Update engine/features.py

engine/features.py

@@ -4,7 +4,6 @@ import numpy as np
 import re
 from typing import Dict, List, Any
 
-
 # ------------------------------------------------------------------
 # Load schema once
 # ------------------------------------------------------------------
@@ -16,7 +15,6 @@ with open(_FEATURE_SCHEMA_PATH, "r", encoding="utf-8") as f:
 
 FEATURES = SCHEMA["features"]
 
-
 # ------------------------------------------------------------------
 # Helper mappings
 # ------------------------------------------------------------------
@@ -26,7 +24,7 @@ PNV_MAP = {
     "negative": -1.0,
     "variable": 0.5,
     "unknown": 0.0,
-    None: 0.0
+    None: 0.0,
 }
 
 SHAPE_MAP = {
@@ -36,7 +34,7 @@ SHAPE_MAP = {
     "spiral": 3.0,
     "yeast": 4.0,
     "variable": 0.5,
-    "unknown": 0.0
+    "unknown": 0.0,
 }
 
 OXYGEN_MAP = {
@@ -45,14 +43,14 @@ OXYGEN_MAP = {
     "facultative anaerobe": 3.0,
     "microaerophilic": 4.0,
     "capnophilic": 5.0,
-    "unknown": 0.0
+    "unknown": 0.0,
 }
 
-
 # ------------------------------------------------------------------
 # Normalisation helpers
 # ------------------------------------------------------------------
 
+
 def _norm(s: Any) -> str:
     if not s:
         return "unknown"
@@ -71,26 +69,29 @@ def _map_oxygen(x: Any) -> float:
     return OXYGEN_MAP.get(_norm(x), 0.0)
 
 
-def _growth_minmax(v: str):
+def _growth_minmax(v: Any) -> tuple[float, float]:
     """
     Convert '30//37' → (30, 37)
-    If missing, return (0, 0)
+    If missing or malformed, return (0, 0).
     """
     if not v:
         return (0.0, 0.0)
+    if not isinstance(v, str):
+        v = str(v)
     m = re.match(r"^\s*(\d+)\s*//\s*(\d+)\s*$", v)
     if not m:
         return (0.0, 0.0)
     return (float(m.group(1)), float(m.group(2)))
 
 
-def _media_flag(media_field: str, medium: str) -> float:
+def _media_flag(media_field: Any, medium: str) -> float:
     """
-    Return 1.0 if medium appears in media list, else 0.0.
+    Return 1.0 if 'medium' appears in 'Media Grown On' field, else 0.0.
+    e.g. medium='MacConkey Agar' and media_field='Blood Agar; MacConkey Agar'
     """
     if not media_field:
         return 0.0
-    mf = media_field.lower()
+    mf = str(media_field).lower()
     return 1.0 if medium.lower() in mf else 0.0
 
 
@@ -98,18 +99,26 @@ def _media_flag(media_field: str, medium: str) -> float:
 # Main public function
 # ------------------------------------------------------------------
 
+
 def extract_feature_vector(fused_fields: Dict[str, Any]) -> np.ndarray:
     """
     Convert fused tri-fusion fields into a fixed-length numeric vector.
     Ordered exactly according to feature_schema.json.
-    Unknowns → 0.0.
+
+    Missing/unknown → 0.0.
     """
     vec: List[float] = []
 
+    # We iterate over schema in order and push one numeric value per feature
+    # (except Growth Temp Min/Max, which split the single 'Growth Temperature'
+    # field into two numeric components).
+    growth_value = fused_fields.get("Growth Temperature")
+
     for f in FEATURES:
         name = f["name"]
         kind = f["kind"]
 
+        # Default: look up by the same name in fused_fields
         value = fused_fields.get(name, "Unknown")
 
         if kind == "pnv":
@@ -122,21 +131,28 @@ def extract_feature_vector(fused_fields: Dict[str, Any]) -> np.ndarray:
             vec.append(_map_oxygen(value))
 
         elif kind == "numeric_from_growth_temp":
-            low, high = _growth_minmax(value)
-            vec.append(low)
-            vec.append(high)
-            # IMPORTANT: skip the next schema feature
-            # (schema should include two entries but model expects two values)
-            continue
+            # We assume:
+            #   - "Growth Temp Min" → min
+            #   - "Growth Temp Max" → max
+            low, high = _growth_minmax(growth_value)
+            if "min" in name.lower():
+                vec.append(low)
+            elif "max" in name.lower():
+                vec.append(high)
+            else:
+                # Fallback: just append 0 if schema name is unexpected
+                vec.append(0.0)
 
         elif kind == "media_flag":
-            # Each media entry in schema specifies the medium name
-            # e.g. "MacConkey Growth"
+            # For media flags we derive the medium name from the feature name.
+            # e.g. "MacConkey Growth" → medium="MacConkey"
+            #      "Nutrient Growth"  → medium="Nutrient"
             medium = name.replace("Growth", "").strip()
-            vec.append(_media_flag(fused_fields.get("Media Grown On"), medium))
+            media_field = fused_fields.get("Media Grown On")
+            vec.append(_media_flag(media_field, medium))
 
         else:
-            # Unknown kind → default numeric 0
+            # Unknown kind → default numeric 0.0
             vec.append(0.0)
 
     return np.array(vec, dtype=float)