优化科室的匹配

docker-compose.yml

@@ -1,6 +1,6 @@
 x-env:
   &template
-  image: fcb_photo_review:1.12.9
+  image: fcb_photo_review:1.12.10
   restart: always
 
 services:

photo_review/__init__.py

@@ -75,7 +75,7 @@ HOSPITAL_ALIAS = {
     "漂水": ["溧水"],
 }
 DEPARTMENT_ALIAS = {
-    "耳鼻喉": ["耳鼻咽喉"]
+    "耳鼻喉": ["耳鼻咽喉"],
 }
 
 """
@@ -84,10 +84,16 @@ DEPARTMENT_ALIAS = {
 # 默认会过滤单字
 HOSPITAL_FILTER = ["医院", "人民", "第一", "第二", "第三", "大学", "附属"]
 
+DEPARTMENT_FILTER = ["医", "病", "病区", "伤"]
+
 """
 分词配置
 """
 jieba.suggest_freq(('肿瘤', '医院'), True)
+jieba.suggest_freq(('心血管', '病'), True)
+jieba.suggest_freq(('骨', '伤'), True)
+jieba.suggest_freq(('感染', '性'), True)
+jieba.suggest_freq(('胆', '道'), True)
 
 """
 模型配置

photo_review/photo_review.py

@@ -17,7 +17,7 @@ from db.mysql import BdYljg, BdYlks, ZxIeResult, ZxIeCost, ZxIeDischarge, ZxIeSe
 from photo_review import PATIENT_NAME, ADMISSION_DATE, DISCHARGE_DATE, MEDICAL_EXPENSES, PERSONAL_CASH_PAYMENT, \
     PERSONAL_ACCOUNT_PAYMENT, PERSONAL_FUNDED_AMOUNT, MEDICAL_INSURANCE_TYPE, HOSPITAL, DEPARTMENT, DOCTOR, \
     ADMISSION_ID, SETTLEMENT_ID, AGE, OCR, SETTLEMENT_IE, DISCHARGE_IE, COST_IE, PHHD_BATCH_SIZE, SLEEP_MINUTES, \
-    UPPERCASE_MEDICAL_EXPENSES, HOSTNAME, HOSPITAL_ALIAS, HOSPITAL_FILTER, DEPARTMENT_ALIAS
+    UPPERCASE_MEDICAL_EXPENSES, HOSTNAME, HOSPITAL_ALIAS, HOSPITAL_FILTER, DEPARTMENT_ALIAS, DEPARTMENT_FILTER
 from ucloud import ufile
 from util import image_util, util
 from util.data_util import handle_date, handle_decimal, parse_department, handle_name, \
@@ -206,7 +206,7 @@ def search_hospital(hospital):
             result.append(result2)
         return result
 
-    cut_list = jieba.lcut(hospital)
+    cut_list = jieba.lcut(hospital, HMM=False)
     session = MysqlSession()
     yljg = session.query(BdYljg.pk_yljg, BdYljg.name).filter(BdYljg.name.like(f"%{'%'.join(cut_list)}%")).all()
     if not yljg:
@@ -222,8 +222,15 @@ def search_hospital(hospital):
 
 
 def search_department(department):
-    cut_list = jieba.lcut(department)
+    def _filter_search_keywords(keywords):
+        keywords = [x for x in keywords if x not in DEPARTMENT_FILTER]
+        return keywords
+
+    cut_list = jieba.lcut(department, HMM=False)
     session = MysqlSession()
+    cut_list = _filter_search_keywords(cut_list)
+    if not cut_list:
+        return None
     ylks = session.query(BdYlks.pk_ylks, BdYlks.name).filter(BdYlks.name.like(f"%{'%'.join(cut_list)}%")).all()
     if not ylks:
         filter_keywords = cut_list
@@ -233,7 +240,11 @@ def search_department(department):
                 break
     session.close()
     ylks = {row.pk_ylks: row.name for row in ylks}
-    best_match = process.extractOne(department, ylks, scorer=fuzz.partial_token_set_ratio)
+    best_match = process.extractOne(department, ylks, scorer=fuzz.token_ratio)
+    if best_match and best_match[0] in ["内科", "外科"]:
+        # 降低内科、外科的优先级
+        best_match = list(best_match)
+        best_match[1] -= 100
     return best_match