Update engine/parser_rules.py

engine/parser_rules.py

@@ -1185,79 +1185,152 @@ def _parse_colony(text_lc: str, parsed: Dict[str, str]) -> None:
                 _set_if_stronger(parsed, "Colony Morphology", pretty)
                 return
 
-# ============================================================
-#  PARSER_PATCH_v1
-#  Extends missing behaviours without modifying core parser.
-#  This patch is applied AFTER all existing parsing logic.
-#
-#  Adds:
-#    • Haemolysis ± without type      ("haemolysis positive")
-#    • Motility ± phrases             ("motility positive")
-#    • Spore Formation ±              ("spore formation negative")
-#    • NaCl tolerance direct ±        ("NaCl tolerant positive")
-#    • Growth Temperature parsing of:
-#         "20/40", "20//40", "20 / 40"
-#    • Multi-media extraction:
-#         "Media grown on: Blood Agar; MSA; MacConkey Agar"
-#      → merges with existing parsed value WITHOUT overwriting
-# ============================================================
-
 def _apply_patches(original_text: str, text_lc: str, parsed: Dict[str, str]) -> Dict[str, str]:
     # ----------------------------------------------
-    # 1. Haemolysis: generic ± without type
+    # helper for P/N/V
     # ----------------------------------------------
+    def _pnv(x: str) -> Optional[str]:
+        x = x.strip().lower()
+        if x in {"positive", "pos", "+", "strongly positive", "weakly positive"}:
+            return "Positive"
+        if x in {"negative", "neg", "-", "no"}:
+            return "Negative"
+        if x in {"variable", "var", "mixed"}:
+            return "Variable"
+        return None
+
+    # ============================================================
+    #  NEW LOGIC: Haemolysis Type detection (alpha/beta/none)
+    # ============================================================
+
+    # alpha
+    m_alpha = re.search(r"(alpha|α)[-\s]*haemolysis", text_lc) or \
+              re.search(r"haemolysis type[: ]*(alpha|α)", text_lc)
+    if m_alpha:
+        if parsed.get("Haemolysis", UNKNOWN) == UNKNOWN:
+            parsed["Haemolysis"] = "Positive"
+        if parsed.get("Haemolysis Type", UNKNOWN) == UNKNOWN:
+            parsed["Haemolysis Type"] = "Alpha"
+
+    # beta
+    m_beta = re.search(r"(beta|β)[-\s]*haemolysis", text_lc) or \
+             re.search(r"haemolysis type[: ]*(beta|β)", text_lc)
+    if m_beta:
+        if parsed.get("Haemolysis", UNKNOWN) == UNKNOWN:
+            parsed["Haemolysis"] = "Positive"
+        if parsed.get("Haemolysis Type", UNKNOWN) == UNKNOWN:
+            parsed["Haemolysis Type"] = "Beta"
+
+    # gamma / none
+    m_gamma = re.search(r"(gamma|γ)[-\s]*haemolysis", text_lc)
+    m_none = re.search(r"(no haemolysis|non[- ]haemolytic|no hemolysis|non[- ]hemolytic)", text_lc)
+    if m_gamma or m_none:
+        if parsed.get("Haemolysis", UNKNOWN) == UNKNOWN:
+            parsed["Haemolysis"] = "Negative"
+        if parsed.get("Haemolysis Type", UNKNOWN) == UNKNOWN:
+            parsed["Haemolysis Type"] = "None"
+
+    # ============================================================
+    # ORIGINAL PATCH v1 LOGIC (fully preserved)
+    # ============================================================
+
+    # 1. Haemolysis: generic ± without type
     m_h = re.search(r"haemolysis\s+(positive|negative|variable|pos|neg|\+|\-)", text_lc)
     if m_h and "Haemolysis" not in parsed:
-        val = _value_from_pnv_context(m_h.group(1))
+        val = _pnv(m_h.group(1))
         if val:
             parsed["Haemolysis"] = val
+            if parsed.get("Haemolysis Type", UNKNOWN) == UNKNOWN and val == "Positive":
+                parsed["Haemolysis Type"] = "Unknown"
 
-    # ----------------------------------------------
     # 2. Motility: generic ±
-    # ----------------------------------------------
     m_mot = re.search(r"motility\s+(positive|negative|variable|pos|neg|\+|\-)", text_lc)
     if m_mot and "Motility" not in parsed:
-        val = _value_from_pnv_context(m_mot.group(1))
+        val = _pnv(m_mot.group(1))
         if val:
             parsed["Motility"] = val
 
-    # ----------------------------------------------
     # 3. Spore formation ±
-    # ----------------------------------------------
     m_sp = re.search(r"spore formation\s+(positive|negative|variable|pos|neg|\+|\-)", text_lc)
     if m_sp and parsed.get("Spore Formation", UNKNOWN) == UNKNOWN:
-        val = _value_from_pnv_context(m_sp.group(1))
+        val = _pnv(m_sp.group(1))
         if val:
             parsed["Spore Formation"] = val
 
