mirror of
https://github.com/proffesor-for-testing/agentic-qe.git
synced 2026-09-19 08:45:47 +08:00
b3d2018ee2
Critical bug fixes and code quality improvements from brutal-honest-review. ## Critical Fixes (9/10 completed, 1 deferred) 1. Fixed Agent Pattern Isolation (CRITICAL) - Added SwarmMemoryManager.queryPatternsByAgent() - SQL filtering by agent_id in metadata - Each agent now isolated 2. Fixed Async Method Cascade - Changed getPatterns() to async - Fixed 10 callers across 6 files - Build passes 3. Fixed Misleading Logging - Added isRealAgentDB() detection - Honest mock vs real reporting 4. Consolidated Embedding Generation - Created EmbeddingGenerator utility - Removed 50+ duplicate lines 5. Added AgentDB Initialization Checks - Empty DB detection - HNSW index verification 6. Integration Test Suite - 26 new test cases - Error path coverage 7. Repository Cleanup - Removed 19 temp test files 8. New QE Skill: sherlock-review - Evidence-based investigative review - 38 QE skills total 9. Deferred: Wire Up Real Test Execution - Deferred to v1.9.0 (documented) ## Statistics - Fixes: 9/10 (90%, 1 deferred) - Files Modified: 16 - Files Created: 3 - Files Deleted: 19 - Integration Tests: 26 - Build: PASSING ## Breaking Changes - getPatterns() now async - add await 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
5.8 KiB
5.8 KiB
AgentDB v2.0 Quick Reference Card
📋 Essential Tables
Core Learning
episodes -- Historical learning data (1,759 rows)
test_patterns -- Reusable test templates
pattern_usage -- Pattern effectiveness tracking
learning_metrics -- Improvement over time
Reinforcement Learning
q_values -- Q-learning state-action values
learning_experiences -- Experience replay buffer
Discovery
pattern_similarity_index -- Pre-computed similarities
pattern_fts -- Full-text search (virtual)
🚀 Common Queries
Find Best Patterns
SELECT pattern_name, success_rate, usage_count
FROM test_patterns
WHERE framework = 'jest' AND success_rate >= 0.8
ORDER BY success_rate DESC LIMIT 10;
Track Agent Progress
SELECT DATE(timestamp, 'unixepoch') as date,
AVG(coverage_percent) as coverage
FROM learning_metrics
WHERE agent_type = 'qe-test-generator'
AND timestamp >= strftime('%s', 'now', '-30 days')
GROUP BY DATE(timestamp, 'unixepoch');
Search Patterns (Full-Text)
SELECT p.pattern_name, bm25(pattern_fts) as relevance
FROM pattern_fts
JOIN test_patterns p ON pattern_fts.rowid = p.rowid
WHERE pattern_fts MATCH 'authentication error'
ORDER BY relevance DESC;
High-Coverage Episodes
SELECT session_id, task, coverage_after, test_count
FROM episodes
WHERE coverage_after >= 80.0
AND success = 1
ORDER BY coverage_after DESC;
📊 Key Views
v_pattern_performance -- Pattern metrics + usage stats
v_agent_learning_progress -- Agent improvement dashboard
v_top_patterns -- High-performing patterns
🔍 Performance Indexes
-- Episodes
idx_episodes_framework -- (test_framework, test_type)
idx_episodes_coverage -- (coverage_after DESC)
-- Patterns
idx_patterns_framework_type -- (framework, pattern_type)
idx_patterns_success_rate -- (success_rate DESC)
idx_patterns_dedup (UNIQUE) -- (code_signature_hash, framework)
-- Learning
idx_metrics_agent -- (agent_id, timestamp)
idx_metrics_improvement -- (improvement_percentage DESC)
🎯 Key Metrics
Episodes Table
- 1,759 episodes migrated
- New columns: test_framework, test_type, coverage_before/after
- Constraints: reward (0-1), coverage (0-100)
Test Patterns Table
- 0 patterns currently (ready for data)
- 10+ indexes for fast queries
- Deduplication: code_signature_hash + framework
- Trend tracking: improving/declining/stable
Performance Targets
- Pattern lookup: < 1ms
- Framework filter: < 10ms
- Full-text search: < 100ms
- Aggregate queries: < 100ms
🔧 Useful Commands
Database Statistics
SELECT COUNT(*) FROM episodes;
SELECT COUNT(*) FROM test_patterns;
SELECT COUNT(*) FROM learning_metrics;
Schema Version
SELECT * FROM schema_version WHERE version = '2.0.0';
Rebuild Performance
REINDEX;
ANALYZE;
VACUUM;
Clear Cache
DELETE FROM pattern_stats_cache
WHERE strftime('%s', 'now') > expires_at;
📈 Data Types
TEXT - Strings, JSON
INTEGER - Timestamps (Unix epoch), counts
REAL - Rates (0-1), percentages (0-100)
BOOLEAN - 0 or 1
JSON - Validated with CHECK(json_valid(...))
BLOB - Vector embeddings
🎨 Pattern Types
edge-case error-handling integration
boundary unit e2e
performance security accessibility
regression smoke contract
mutation
🧪 Frameworks Supported
jest mocha vitest
playwright cypress ava
tape jasmine qunit
testcafe
📊 Metric Types
coverage -- Test coverage %
quality -- Test quality score
latency -- Execution time
success_rate -- Success percentage
improvement -- Overall improvement
accuracy -- Prediction accuracy
test_count -- Number of tests
pattern_reuse -- Pattern reuse rate
code_quality -- Code quality score
🔄 Triggers (Auto-Execute)
update_pattern_stats -- Updates on pattern usage
update_pattern_trend -- Detects improving/declining
patterns_ai/ad/au -- Syncs FTS index
💾 SQLite Settings
PRAGMA journal_mode = WAL; -- Concurrent access
PRAGMA cache_size = -64000; -- 64MB cache
PRAGMA synchronous = NORMAL; -- Balanced performance
PRAGMA mmap_size = 268435456; -- 256MB mmap
🚨 Common Troubleshooting
Database Locked
lsof | grep agentdb.db
pkill -f agentdb
Slow Queries
EXPLAIN QUERY PLAN SELECT ...;
REINDEX;
ANALYZE;
Check Integrity
echo "PRAGMA integrity_check;" | sqlite3 agentdb.db
📁 File Locations
agentdb.db -- Main database
docs/database/schema-v2.sql -- Schema definition
docs/database/schema-v2.md -- Full documentation
docs/database/migration-v1-to-v2.md -- Migration guide
docs/database/example-queries.sql -- Query examples
🎯 Quick Migration
# 1. Backup
cp agentdb.db agentdb-backup.db
# 2. Create v2.0
node create-v2-schema.js
# 3. Migrate episodes
node migrate-episodes.js
# 4. Verify
node verify-migration.js
# 5. Switch
mv agentdb.db agentdb-v1.db
mv agentdb-v2.db agentdb.db
Time: 3-5 minutes Data Loss: None (100% preserved)
📞 Support
Documentation: /workspaces/agentic-qe-cf/docs/database/
Examples: example-queries.sql
Schema: schema-v2.sql (500+ lines)
Guide: schema-v2.md (1,200+ lines)
Version: 2.0.0 Date: 2025-11-16 Status: Production Ready Episodes Preserved: 1,759 / 1,759 ✅