1 files changed, 177 insertions, 0 deletions
diff --git a/users.py b/users.py
new file mode 100755
index 0000000..e61d3c2
--- /dev/null
+++ b/users.py
@@ -0,0 +1,177 @@
+#!/usr/bin/python3
+# -*- coding: utf-8 -*-
+
+
+import tweepy
+from config import accounts
+from db import open_db, LikeMemory, TrackMemory
+from taste import analyse
+import os
+import pickle
+from random import shuffle
+
+
+CACHE_FILENAME = 'ids.cache'
+
+
+class UsersListener():
+    """A listener handles tweets are the received from the stream."""
+
+    def __init__(self, api):
+        """Build the Python object."""
+
+        super().__init__()
+
+        self._api = api
+
+        open_db()
+
+        self._memory = LikeMemory(api)
+        self._tracker = TrackMemory()
+
+        self._compute_ids_to_follow(True)
+
+
+    def _compute_ids_to_follow(self, cached):
+        """Get the list of accounts to track."""
+
+        if not os.path.isfile(CACHE_FILENAME):
+            cached = False
+
+        if cached:
+
+            ids = pickle.load(open(CACHE_FILENAME, 'rb'))
+
+            print('[i] Reloaded %u accounts' % len(ids))
+
+        else:
+
+            ids = []
+
+            for master in accounts.split(' '):
+
+                count = 0
+
+                try:
+
+                    for page in tweepy.Cursor(self._api.followers_ids, screen_name=master).pages():
+
+                        count += len(page)
+                        ids.extend(page)
+
+                    print('[i] Got %u accounts following %s' % (count, master))
+
+                except:
+
+                    print('[!] Error while receiving followers for %s...' % master)
+
+
+            pickle.dump(ids, open(CACHE_FILENAME, 'wb'))
+
+            print('[i] Loaded %u accounts' % len(ids))
+
+            ids = list(set(ids))
+
+            print('[i] Kept %u accounts' % len(ids))
+
+        # Remove all account natively followed
+
+        already = []
+
+        for page in tweepy.Cursor(self._api.followers_ids, screen_name=self._api.me().name).pages():
+            already.extend(page)
+
+        print('[i] I am followed by %u accounts' % len(already))
+
+        self._ids = [ x for x in ids if x not in already ]
+
+        shuffle(self._ids)
+
+        print('[i] Tracking %u accounts...' % len(self._ids))
+
+
+    def start(self, auth):
+        """Start the listener."""
+
+        while True:
+
+            for uid in self._ids:
+
+                since = self._tracker.get_last_seen_for(uid)
+
+                last = []
+
+                try:
+
+                    last = self._api.user_timeline(uid, since)
+
+                except tweepy.error.TweepError as e:
+
+                    # Private account
+                    # tweepy.error.TweepError: Not authorized.
+                    if e.response.status_code == 401:
+                        pass
+
+                    # Nothing new!
+                    # tweepy.error.TweepError: [{'message': 'Sorry, that page does not exist.', 'code': 34}]
+                    elif e.response.status_code == 404:
+                        pass
+
+                    else:
+                        print(e, e.response.status_code)
+                        assert(False)
+
+                first = None
+
+                for status in last:
+
+                    sid = status.id
+                    uid = status.author.id
+                    username = status.author.screen_name
+
+                    while hasattr(status, 'retweeted_status'):
+                        status = status.retweeted_status
+
+                    analyse(sid, username, status.text, self._api, self._memory)
+
+                    if first is None:
+                        first = uid, username, sid
+
+                if not(first is None):
+                    uid, username, sid = first
+                    self._tracker.set_last_seen_for(uid, username, sid)
+
+
+
+def listen_to_users(auth, api):
+    """Track all tweets written by users."""
+
+
+    data = api.rate_limit_status()
+
+    for c in data['resources'].keys():
+
+        print('%s' % c)
+
+        category = data['resources'][c]
+
+        for p in category.keys():
+
+            props = category[p]
+            changed = props['remaining'] != props['limit']
+
+            print(' %s %s: %d / %d' % ('!!' if changed else '  ', p, props['remaining'], props['limit']))
+
+
+
+
+
+
+    if True:
+
+        listener = UsersListener(api)
+
+        ####listener.start(auth)
+
+        #stream = Stream(auth, listener)
+        #stream.filter(follow=new)