Saved a first version of HTT.

3 files changed, 219 insertions, 0 deletions

diff --git a/config.py b/config.py
new file mode 100644
index 0000000..b1a3471
--- /dev/null
+++ b/config.py
@@ -0,0 +1,12 @@
+#!/usr/bin/python3
+# -*- coding: utf-8 -*-
+
+
+# List of space-separated hashtags to follow, for instance #python #bot
+hashtags = '#python #bot'
+
+# Keywords to find in Tweets we want to highlight
+white_kwds = 'you got the idea'
+
+# Age of old Tweets to get purged in days
+max_age = 14
diff --git a/db.py b/db.py
new file mode 100644
index 0000000..ad497a0
--- /dev/null
+++ b/db.py
@@ -0,0 +1,84 @@
+#!/usr/bin/python3
+# -*- coding: utf-8 -*-
+
+
+from config import max_age
+import sqlite3
+import time
+import tweepy
+
+
+class LikeMemory():
+    """Track all liked Tweets."""
+
+    def __init__(self, api):
+        """Build the Python object."""
+
+        self._api = api
+
+        self._db = sqlite3.connect('HTT.db', detect_types=sqlite3.PARSE_DECLTYPES)
+
+        sqlite3.register_adapter(bool, int)
+        sqlite3.register_converter("BOOLEAN", lambda v: bool(int(v)))
+
+        sql = '''
+            CREATE TABLE IF NOT EXISTS LikedTweets(
+                sid INTEGER PRIMARY KEY,
+                username TEXT,
+                timestamp INTEGER,
+                purged BOOLEAN
+            )
+        '''
+
+        cursor = self._db.cursor()
+        cursor.execute(sql)
+        self._db.commit()
+
+
+    def save_liked_status(self, sid, username):
+        """Remember a given liked status."""
+
+        timestamp = int(time.time())
+
+        values = (sid, username, timestamp, False)
+
+        cursor = self._db.cursor()
+        cursor.execute('INSERT INTO LikedTweets VALUES (?, ?, ?, ?)', values)
+        self._db.commit()
+
+        print(timestamp)
+
+
+    def purge_old_status(self):
+        """Purge old seen statuses."""
+
+        timestamp = int(time.time()) - max_age * 24 * 60 * 60
+
+        values = (timestamp, False)
+
+        cursor = self._db.cursor()
+        cursor.execute('SELECT sid FROM LikedTweets WHERE timestamp < ? AND purged = ?', values)
+
+        rows = cursor.fetchall()
+
+        for row in rows:
+
+            sid = row[0]
+
+            try:
+
+                self._api.destroy_favorite(sid)
+
+            # tweepy.error.TweepError: [{'code': 144, 'message': 'No status found with that ID.'}]
+            except tweepy.error.TweepError as err:
+
+                pass
+
+            values = (True, sid)
+
+            cursor = self._db.cursor()
+            cursor.execute('UPDATE LikedTweets SET purged = ? WHERE sid = ?', values)
+
+        self._db.commit()
+
+        print('Purged %d liked Tweet%s!' % (len(rows), '' if len(rows) <= 1 else 's'))
diff --git a/htt.py b/htt.py
new file mode 100755
index 0000000..cc82531
--- /dev/null
+++ b/htt.py
@@ -0,0 +1,123 @@
+#!/usr/bin/python3
+# -*- coding: utf-8 -*-
+
+
+import tweepy
+from tweepy import OAuthHandler
+from tweepy import Stream
+from tweepy.streaming import StreamListener
+from auth import *
+from config import hashtags, white_kwds
+from db import LikeMemory
+import json
+import sys
+
+
+class StdOutListener(StreamListener):
+    """A listener handles tweets are the received from the stream."""
+
+    def __init__(self, api, memory):
+        """Build the Python object."""
+
+        super().__init__()
+
+        self._api = api
+        self._memory = memory
+
+        self._white = [ s.lower() for s in white_kwds.split(' ') ]
+
+
+    def get_status_info(self, data):
+        """Parse status data to get information about its author and content."""
+
+        # Do not rely on https://dev.twitter.com/overview/api/tweets
+        # as the specs seem outdated...
+
+        sid = data['id']
+        username = data['user']['screen_name']
+
+        if 'extended_tweet' in data:
+            content = data['extended_tweet']['full_text']
+        else:
+            content = data['text']
+
+        content = content.replace('\n', '')
+
+        return sid, username, content
+
+
+    def on_data(self, data):
+        """Receive Tweets matching the given hashtags."""
+
+        decoded = json.loads(data)
+
+        if 'retweeted_status' in decoded:
+            sid, username, content = self.get_status_info(decoded['retweeted_status'])
+        else:
+            sid, username, content = self.get_status_info(decoded)
+
+        like = False
+
+        words = content.split(' ')
+
+        for kwd in self._white:
+
+            for w in words:
+                if w.lower() == kwd:
+                    like = True
+                    break
+
+            if like:
+                break
+
+        if like:
+
+            try:
+
+                self._api.create_favorite(sid)
+
+                self._memory.save_liked_status(sid, username)
+
+                print('@%s: "%s" (id=%d)' % (username, content, sid))
+                print(' -> https://twitter.com/%s/status/%d' % (username, sid))
+
+            except tweepy.error.TweepError:
+
+                pass
+
+        else:
+
+            print('Reject "%s"' % content)
+
+        return True
+
+
+    def on_error(self, status):
+        """Handle errors."""
+
+        print('Error:', status)
+
+        if status_code == 420:
+            #returning False in on_data disconnects the stream
+            return False
+
+
+if __name__ == '__main__':
+    """Start of the script."""
+
+    auth = OAuthHandler(CONSUMER_KEY, CONSUMER_SECRET)
+    auth.set_access_token(ACCESS_KEY, ACCESS_SECRET)
+
+    api = tweepy.API(auth)
+    memory = LikeMemory(api)
+
+    if len(sys.argv) > 1 and sys.argv[1] == '--purge':
+
+        memory.purge_old_status()
+
+    else:
+
+        listener = StdOutListener(api, memory)
+
+        stream = Stream(auth, listener)
+        stream.filter(track=hashtags.split(' '))