Merge branch 'master' into dev

Twint.py

@@ -112,7 +112,10 @@ def initialize(args):
     c.Profile_full = args.profile_full
     c.Store_pandas = args.store_pandas
     c.Pandas_type = args.pandas_type
-    c.search_name = args.search_name
+    c.Search_name = args.search_name
+    c.Index_tweets = args.index_tweets
+    c.Index_follow = args.index_follow
+    c.Index_users = args.index_users
     return c
 
 def options():
@@ -165,7 +168,10 @@ def options():
             action="store_true")
     ap.add_argument("--store-pandas", help="Save Tweets in a DataFrame (Pandas) file.")
     ap.add_argument("--pandas-type", help="Specify HDF5 or Pickle (HDF5 as default)")
-    ap.add_argument("--search_name", help="name for identify the search like -3dprinter stuff- only for mysql")
+    ap.add_argument("--search_name", help="Name for identify the search like -3dprinter stuff- only for mysql")
+    ap.add_argument("-it", "--index-tweets", help="Custom Elasticsearch Index name for Tweets.")
+    ap.add_argument("-if", "--index-follow", help="Custom Elasticsearch Index name for Follows.")
+    ap.add_argument("-iu", "--index-users", help="Custom Elasticsearch Index name for Users.")
     args = ap.parse_args()
 
     return args
@@ -180,6 +186,15 @@ def main():
     if not args.pandas_type:
         args.pandas_type = "HDF5"
 
+    if not args.index_tweets:
+        args.index_tweets = "twint"
+
+    if not args.index_follow:
+        args.index_follow = "twintGraph"
+
+    if not args.index_users:
+        args.index_users = "twintUser"
+
     c = initialize(args)
 
     if args.favorites:

twint/config.py

@@ -41,4 +41,7 @@ class Config:
     Store_pandas = False
     Pandas_type = None
     Pandas = False
-    search_name = "-" #for identify a records in mysql with the search it provides from.  it cannot be null for DB requirements. a tweet must be in several search so the PK are tweet ID and search_name
+    Search_name = "-" #for identify a records in mysql with the search it provides from.  it cannot be null for DB requirements. a tweet must be in several search so the PK are tweet ID and search_name
+    Index_tweets = None
+    Index_follow = None
+    Index_users = None

twint/elasticsearch.py

@@ -30,7 +30,7 @@ def weekday(day):
 def hour(datetime):
     return strftime("%H", localtime(datetime))
 
-def Tweet(Tweet, es, session):
+def Tweet(Tweet, config):
     day = weekday(strftime("%A", localtime(Tweet.datetime)))
 
     actions = []
@@ -41,9 +41,9 @@ def Tweet(Tweet, es, session):
     dt = "{} {}".format(Tweet.datestamp, Tweet.timestamp)
 
     j_data = {
-            "_index": "twint",
+            "_index": config.Index_tweets,
             "_type": "items",
-            "_id": Tweet.id + "_raw_" + str(session),
+            "_id": Tweet.id + "_raw_" + config.Essid,
             "_source": {
                 "id": Tweet.id,
                 "date": dt,
@@ -58,16 +58,16 @@ def Tweet(Tweet, es, session):
                 "link": Tweet.link,
                 "retweet": Tweet.retweet,
                 "user_rt": Tweet.user_rt,
-                "essid": str(session)
+                "essid": config.Essid
                 }
             }
     actions.append(j_data)
 
     for l in range(int(Tweet.likes)):
         j_data = {
-                "_index": "twint",
+                "_index": config.Index_tweets,
                 "_type": "items",
-                "_id": Tweet.id + "_likes_" + str(nLikes) + "_" + str(session),
+                "_id": Tweet.id + "_likes_" + str(nLikes) + "_" + config.Essid,
                 "_source": {
                     "id": Tweet.id,
                     "date": dt,
@@ -83,7 +83,7 @@ def Tweet(Tweet, es, session):
                     "link": Tweet.link,
                     "retweet": Tweet.retweet,
                     "user_rt": Tweet.user_rt,
-                    "essid": str(session)
+                    "essid": config.Essid
                     }
                 }
         actions.append(j_data)
@@ -91,9 +91,9 @@ def Tweet(Tweet, es, session):
 
