Multicatches and improvements on /botprobability

app/models/models.py

@@ -110,16 +110,20 @@ def predict(self, users_data, timeline_data, path_input_model="app/models/pegabo
         return predicted_proba
 
     def botProbability(self, handle, twitterTimeline, twitterUserData):
+        analise = 0
         try:
             df_timeline = pd.DataFrame.from_dict(twitterTimeline)
-            df_user_data = pd.DataFrame.from_dict(twitterUserData)
+            df_user_data = pd.DataFrame([twitterUserData])
             analise = self.predict(df_user_data, df_timeline)
             self.total = round(analise[0][1]*100, 2)
+
         except:
-            self.total = -1
+             raise Exception("Problem(s) probably in predict function")    
+
 
         return edict({
             'pegabot_version': 'version-1.0',
             'handle': handle,
-            'total': self.total
+            'total': self.total,
+            'analysis': analise.tolist()
         })

app/models/prepare_data.py

@@ -26,6 +26,9 @@ def prepare_data(self, df_users, df_timeline, path_trendtopics_data = "app/model
         df_users['É bot?'] = ''
 
         #Extrai as informações de retweet
+        if not('tweet_is_retweet' in df_timeline.columns and 'tweet_text' in df_timeline.columns):
+            raise Exception("Problems on tweets")
+
         df_timeline['retweet_tratado'] = df_timeline['tweet_is_retweet'].apply(lambda x: "sim" if (x == 'True' or x == True) else "não")
         df_timeline['tweet_com_rt_tratado'] = df_timeline['tweet_text'].apply(lambda x: "sim" if x.find("RT @") != -1 else "não" )
 
@@ -91,7 +94,7 @@ def reune_rt(retweet,rt):
         df = df_result_merge
 
         #Monta o conjunto de treinamento
-        feature_cols = ['followers_count', 'friends_count', 'Tempo mediano', 'Tempo menor']
+        feature_cols = ['twitter_followers_count', 'twitter_friends_count', 'Tempo mediano', 'Tempo menor']
         x = df[feature_cols]
 
         qtd_hashtags = df['tweet_hashtags'].apply(lambda x: len(x.replace("[","").replace("]","").replace(", \'","$").split("$")))
@@ -105,7 +108,7 @@ def reune_rt(retweet,rt):
 
         #O tamanho do nome e do login
         tam_username = df['handle'].apply(lambda x: len(str(x)))
-        tam_nome = df['name'].apply(lambda x: len(str(x)))
+        tam_nome = df['twitter_user_name'].apply(lambda x: len(str(x)))
         x['Tamanho do username'] = np.array(list(tam_username))
         x['Tamanho do nome'] = np.array(list(tam_nome))
 

app/routes/routes.py

@@ -2,7 +2,7 @@
 from flask import jsonify, request
 from app.models.models import Analises, AnaliseSchema
 from app.services.botometer_service import BotometerService
-
+from concurrent.futures import ThreadPoolExecutor
 
 @app.get("/catch")
 def catch():
@@ -31,3 +31,31 @@ def complete():
 def feedback():
     return jsonify("feedback")
 
+@app.get('/multicatches')
+def multicatches():
+    handle = str(request.args.get('profiles'))
+    users = handle.split(',')
+
+    results = list()
+    for user in users:
+        botometer_service = BotometerService()
+        response = botometer_service.catch(user)
+        results.append(response)
+
+    return jsonify(results), 200
+
+@app.get('/multicatchesparallel')
+def multicatches2():
+    handle = str(request.args.get('profiles'))
+    users = handle.split(',')
+    results = list()
+
+    def getResult(username):
+        botometer_service = BotometerService()
+        response = botometer_service.catch(username)
+        results.append(response)
+
+    with ThreadPoolExecutor(max_workers=10) as pool:
+        pool.map(getResult, users)
+
+    return jsonify(results), 200

app/services/botometer_service.py

@@ -105,7 +105,10 @@ def check_cache_validity(self, analise, handle):
                 db.session.add(analise)
                 db.session.commit()
 
-    def botProbability(self, handle, user, timeline):
+    def botProbability(self, handle):
         p = BotProbability()
+        user = self.findUserAnalisisByHandle(handle=handle)
+        response = self.twitter_handler.findByHandle(handle=handle)
+        timeline = self.twitter_handler.getUserTimeline(response.twitter_id)
         response = p.botProbability(handle=handle, twitterTimeline=timeline, twitterUserData=user)
         return response