """Extract should return 20-element vector (v2.8)."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feature_extractor import FeatureExtractor
-        fe = FeatureExtractor()
-        features = fe.extract_features(
-            {'id': 1, 'content': 'test', 'importance': 5},
-            'test'
-        )
-        assert len(features) == 20
-
-    def test_signal_features_with_stats(self):
-        """Signal features should use provided stats."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feature_extractor import FeatureExtractor
-        fe = FeatureExtractor()
-        fe.set_context(signal_stats={
-            '42': {'count': 8, 'avg_value': 0.9},
-        })
-        features = fe.extract_features(
-            {'id': 42, 'content': 'test', 'importance': 5},
-            'test'
-        )
-        assert features[10] == 0.8  # count=8, 8/10=0.8
-        assert features[11] == 0.9  # avg_value=0.9
-
-    def test_signal_features_without_stats(self):
-        """Signal features should default safely without stats."""
-        sys.path.insert(0, str(Path(__file__).parent.parent))
-        from feature_extractor import FeatureExtractor
-        fe = FeatureExtractor()
-        # No set_context call — signal_stats empty
-        features = fe.extract_features(
-            {'id': 99, 'content': 'test', 'importance': 5},
-            'test'
-        )
-        assert features[10] == 0.0  # No stats → 0.0
-        assert features[11] == 0.5  # No stats → neutral 0.5