feat(prediction): fleet_tracker NAME_FUZZY 매칭 추가 (A-2)

_normalize_vessel_name()로 선박번호 suffix(호/號/号/NoN/#N) +
공백/구두점 제거 후 upper() 통일. EXACT 실패 시 FUZZY 단계로
매칭 시도. 동명이 후보 2개 이상이거나 이미 다른 MMSI 할당된 vid는
제외하여 중복 방지.

- match_ais_to_registry: NAME_EXACT → NAME_FUZZY (confidence 0.80)
- track_gear_identity: parent_name 매칭에도 FUZZY 적용
- _name_fuzzy_map 캐시로 1회 lookup

검증 목표: fleet_vessels.mmsi 매칭률 8.7% → 30%+

prediction/fleet_tracker.py

@@ -23,13 +23,40 @@ GEAR_IDENTITY_LOG = qualified_table('gear_identity_log')
 GEAR_CORRELATION_SCORES = qualified_table('gear_correlation_scores')
 FLEET_TRACKING_SNAPSHOT = qualified_table('fleet_tracking_snapshot')
 
+# 선박명 정규화: 공백/구두점 제거, 선박번호 suffix 제거, upper() 통일.
+# AIS 이름과 fishery_permit_cn 이름 간 suffix/공백 차이로 NAME_EXACT 매칭률 8.7% → 정규화 기반 매칭으로 회복.
+_NAME_STRIP_SUFFIX = re.compile(
+    r'(?:'
+    r'[\s_]*(?:NO\.?|#|＃)?[\s_]*\d+\s*(?:호|號|号)?'  # 123 / No.123 / #123 / 123호 / 12号
+    r'|[\s_]*(?:호|號|号)'                              # 말미 호/號/号 단독
+    r')\s*$',
+    re.IGNORECASE,
+)
+_NAME_STRIP_CHARS = re.compile(r'[\s\-_./,()\[\]·•\u3000]+')
+
+
+def _normalize_vessel_name(name: Optional[str]) -> str:
+    """선박명을 매칭용으로 정규화.
+
+    1. upper() + strip
+    2. 말미 선박번호 패턴 제거 (호/號/号, No.N, #N, 공백숫자)
+    3. 남은 공백/구두점 제거
+    """
+    if not name:
+        return ''
+    s = name.strip().upper()
+    s = _NAME_STRIP_SUFFIX.sub('', s)
+    s = _NAME_STRIP_CHARS.sub('', s)
+    return s
+
 
 class FleetTracker:
     def __init__(self) -> None:
         self._companies: dict[int, dict] = {}       # id → {name_cn, name_en}
         self._vessels: dict[int, dict] = {}         # id → {permit_no, name_cn, ...}
-        self._name_cn_map: dict[str, int] = {}      # name_cn → vessel_id
-        self._name_en_map: dict[str, int] = {}      # name_en(lowercase) → vessel_id
+        self._name_cn_map: dict[str, int] = {}      # name_cn → vessel_id (정확일치)
+        self._name_en_map: dict[str, int] = {}      # name_en(lowercase) → vessel_id (정확일치)
+        self._name_fuzzy_map: dict[str, list[int]] = {}  # 정규화 이름 → [vessel_id, ...]
         self._mmsi_to_vid: dict[str, int] = {}      # mmsi → vessel_id (매칭된 것만)
         self._gear_active: dict[str, dict] = {}     # mmsi → {name, parent_mmsi, ...}
         self._last_registry_load: float = 0.0
@@ -54,6 +81,7 @@ class FleetTracker:
         self._vessels = {}
         self._name_cn_map = {}
         self._name_en_map = {}
+        self._name_fuzzy_map = {}
         self._mmsi_to_vid = {}
 
         for r in cur.fetchall():
@@ -73,8 +101,14 @@ class FleetTracker:
             self._vessels[vid] = v
             if r[3]:
                 self._name_cn_map[r[3]] = vid
+                key_cn = _normalize_vessel_name(r[3])
+                if key_cn:
+                    self._name_fuzzy_map.setdefault(key_cn, []).append(vid)
             if r[4]:
                 self._name_en_map[r[4].lower().strip()] = vid
+                key_en = _normalize_vessel_name(r[4])
+                if key_en:
+                    self._name_fuzzy_map.setdefault(key_en, []).append(vid)
             if r[9]:
                 self._mmsi_to_vid[r[9]] = vid
 
