Merge pull request #14 from PNU-Sinbaram/server-recommend1

Implement recommend algorithm 1 at main

server/.gitignore

@@ -291,4 +291,7 @@ $RECYCLE.BIN/
 # Windows shortcuts
 *.lnk
 
+#hide api key
+config.py
+
 # End of https://www.toptal.com/developers/gitignore/api/windows,linux,vim,django,python,visualstudiocode

server/recommend/recommend1/recommend1.py

@@ -1,12 +1,130 @@
+# -*-coding:utf-8-*-
+
+import os
+import requests
+import json
+from googletrans import Translator
+
+
 class recommend1:
     def __init__(self):
-        """Will be implemented later."""
-        pass
+        """Set radius and api key for google nearby search."""
+        self.radius = 100
+        self.key = os.environ["GOOGLE_PLACES_KEY"]
+
+    def call_api(self, lat, lng):
+        url = "https://maps.googleapis.com/maps/api/place/nearbysearch/json?"\
+                   + "location=" + str(lat) + "," + str(lng)\
+                   + "&radius=" + str(self.radius)\
+                   + "&key=" + self.key
+
+        response = requests.get(url)
+
+        if(response.status_code == 200):
+            response_body = response.text
+            return(response_body)
+        else:
+            print("Error Code:" + response.status_code)
+
+    def call_review(self, place_key):
+        url = "https://maps.googleapis.com/maps/api/place/details/json?"\
+              + "place_id=" + place_key\
+              + "&fields=name,rating,review"\
+              + "&key=" + self.key
+
+        response = requests.get(url)
+
+        if(response.status_code == 200):
+            response_body = response.text
+            return(response_body)
+        else:
+            print("Error Code:" + response.status_code)
+
+    @classmethod
+    def LocalAlignment(cls, String1, String2, GapPenalty):
+        Match, MisMatch = 15, -5
+        DP = []
+        for _ in range(len(String2)+1):
+            DP.append([0]*(len(String1)+1))
+
+        for y in enumerate(String2):
+            for x in enumerate(String1):
+                MatchTest = 0
+
+                if(String1[x] == String2[y]):
+                    MatchTest = Match
+                else:
+                    MatchTest = MisMatch
+
+                DP[y+1][x+1] = max(DP[y][x] + MatchTest,
+                                   DP[y][x+1] + GapPenalty,
+                                   DP[y+1][x] + GapPenalty, 0)
+
+        return DP
+
+    @classmethod
+    def word_tanslate(cls, word):
+        translator = Translator()
+        result = translator.translate(word, dest="en")
+        return result.text
+
+    @classmethod
+    def json2list(cls, js):
+        return json.loads(js)
+
+    def recommend(self, lat, lng, keyword, epsilon):
+        filtering_result = []
+        translated_keyword = self.word_tanslate(keyword)
+
+        places_info = []
+        api_result = self.call_api(lat, lng)
+        places = self.json2list(api_result)
+        # analyze each place infomation
+        for place in places['results']:
+            place_info = []
+
+            place_info.append(place['name'])
+            place_info.append(place['place_id'])
+
+            try:
+                if(place['business_status'] != 'OPERATIONAL'):
+                    continue
+            except KeyError:
+                continue
+
+            place_info.append(place['geometry']['location']['lat'])
+            place_info.append(place['geometry']['location']['lng'])
+
+            places_info.append(place_info)
+
+        for x, _ in enumerate(places_info):
+            place_key = places_info[x][1]
+            review_result = self.call_review(place_key)
+            reviews = self.json2list(review_result)
+            alig_max = 0
+            for y in range(len(reviews['result'])):
+                try:
+                    reviewtext = reviews['result']['reviews'][y]['text']
+                    reviewtext = reviewtext.replace(" ", "").lower()
+                    transtext = translated_keyword.replace(" ", "").lower()
+                    alig_score = self.LocalAlignment(reviewtext,
+                                                     transtext,
+                                                     (epsilon-101))
+                    if(max(map(max, alig_score)) > alig_max):
+                        alig_max = max(map(max, alig_score))
+                except Exception:
+                    break
+            places_info[x].insert(0, alig_max)
+        places_info.sort(reverse=True)
 
-    @staticmethod
-    def recommend(lat, lng, keyword, epsilon):
-        """Returns list of place recommendation using local alignment.
+        while(len(filtering_result) < 5):
+            try:
+                rank_place = places_info.pop(0)
+                filtering_result.append({"name": rank_place[1],
+                                         "lat": rank_place[3],
+                                         "long": rank_place[4],
+                                         "filtering": 1})
+            except Exception:
+                break
 
-        This function will be implemented later.
-        """
-        return([["test", 1.0, 1.0], ["test2", 2.0, 2.0]])
+        return(filtering_result)

server/recommend/views.py

@@ -15,8 +15,8 @@
 class STCViewSet(viewsets.ViewSet):
     @classmethod
     def list(self, request):
-        # recommender1 = recommend1()
-        # recommender2 = recommend2()
+        recommender1 = recommend1()
+
         UserID = request.query_params["User_ID"]
         latitude = request.query_params["lat"]
         longitude = request.query_params["long"]
@@ -30,7 +30,7 @@ def list(self, request):
                        "epsilon": epsilon}
         serializer = PlaceRequestSerializer(data=requestData)
         if serializer.is_valid():
-            requestResult1 = recommend1.recommend(
+            requestResult1 = recommender1.recommend(
                 latitude, longitude, keywords, epsilon
             )
             requestResult2 = recommend2.recommend(

server/requirements.txt

@@ -11,3 +11,4 @@ six
 urllib3
 tensorflow-gpu
 typing
+googletrans==4.0.0rc1