[KAN-87] naver geo api를 활용하여 좌표 데이터 추가 (#20)

csv-to-es.py

@@ -41,6 +41,8 @@
                 "image_url": {"type": "text"},
                 "category": {"type": "text", "analyzer": "korean"},
                 "discount_content": {"type": "text", "analyzer": "korean"},
+                "longitude": {"type": "float"},
+                "latitude": {"type": "float"},
                 "menus": {
                     "type": "nested",
                     "properties": {
@@ -86,8 +88,18 @@
     else:
         discount_content = None
 
+    if pd.notna(row['longitude']):
+        longitude = float(row['longitude'])
+    else:
+        longitude = None
+
+    if pd.notna(row['latitude']):
+        latitude = float(row['latitude'])
+    else:
+        latitude = None
+
     print(row['name'], row['category'], row['review_count'], row['address'], rating, number, restaurant_image_url,
-          menus)
+          menus, longitude, latitude)
     data = {
         "id": row['id'],
         "name": row['name'],
@@ -99,6 +111,8 @@
         "image_url": restaurant_image_url,
         "category": row['custom_category'],
         "discount_content": discount_content,
+        "longitude": longitude,
+        "latitude": latitude,
         "menus": menus,
     }
     if data.get("discount_content") is None:

csv-to-geoapi.py

@@ -0,0 +1,41 @@
+import os
+
+import pandas as pd
+import requests
+
+# 네이버 지오코딩 API 키 설정
+client_id = os.getenv('NAVER_CLIENT_ID')
+client_secret = os.getenv('NAVER_CLIENT_SECRET')
+
+df = pd.read_csv('restaurants.csv')
+
+
+def get_coordinates(address):
+  url = f"https://naveropenapi.apigw.ntruss.com/map-geocode/v2/geocode?query={address}"
+  headers = {
+    "X-NCP-APIGW-API-KEY-ID": client_id,
+    "X-NCP-APIGW-API-KEY": client_secret
+  }
+  response = requests.get(url, headers=headers)
+  if response.status_code == 200:
+    data = response.json()
+    if data['addresses']:
+      x = data['addresses'][0]['x']
+      y = data['addresses'][0]['y']
+      return x, y
+  else:
+    print("Error Code:", response.status_code)
+    print(response.text)
+  return None, None
+
+
+df['longitude'] = ''
+df['latitude'] = ''
+
+for idx, row in df.iterrows():
+  address = row['address']
+  x, y = get_coordinates(address)
+  df.at[idx, 'longitude'] = x
+  df.at[idx, 'latitude'] = y
+
+df.to_csv('restaurants.csv', index=False)

mysql_batch/create_table.py

@@ -1,20 +1,10 @@
-from create_table_list import *
 from delete_table_list import *
 
 def create_table(cursor):
 
-    # table deletion query
     cursor.execute(delete_table_restaurants)
     cursor.execute(delete_table_categories)
     cursor.execute(delete_table_restaurant_categories)
     cursor.execute(delete_table_operating_infos)
     cursor.execute(delete_table_menus)
-
-    # table creation query
-    cursor.execute(create_table_restaurants)
-    cursor.execute(create_table_restaurant_likes)
-    cursor.execute(create_table_categories)
-    cursor.execute(create_table_restaurant_categories)
-    cursor.execute(create_table_operating_infos)
-    cursor.execute(create_table_menus)
 

mysql_batch/insert.py

@@ -11,10 +11,12 @@ def insert_into_restaurants(cursor, restaurant):
       `rating_avg`, 
       `representative_image_url`, 
       `view_count`,
-      `discount_content`
+      `discount_content`,
+      `longitude`,
+      `latitude`
     ) 
     VALUES (
-      %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s
+      %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s
     );
     """
 
@@ -29,7 +31,9 @@ def insert_into_restaurants(cursor, restaurant):
     0,
     restaurant['image_url'],
     0,
-    restaurant['discount_content']
+    restaurant['discount_content'],
+    restaurant['longitude'],
+    restaurant['latitude']
   ))