     for rep in range(int(Tweet.replies)):
         j_data = {
-                "_index": "twint",
+                "_index": config.Index_tweets,
                 "_type": "items",
-                "_id": Tweet.id + "_replies_" + str(nReplies) + "_" + str(session),
+                "_id": Tweet.id + "_replies_" + str(nReplies) + "_" + config.Essid,
                 "_source": {
                     "id": Tweet.id,
                     "date": dt,
@@ -109,7 +109,7 @@ def Tweet(Tweet, es, session):
                     "link": Tweet.link,
                     "retweet": Tweet.retweet,
                     "user_rt": Tweet.user_rt,
-                    "essid": str(session)
+                    "essid": config.Essid
                     }
                 }
         actions.append(j_data)
@@ -117,9 +117,9 @@ def Tweet(Tweet, es, session):
 
     for ret in range(int(Tweet.retweets)):
         j_data = {
-                "_index": "twint",
+                "_index": config.Index_tweets,
                 "_type": "items",
-                "_id": Tweet.id + "_retweets_" + str(nRetweets) + "_" + str(session),
+                "_id": Tweet.id + "_retweets_" + str(nRetweets) + "_" + config.Essid,
                 "_source": {
                     "id": Tweet.id,
                     "date": dt,
@@ -135,44 +135,44 @@ def Tweet(Tweet, es, session):
                     "link": Tweet.link,
                     "retweet": Tweet.retweet,
                     "user_rt": Tweet.user_rt,
-                    "essid": str(session)
+                    "essid": config.Essid
                     }
                 }
         actions.append(j_data)
         nRetweets += 1
 
-    es = Elasticsearch(es)
+    es = Elasticsearch(config.Elasticsearch)
     with nostdout():
         helpers.bulk(es, actions, chunk_size=2000, request_timeout=200)
     actions = []
 
-def Follow(es, user, follow, session):
+def Follow(user, config):
     actions = []
 
     j_data = {
-            "_index": "twintGraph",
+            "_index": config.Index_follow,
             "_type": "items",
-            "_id": user + "_" + follow + "_" + str(session),
+            "_id": user + "_" + config.Username + "_" + config.Essid,
             "_source": {
                 "user": user,
-                "follow": follow,
-                "essid": str(session)
+                "follow": config.Username,
+                "essid": config.Essid
                 }
             }
     actions.append(j_data)
 
-    es = Elasticsearch(es)
+    es = Elasticsearch(config.Elasticsearch)
     with nostdout():
         helpers.bulk(es, actions, chunk_size=2000, request_timeout=200)
     actions = []
 
-def UserProfile(es, user, follow, session):
+def UserProfile(user, config):
     actions = []
 
     j_data = {
-            "_index": "twintUser",
+            "_index": config.Index_users,
             "_type": "items",
-            "_id": user.id + "_" + user.join_date + "_" + user.join_time + "_" + str(session),
+            "_id": user.id + "_" + user.join_date + "_" + user.join_time + "_" + config.Essid,
             "_source": {
                 "id": user.id,
                 "name": user.name,
@@ -191,12 +191,12 @@ def UserProfile(es, user, follow, session):
                 "private": user.is_private,
                 "verified": user.is_verified,
                 "avatar": user.avatar,
-                "session": str(session)
+                "session": config.Essid
                 }
             }
     actions.append(j_data)
 
-    es = Elasticsearch(es)
+    es = Elasticsearch(config.Elasticsearch)
     with nostdout():
         helpers.bulk(es, actions, chunk_size=2000, request_timeout=200)
     actions = []

twint/output.py

@@ -56,7 +56,7 @@ async def Tweets(tw, location, config, conn):
             elif config.Database:
                 db.tweets(conn, tweet, config)
             if config.Elasticsearch:
-                elasticsearch.Tweet(tweet, config.Elasticsearch, config.Essid)
+                elasticsearch.Tweet(tweet, config)
             _output(tweet, output, config)
 
 async def Users(u, config, conn):
@@ -73,8 +73,7 @@ async def Users(u, config, conn):
         _save_time = user.join_time
         user.join_date = str(datetime.strptime(user.join_date, "%d %b %Y")).split()[0]
         user.join_time = str(datetime.strptime(user.join_time, "%I:%M %p")).split()[1]
-        elasticsearch.UserProfile(config.Elasticsearch, user,
-                config.Username, config.Essid)
+        elasticsearch.UserProfile(user, config)
         user.join_date = _save_date
         user.join_time = _save_time
 
@@ -87,7 +86,6 @@ async def Username(username, config, conn):
         db.follow(conn, config.Username, config.Followers, username)
 
     if config.Elasticsearch:
-        elasticsearch.Follow(config.Elasticsearch, username,
-                config.Username, config.Essid)
+        elasticsearch.Follow(username, config)
 
     _output(username, username, config)