feat(ai-services): add building_coverage, loading_docks, zoning to industrial AVM

Completes the industrial-specific feature set required for AVM industrial valuation. Adds heuristic adjustments for all three new features and 4 new tests covering zoning premiums, loading docks, and coverage ratio. Co-Authored-By: Paperclip <noreply@paperclip.ing>
2026-04-16 17:06:27 +07:00
parent 8592fb436c
commit 13bd76ac5d
3 changed files with 115 additions and 8 deletions
--- a/libs/ai-services/app/models/avm_industrial.py
+++ b/libs/ai-services/app/models/avm_industrial.py
@@ -37,6 +37,20 @@ class IndustrialAVMRequest(BaseModel):
    power_capacity_kva: float = Field(
        0.0, ge=0, description="Allocated power capacity in kVA"
    )
+    building_coverage: float = Field(
+        0.0,
+        ge=0,
+        le=1,
+        description="Building coverage ratio — footprint area / lot area (0-1)",
+    )
+    loading_docks: int = Field(
+        0, ge=0, description="Number of loading docks / bays"
+    )
+    zoning: str = Field(
+        "general_industrial",
+        description="Industrial zoning category: general_industrial, heavy_industrial, "
+        "light_industrial, logistics, free_trade_zone, high_tech",
+    )
    industry_demand_index: float = Field(
        0.5, ge=0, le=1, description="Local industry demand index (0-1)"
    )
--- a/libs/ai-services/app/services/avm_industrial_service.py
+++ b/libs/ai-services/app/services/avm_industrial_service.py
@@ -34,6 +34,9 @@ INDUSTRIAL_FEATURE_NAMES = [
    "ceiling_height_m",
    "floor_load_ton_m2",
    "power_capacity_kva",
+    "building_coverage",
+    "loading_docks",
+    "zoning_encoded",
    "industry_demand_index",
    "fdi_province_musd",
    "labor_cost_province_vnd",
@@ -56,6 +59,15 @@ PROPERTY_TYPE_MAP = {
    "office_in_park": 5,
 }

+ZONING_MAP = {
+    "general_industrial": 0,
+    "heavy_industrial": 1,
+    "light_industrial": 2,
+    "logistics": 3,
+    "free_trade_zone": 4,
+    "high_tech": 5,
+}
+
 # ── Province-level rent baselines (USD/m²/month) ────────────────
 # Based on Vietnamese industrial real estate market data
 PROVINCE_BASELINE: dict[str, float] = {
@@ -113,6 +125,9 @@ def _encode_features(req: IndustrialAVMRequest) -> np.ndarray:
            req.ceiling_height_m,
            req.floor_load_ton_m2,
            req.power_capacity_kva,
+            req.building_coverage,
+            req.loading_docks,
+            ZONING_MAP.get(req.zoning.lower(), 0),
            req.industry_demand_index,
            req.fdi_province_musd,
            req.labor_cost_province_vnd,
@@ -253,6 +268,22 @@ class IndustrialAVMService:
        floor_load_adj = 1.0 + max(0.0, (req.floor_load_ton_m2 - 2.0) * 0.03)
        power_adj = 1.0 + min(0.10, req.power_capacity_kva / 5000 * 0.10)

+        # Building coverage — higher coverage = more usable space = premium
+        coverage_adj = 1.0 + max(0.0, (req.building_coverage - 0.4) * 0.15)
+
+        # Loading docks — each dock adds a small premium, diminishing returns
+        docks_adj = 1.0 + min(0.12, req.loading_docks * 0.02)
+
+        # Zoning premium — specialized zones command higher rents
+        zoning_mult = {
+            "general_industrial": 1.00,
+            "heavy_industrial": 0.95,
+            "light_industrial": 1.05,
+            "logistics": 1.10,
+            "free_trade_zone": 1.20,
+            "high_tech": 1.25,
+        }.get(req.zoning.lower(), 1.0)
+
        # Economic indicators
        demand_adj = 1.0 + (req.industry_demand_index - 0.5) * 0.25
        fdi_adj = 1.0 + min(0.15, req.fdi_province_musd / 5000 * 0.15)
@@ -280,6 +311,9 @@ class IndustrialAVMService:
            * ceiling_adj
            * floor_load_adj
            * power_adj
+            * coverage_adj
+            * docks_adj
+            * zoning_mult
            * demand_adj
            * fdi_adj
            * labor_adj
@@ -291,14 +325,17 @@ class IndustrialAVMService:

