feat(ai-services): add AVM v2 A/B comparison endpoint and tests

Add POST /avm/v2/compare-v1 endpoint that runs both v1 (single-model)
and v2 (ensemble) AVM predictions on the same property and returns a
side-by-side comparison with price diff, confidence delta, and a
recommendation on which model to prefer.

- ABComparisonRequest/Response schemas in avm_v2 models
- compare_v1() method in AVMv2EnsembleService
- 4 new integration tests for the comparison endpoint
- All 47 Python tests pass

Co-Authored-By: Paperclip <noreply@paperclip.ing>

libs/ai-services/app/routers/avm_v2.py

@@ -3,6 +3,8 @@
 from fastapi import APIRouter
 
 from app.models.avm_v2 import (
+    ABComparisonRequest,
+    ABComparisonResponse,
     AVMv2ModelInfo,
     AVMv2PredictRequest,
     AVMv2PredictResponse,
@@ -33,6 +35,16 @@ def train_v2(req: AVMv2TrainRequest) -> AVMv2TrainResponse:
     return avm_v2_service.train(req)
 
 
+@router.post("/compare-v1", response_model=ABComparisonResponse)
+def compare_v1(req: ABComparisonRequest) -> ABComparisonResponse:
+    """Compare v1 (single-model) vs v2 (ensemble) predictions side by side.
+
+    Runs both models on the same property and returns price difference,
+    confidence delta, and a recommendation on which to prefer.
+    """
+    return avm_v2_service.compare_v1(req)
+
+
 @router.get("/model-info", response_model=AVMv2ModelInfo)
 def model_info_v2() -> AVMv2ModelInfo:
     """Get current active ensemble model information."""