@@ -165,7 +199,8 @@ class FleetTracker:
         ais_vessels: [{mmsi, name, lat, lon, sog, cog}, ...]
         """
         cur = conn.cursor()
-        matched = 0
+        matched_exact = 0
+        matched_fuzzy = 0
 
         for v in ais_vessels:
             mmsi = v.get('mmsi', '')
@@ -185,22 +220,47 @@ class FleetTracker:
             vid: Optional[int] = self._name_cn_map.get(name)
             if not vid:
                 vid = self._name_en_map.get(name.lower().strip())
+            method = 'NAME_EXACT'
+            confidence = 0.95
+
+            # NAME_FUZZY 매칭 (정규화 후 lookup)
+            if not vid:
+                key = _normalize_vessel_name(name)
+                if key:
+                    candidates = self._name_fuzzy_map.get(key, [])
+                    # 이미 다른 MMSI에 할당된 vid 제외 → 동명이 중복 매칭 방지
+                    unassigned = [
+                        c for c in candidates
+                        if not self._vessels.get(c, {}).get('mmsi')
+                        or self._vessels[c].get('mmsi') == mmsi
+                    ]
+                    if len(unassigned) == 1:
+                        vid = unassigned[0]
+                        method = 'NAME_FUZZY'
+                        confidence = 0.80
 
             if vid:
                 cur.execute(
                     f"""UPDATE {FLEET_VESSELS}
-                       SET mmsi = %s, match_confidence = 0.95, match_method = 'NAME_EXACT',
+                       SET mmsi = %s, match_confidence = %s, match_method = %s,
                            last_seen_at = NOW(), updated_at = NOW()
                        WHERE id = %s AND (mmsi IS NULL OR mmsi = %s)""",
-                    (mmsi, vid, mmsi),
+                    (mmsi, confidence, method, vid, mmsi),
                 )
                 self._mmsi_to_vid[mmsi] = vid
-                matched += 1
+                self._vessels[vid]['mmsi'] = mmsi
+                if method == 'NAME_FUZZY':
+                    matched_fuzzy += 1
+                else:
+                    matched_exact += 1
 
         conn.commit()
         cur.close()
-        if matched > 0:
-            logger.info('AIS→registry matched: %d vessels', matched)
+        if matched_exact or matched_fuzzy:
+            logger.info(
+                'AIS→registry matched: exact=%d, fuzzy=%d',
+                matched_exact, matched_fuzzy,
+            )
 
     def track_gear_identity(self, gear_signals: list[dict], conn) -> None:
         """어구/어망 정체성 추적.
@@ -242,13 +302,19 @@ class FleetTracker:
             if not is_trackable_parent_name(effective_parent_name):
                 continue
 
-            # 모선 매칭
+            # 모선 매칭 (EXACT → FUZZY 순)
             parent_mmsi: Optional[str] = None
             parent_vid: Optional[int] = None
             if parent_name:
                 vid = self._name_cn_map.get(parent_name)
                 if not vid:
                     vid = self._name_en_map.get(parent_name.lower())
+                if not vid:
+                    key = _normalize_vessel_name(parent_name)
+                    if key:
+                        candidates = self._name_fuzzy_map.get(key, [])
+                        if len(candidates) == 1:
+                            vid = candidates[0]
                 if vid:
                     parent_vid = vid
                     parent_mmsi = self._vessels[vid].get('mmsi')