fix code, implement twint stuff

src/catchup.py

@@ -5,18 +5,25 @@
 # We should post, at the fastest, one tweet per minute.
 
 import traceback
+import datetime
 import sys
 import os
 import asyncio
 
 import twint
-import tweepy
 
 from util import *
 from talent_lists import *
 from twapi import TwAPI
 import talenttweet as tt
 
+def write_user_date(user_id, file, date_str = None, error = False):
+    if date_str is None:
+        date_str = util.datetime_to_tdate(datetime.datetime.now())
+
+    file.write(f'# {user_id} {date_str if not error else "-1"}\n')
+    pass
+
 def get_queue_file():
     return f'{util.get_project_dir()}/queue.txt'
 
@@ -25,7 +32,7 @@ def get_local_queue():
     pass
 
 ## Returns the ID of all tweets (up to limit) from a user ID.
-def get_user_tweets(id, limit=None):
+def get_user_tweets(id, since_date='', limit=None):
     tweets = list()
     c = twint.Config()
     c.User_id = id
@@ -33,6 +40,7 @@ def get_user_tweets(id, limit=None):
     c.Store_object = True
     c.Store_object_tweets_list = tweets
     c.Hide_output = True
+    c.Since = since_date
     
     user_str = f'{util.get_username(id)}'
     print(f'Scraping tweets from {user_str}...')
@@ -47,23 +55,24 @@ def get_user_tweets(id, limit=None):
 # If queue.txt doesn't exist, creates and populates it.
 # Returns a list of sorted and filtered TalentTweets (should
 # be equivalent to queue.txt)
-async def get_cross_talent_tweets(queue_file):
-    finished_users = set()
+async def get_cross_talent_tweets(queue_path):
+    finished_user_tdates = dict()
     ttweets_dict = dict()
 
     # Populate structures with existing data from queue.txt
     try:
         print('Processing existing data in queue.txt...')
-        with open(queue_file, 'r') as f:
+        with open(queue_path, 'r') as f:
             # Check for finished and incomplete accounts
-            # LINE FORMAT: "# {user_id} {status_num}"
+            # LINE FORMAT: "# {user_id} {status_num} (TODO: use date of retrival YYYY-MM-DD)
             for line in f:
                 tokens = line.split()
                 if len(tokens) != 3 or tokens[0][0] != '#':
                     # reached end of accounts list
                     break
-                if tokens[2] == 0:
-                    finished_users.add(tokens[1])
+
+                if tokens[2] != '-1':
+                    finished_user_tdates[int(tokens[1])] = tokens[2]
             
             # Add existing serialized TalentTweets into ttweets
             for line in f:
@@ -75,7 +84,34 @@ async def get_cross_talent_tweets(queue_file):
     except FileNotFoundError:
         print('Couldn\'t find queue.txt.')
 
-    # TODO: implement ordered cross-company ttweets dict creation using twint
+    # Pull tweets from twint
+    with open(queue_path, 'w') as f:
+        # for talent_id in talent_lists.talents:
+        for talent_id in talent_lists.test_talents:
+            print('using test_talents')
+            if talent_id not in finished_user_tdates or \
+                finished_user_tdates[talent_id] != util.datetime_to_tdate(datetime.datetime.today()):
+                try:
+                    tweets = get_user_tweets(talent_id, since_date=finished_user_tdates.get(talent_id, None))
+                    for tweet in tweets:
+                        ttweet = await tt.TalentTweet.create_from_twint_tweet(tweet)
+                        if ttweet.is_cross_company():
+                            ttweets_dict[ttweet.tweet_id] = ttweet
+                except:
+                    print('Error occurred processing tweet data. Traceback:')
+                    print(traceback.format_exc())
+                    write_user_date(user_id=talent_id, file=f, error=True)
+                else:
+                    write_user_date(user_id=talent_id, file=f)
+            else:
+                print(f'Skipping already completed {util.get_username(talent_id)}')
+                write_user_date(user_id=talent_id, file=f, date_str=finished_user_tdates[talent_id])
+        f.write('\n')
+        ttweets_dict = dict(sorted(ttweets_dict.items()))
+        for ttweet in ttweets_dict.values():
+            f.write(f'{ttweet.serialize()}\n')
+    
+    return ttweets_dict
 
 def process_queue(file):
     print('TODO: implement process_queue')
@@ -101,5 +137,4 @@ async def run():
 
     queue_path = get_queue_file()
     ttweet_dict = await get_cross_talent_tweets(queue_path)
-    for ttweet in ttweet_dict.values():
-        print(ttweet)
+    print(f'got {len(ttweet_dict)} tweets')

src/talent_lists.py

@@ -6,6 +6,8 @@ niji_en = dict()
 niji_exid = dict()
 talents = dict()
 