        # Heuristic feature importance
        drivers = [
-            FeatureImportance(feature="province_baseline", importance=0.20),
-            FeatureImportance(feature="property_type", importance=0.15),
-            FeatureImportance(feature="park_occupancy_rate", importance=0.12),
-            FeatureImportance(feature="logistics_connectivity_score", importance=0.10),
-            FeatureImportance(feature="industry_demand_index", importance=0.10),
-            FeatureImportance(feature="fdi_province_musd", importance=0.08),
-            FeatureImportance(feature="distance_to_port_km", importance=0.07),
-            FeatureImportance(feature="area_m2", importance=0.06),
+            FeatureImportance(feature="province_baseline", importance=0.16),
+            FeatureImportance(feature="property_type", importance=0.12),
+            FeatureImportance(feature="zoning", importance=0.11),
+            FeatureImportance(feature="park_occupancy_rate", importance=0.10),
+            FeatureImportance(feature="logistics_connectivity_score", importance=0.09),
+            FeatureImportance(feature="industry_demand_index", importance=0.09),
+            FeatureImportance(feature="building_coverage", importance=0.07),
+            FeatureImportance(feature="loading_docks", importance=0.06),
+            FeatureImportance(feature="fdi_province_musd", importance=0.06),
+            FeatureImportance(feature="distance_to_port_km", importance=0.05),
+            FeatureImportance(feature="area_m2", importance=0.05),
        ]

        return IndustrialAVMResponse(
--- a/libs/ai-services/tests/test_avm_industrial.py
+++ b/libs/ai-services/tests/test_avm_industrial.py
@@ -109,6 +109,62 @@ def test_predict_industrial_total_rent():
    assert abs(resp["total_monthly_rent_usd"] - expected_total) < 1.0


+def test_predict_industrial_free_trade_zone_premium():
+    """Free-trade-zone zoning should command higher rent than general_industrial."""
+    general = client.post(
+        "/avm/industrial/predict",
+        json={**_PREDICT_PAYLOAD, "zoning": "general_industrial"},
+    ).json()
+    ftz = client.post(
+        "/avm/industrial/predict",
+        json={**_PREDICT_PAYLOAD, "zoning": "free_trade_zone"},
+    ).json()
+
+    assert ftz["estimated_rent_usd_m2"] > general["estimated_rent_usd_m2"]
+
+
+def test_predict_industrial_high_tech_zone_premium():
+    """High-tech zoning should command higher rent than general_industrial."""
+    general = client.post(
+        "/avm/industrial/predict",
+        json={**_PREDICT_PAYLOAD, "zoning": "general_industrial"},
+    ).json()
+    ht = client.post(
+        "/avm/industrial/predict",
+        json={**_PREDICT_PAYLOAD, "zoning": "high_tech"},
+    ).json()
+
+    assert ht["estimated_rent_usd_m2"] > general["estimated_rent_usd_m2"]
+
+
+def test_predict_industrial_loading_docks_premium():
+    """More loading docks should increase rent."""
+    no_docks = client.post(
+        "/avm/industrial/predict",
+        json={**_PREDICT_PAYLOAD, "loading_docks": 0},
+    ).json()
+    many_docks = client.post(
+        "/avm/industrial/predict",
+        json={**_PREDICT_PAYLOAD, "loading_docks": 6},
+    ).json()
+
+    assert many_docks["estimated_rent_usd_m2"] > no_docks["estimated_rent_usd_m2"]
+
+
+def test_predict_industrial_building_coverage_premium():
+    """Higher building coverage should increase rent."""
+    low_cov = client.post(
+        "/avm/industrial/predict",
+        json={**_PREDICT_PAYLOAD, "building_coverage": 0.3},
+    ).json()
+    high_cov = client.post(
+        "/avm/industrial/predict",
+        json={**_PREDICT_PAYLOAD, "building_coverage": 0.7},
+    ).json()
+
+    assert high_cov["estimated_rent_usd_m2"] > low_cov["estimated_rent_usd_m2"]
+
+
 def test_predict_industrial_validation_error():
    """Missing required fields should return 422."""
    resp = client.post("/avm/industrial/predict", json={"area_m2": 5000})