-    # ----------------------------------------------
-    # 4. NaCl tolerance ± (simple)
-    # ----------------------------------------------
-    m_nacl = re.search(r"nacl(?:\s+tolerant)?\s+(positive|negative|variable|pos|neg|\+|\-)", text_lc)
-    if m_nacl and parsed.get("NaCl Tolerant (>=6%)", UNKNOWN) == UNKNOWN:
-        val = _value_from_pnv_context(m_nacl.group(1))
-        if val:
-            parsed["NaCl Tolerant (>=6%)"] = val
+    # ============================================================
+    #  FIXED NaCl tolerant logic (patch upgrade)
+    # ============================================================
+    if parsed.get("NaCl Tolerant (>=6%)", UNKNOWN) == UNKNOWN:
 
-    # ----------------------------------------------
-    # 5. Growth Temperature patterns like:
-    #       "20/40", "20 / 40", "20//40"
-    # ----------------------------------------------
+        # direct p/n/v
+        m_nacl = re.search(
+            r"(?:nacl\s*(?:tolerant|tolerance)?|growth\s+in\s+6\%[\s]*nacl)"
+            r"\s*(positive|negative|variable|pos|neg|\+|\-)",
+            text_lc
+        )
+        if m_nacl:
+            val = _pnv(m_nacl.group(1))
+            if val:
+                parsed["NaCl Tolerant (>=6%)"] = val
+
+        # "no growth in 6% nacl"
+        if parsed.get("NaCl Tolerant (>=6%)", UNKNOWN) == UNKNOWN:
+            if re.search(r"no\s+growth\s+in\s+(?:>=)?\s*6\%?\s*nacl", text_lc):
+                parsed["NaCl Tolerant (>=6%)"] = "Negative"
+
+        # "grows in 6% nacl"
+        if parsed.get("NaCl Tolerant (>=6%)", UNKNOWN) == UNKNOWN:
+            if re.search(r"grows?\s+in\s+(?:>=)?\s*6\%?\s*nacl", text_lc):
+                parsed["NaCl Tolerant (>=6%)"] = "Positive"
+
+    # ============================================================
+    #  Growth Temperature patterns (20/40, 20//40, 20 / 40)
+    # ============================================================
     m_temp = re.search(r"\b(\d{1,3})\s*[/]{1,2}\s*(\d{1,3})\b", text_lc)
-    if m_temp:
-        low, high = m_temp.group(1), m_temp.group(2)
-        # Only set if no better structured temp already present
-        if parsed.get("Growth Temperature", UNKNOWN) == UNKNOWN:
-            parsed["Growth Temperature"] = f"{low}//{high}"
+    if m_temp and parsed.get("Growth Temperature", UNKNOWN) == UNKNOWN:
+        parsed["Growth Temperature"] = f"{m_temp.group(1)}//{m_temp.group(2)}"
+
+    # ============================================================
+    #  Colony Morphology STRICT LIST extraction
+    # ============================================================
+    COLONY_TRIGGERS = [
+        "colony morphology",
+        "colonies are",
+        "colonies appear",
+        "colonies look",
+        "colony appearance",
+        "colony characteristics",
+    ]
+    if any(t in text_lc for t in COLONY_TRIGGERS):
+        m_col = re.search(
+            r"(?:colony morphology|colonies are|colonies appear|colonies look|colony appearance|colony characteristics)"
+            r"[: ]+([a-z0-9 ,;/\-]+)",
+            text_lc
+        )
+        if m_col:
+            segment = m_col.group(1)
+            parts = [x.strip() for x in re.split(r"[;,/]", segment) if x.strip()]
 
-    # ----------------------------------------------
-    # 6. Multi-media patch:
-    #       extract all media separated by ',' or ';'
-    #       preserve order from text
-    #       merge with parsed["Media Grown On"]
-    # ----------------------------------------------
+            clean_desc = [p.capitalize() for p in parts if len(p) > 1]
+
+            if clean_desc:
+                existing = parsed.get("Colony Morphology", "")
+                existing_list = [x.strip() for x in existing.split(";")] if existing else []
+
+                merged = []
+                for x in existing_list:
+                    if x not in merged:
+                        merged.append(x)
+                for x in clean_desc:
+                    if x not in merged:
+                        merged.append(x)
+
+                parsed["Colony Morphology"] = "; ".join(merged)
+
+    # ============================================================
+    #  ORIGINAL MULTI-MEDIA PATCH (unchanged)
+    # ============================================================
     if "media grown on" in text_lc or "grown on" in text_lc:
-        # capture segment after "grown on"
         mm = re.search(r"(?:media\s+grown\s+on|grown\s+on)[: ]+([a-z0-9 ,;/\-]+)", text_lc)
         if mm:
             segment = mm.group(1)
@@ -1273,11 +1346,8 @@ def _apply_patches(original_text: str, text_lc: str, parsed: Dict[str, str]) ->
 
             if detected_media:
                 existing = parsed.get("Media Grown On", "")
-                existing_list = []
-                if existing:
-                    existing_list = [x.strip() for x in existing.split(";")]
+                existing_list = [x.strip() for x in existing.split(";")] if existing else []
 
-                # Merge while preserving order from the text
                 merged = []
                 for m in existing_list:
                     if m not in merged:
@@ -1289,6 +1359,7 @@ def _apply_patches(original_text: str, text_lc: str, parsed: Dict[str, str]) ->
                 parsed["Media Grown On"] = "; ".join(merged)
 
     return parsed
+
 # ------------------------------------------------------------
 # PUBLIC API
 # ------------------------------------------------------------