+test_talents = dict()
+
 def __create_dict(file, _dict):
     print(f'Initializing talents\' account list from {file}...')
     global talents
@@ -23,6 +25,7 @@ def init():
     global holo_id
     global niji_en
     global niji_exid
+    global test_talents
 
     # holoEN
     __create_dict(f'{util.get_project_dir()}/lists/holoen.txt', holo_en)
@@ -33,3 +36,7 @@ def init():
     # nijiexID
     __create_dict(f'{util.get_project_dir()}/lists/nijiexid.txt', niji_exid)
 
+    test_talents = {
+        1390637197167038464: 'PomuRainpuff'
+    }
+

src/talenttweet.py

@@ -43,24 +43,30 @@ class TalentTweet:
         )
     
     @staticmethod
-    def create_from_twint_tweet(tweet):
+    async def create_from_twint_tweet(tweet):
         # qrt
-        if tweet.quote_url != '':
-            return TalentTweet(tweet_id=tweet.id)
+        # -- COMMENTED OUT FOR TESTING PURPOSES --
+        # TODO: uncomment
+        # if tweet.quote_url != '':
+        #     api_ttweet = await TalentTweet.create_from_id(tweet.id)
+        #     return api_ttweet
 
         # MRQ (Q is guaranteed to be None)
         mentions = set()
         reply_to = None
         
         # reply_to/mentions
-        is_reply = tweet.id == int(tweet.conversation_id)
-        if is_reply:
-            reply_to = tweet.reply_to[0]
-            mentions = set(tweet.reply_to[1:])
-        mentions.add(*tweet.mentions)
+        is_reply = tweet.id != int(tweet.conversation_id)
+        mentions = set([x['id'] for x in tweet.mentions])
+        if is_reply and len(tweet.reply_to) > 0: # FIXME: QRT = is_reply and len(tweet.reply_to) == 0?
+            reply_to = tweet.reply_to[0]['id']
+            reply_others = [x['id'] for x in tweet.reply_to[1:]]
+            mentions.update(reply_others)
+            try: mentions.remove(reply_to)
+            except: pass
 
-        datetime = datetime.strptime(tweet.datetime, '%Y-%m-%d %H:%M:%S %Z')
-        return TalentTweet(tweet_id=tweet.id, author_id=tweet.user_id, date_time=datetime, mrq=(mentions, reply_to, None))
+        date_time = datetime.strptime(tweet.datetime, '%Y-%m-%d %H:%M:%S %Z')
+        return TalentTweet(tweet_id=tweet.id, author_id=tweet.user_id, date_time=date_time, mrq=(mentions, reply_to, None))
 
 
     @staticmethod
@@ -79,9 +85,9 @@ class TalentTweet:
     def __init__(self, tweet_id: int, author_id: int,date_time: datetime, mrq: tuple):
         self.tweet_id, self.author_id = tweet_id, author_id
         self.date_time = date_time
-        self.mentions = mrq[0]
-        self.reply_to = mrq[1]
-        self.quote_retweeted = mrq[2]
+        self.mentions = tuple(int(x) for x in mrq[0])
+        self.reply_to = int(mrq[1]) if mrq[1] is not None else None
+        self.quote_retweeted = int(mrq[2]) if mrq[2] is not None else None
 
         # all users involved, except for the author
         self.all_parties = {self.reply_to, self.quote_retweeted}

src/util.py

@@ -1,5 +1,6 @@
 ## Shared utility functions.
 
+import datetime
 import os
 
 import twint
@@ -16,6 +17,12 @@ def get_project_dir():
 def clamp(n, smallest, largest):
     return max(smallest, min(n, largest))
 
+def datetime_to_tdate(date_time: datetime.datetime):
+    return date_time.strftime("%Y-%m-%d")
+
+def tdate_to_datetime(tdate: str):
+    return datetime.datetime.strptime("%Y-%m-%d")
+
 async def create_ttweet_image(ttweet):
     tc = TweetCapture()
     filename = 'img.png'