File size: 6,329 Bytes
d574a3d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 | # Phase 7 Web Server Launch Guide
**Ready**: Phase 7 MVP is fully integrated into codette_server.py
## What Happens When You Launch
```bash
codette_web.bat
```
### Initialization Sequence (Expected Console Output)
```
============================================================
Codette v2.0 - Phase 7 Executive Control Architecture
============================================================
Starting with intelligent component routing...
- Phase 7: Executive Controller (query routing)
- Phase 6: ForgeEngine (semantic tension, specialization)
- Phases 1-5: Core reasoning infrastructure
Initializing:
* CodetteOrchestrator with 8 domain LoRA adapters
* ForgeEngine with Query Classifier
* Executive Controller for intelligent routing
Testing locally at: http://localhost:7860
============================================================
Loading CodetteOrchestrator...
... (model loading, ~60-90 seconds first time)
Orchestrator ready: [newton, davinci, empathy, philosophy, quantum, consciousness, multi_perspective, systems_architecture]
Phase 6 bridge initialized
Phase 7 Executive Controller initialized
✓ Server ready on http://localhost:7860
```
### What's Working
✅ Phase 7 Executive Controller auto-initialized
✅ Phase 6 ForgeEngine wrapped behind bridge
✅ All 8 domain-specific LoRA adapters loaded
✅ Intelligent routing ready
---
## Testing Phase 7 in the Web UI
Once the server is running, **try these queries** to observe Phase 7 routing:
### Test 1: SIMPLE Query (Should be ~150-200ms)
```
"What is the speed of light?"
```
**Expected in Response**:
- Fast response (150-200ms actual)
- `phase7_routing.components_activated` should show all FALSE
- `phase7_routing.reasoning`: "SIMPLE factual query - orchestrator direct inference"
- No debate, no semantic tension, no conflicts
---
### Test 2: MEDIUM Query (Should be ~900ms-1200ms)
```
"How does quantum mechanics relate to consciousness?"
```
**Expected in Response**:
- Moderate latency (~900ms-1200ms)
- `phase7_routing.components_activated`:
- `debate`: TRUE (1 round)
- `semantic_tension`: TRUE
- `specialization_tracking`: TRUE
- `preflight_predictor`: FALSE (skipped for MEDIUM)
- Some conflicts detected (10-20 range)
---
### Test 3: COMPLEX Query (Should be ~2000-3000ms)
```
"Can machines be truly conscious? And how should we ethically govern AI?"
```
**Expected in Response**:
- Longer processing (~2000-3000ms)
- `phase7_routing.components_activated`: ALL TRUE
- Full debate (3 rounds)
- Higher conflict count (20-40 range)
- Deep synthesis with multiple perspectives
---
## Interpreting Response Metadata
Every response will include a `phase7_routing` section:
```json
{
"response": "The answer to your question...",
"phase7_routing": {
"query_complexity": "simple",
"components_activated": {
"debate": false,
"semantic_tension": false,
"specialization_tracking": false,
"preflight_predictor": false,
"memory_weighting": false,
"gamma_monitoring": false,
"synthesis": false
},
"reasoning": "SIMPLE factual query - avoided heavy machinery for speed",
"latency_analysis": {
"estimated_ms": 150,
"actual_ms": 142,
"savings_ms": 8
},
"correctness_estimate": 0.95,
"compute_cost": {
"estimated_units": 3,
"unit_scale": "1=classifier, 50=full_machinery"
},
"metrics": {
"conflicts_detected": 0,
"gamma_coherence": 0.95
}
}
}
```
### Key Fields to Watch
| Field | Meaning |
|-------|---------|
| `query_complexity` | SIMPLE/MEDIUM/COMPLEX classification |
| `components_activated` | Which Phase 1-6 components ran |
| `actual_ms` vs `estimated_ms` | Real latency vs prediction |
| `conflicts_detected` | How many conflicts were found |
| `gamma_coherence` | Coherence score (higher = more consistent) |
---
## Success Criteria for Phase 7 Validation
- [ ] Server launches with "Phase 7 Executive Controller initialized"
- [ ] SIMPLE queries complete in 150-250ms (2-3x faster than MEDIUM)
- [ ] MEDIUM queries complete in 800-1200ms
- [ ] COMPLEX queries complete in 2000-3500ms (uses full machinery)
- [ ] Response metadata shows correct component activation
- [ ] `phase7_routing.reasoning` matches expected routing decision
---
## If Something Goes Wrong
**Problem**: Server doesn't mention Phase 7
- Check: Is "Phase 7 Executive Controller initialized" in console?
- If missing: ForgeEngine failed to load (check model files)
**Problem**: All queries treated as COMPLEX
- Check: QueryClassifier patterns in `reasoning_forge/query_classifier.py`
- Common issue: Regex patterns too broad
**Problem**: Latencies not improving
- Check: Is `phase7_routing.components_activated.debate` FALSE for SIMPLE?
- If debate=TRUE on simple queries: Classifier misclassifying
**Problem**: Response metadata missing phase7_routing
- Check: Is `phase7_used` set to TRUE in response?
- If FALSE: Bridge fallback happened (check console errors)
---
## Next Steps After Testing
### If Validation Successful (Expected Path)
1. ✅ Document actual latencies (compare to estimates)
2. ✅ Verify correctness not degraded on MEDIUM/COMPLEX
3. → Move to **Path B: Benchmarking** to quantify improvements
### If Issues Found
1. Document the specific problem
2. Check console logs for error messages
3. Fix and retest with `python run_phase7_demo.py` first
---
## Browser Tool UI Notes
The web interface will show:
- **Response** - The actual answer
- **Metadata** - Below response, includes phase7_routing
- **Latency** - Actual time taken (compare to estimated_ms)
Scroll down to see full phase7_routing metadata in JSON format.
---
## Ready to Launch?
```bash
codette_web.bat
```
Open browser to: **http://localhost:7860**
Test with one of the queries above and look for:
- ✅ Phase 7 routing metadata in response
- ✅ Latency improvements on SIMPLE queries
- ✅ Component activation matching query complexity
**Questions during testing?** Check the metadata for clues about routing decisions.
|