├── database_settings.py.example
├── README.rst
├── importer.py
├── imdbapi.py
├── database.py
└── index.tpl


/database_settings.py.example:
--------------------------------------------------------------------------------
1 | CONNECTION_STRING = "sqlite:///imdbapi.db"
2 | 


--------------------------------------------------------------------------------
/README.rst:
--------------------------------------------------------------------------------
 1 | =======
 2 | imdbapi
 3 | =======
 4 | 
 5 | .. image:: https://www.codeshelter.co/static/badges/badge-flat.svg
 6 |     :target: www.codeshelter.co
 7 | 
 8 | imdbapi is a web-based script that parses the IMDB datafiles and exposes
 9 | an API to retrieve information from them.
10 | 
11 | Dependencies
12 | ------------
13 | 
14 | To run imdbapi, you need the `bottle framework <http://bottle.paws.de/>`_
15 | and the `SQLAlchemy ORM <http://www.sqlalchemy.org/>`_.
16 | 
17 | License
18 | -------
19 | 
20 | imdbapi is made available under the GPL (version 3 or later).
21 | 
22 | 


--------------------------------------------------------------------------------
/importer.py:
--------------------------------------------------------------------------------
 1 | from __future__ import with_statement
 2 | from sqlalchemy.ext.declarative import declarative_base
 3 | from sqlalchemy.orm import sessionmaker, relation, backref
 4 | from database import Show, Episode, init_db
 5 | 
 6 | import codecs
 7 | import re
 8 | import sqlalchemy
 9 | 
10 | session = init_db(transactional=True)
11 | 
12 | def import_data(filename):
13 |     """Import episode names and ratings from a file."""
14 |     regex = re.compile(""""(?P<show_name>.*?)"\s+\((?P<year>\d+)(?:|/.*?)\)\s+\{(?P<episode_name>.*?)\s?\(\#(?P<season_no>\d+)\.(?P<episode_no>\d+)\)\}""")
15 | 
16 |     with codecs.open(filename, "r", "latin-1") as ratings:
17 |         # Generate all the lines that matched.
18 |         matches = (match for match in (regex.search(line.strip()) for line in ratings) if match)
19 |         counter = 0
20 |         for match in matches:
21 |             counter += 1
22 |             if not counter % 100:
23 |                 print counter
24 |             episode = {}
25 |             for field in ["show_name", "year", "episode_name", "episode_no", "season_no"]:
26 |                 episode[field] = match.group(field)
27 | 
28 |             # If the episode has no name it is given the same name as on imdb.com for consistency.
29 |             if not episode["episode_name"]:
30 |                 episode["episode_name"] = "Episode #%s.%s" % (episode["season_no"], episode["episode_no"])
31 | 
32 |             try:
33 |                 show = session.query(Show).filter_by(name=episode["show_name"], year=episode["year"]).one()
34 |             except sqlalchemy.orm.exc.NoResultFound:
35 |                 show = Show(episode["show_name"], episode["year"])
36 |                 session.add(show)
37 | 
38 |             try:
39 |                 episode = session.query(Episode).filter_by(name=episode["episode_name"], show=show).one()
40 |             except sqlalchemy.orm.exc.NoResultFound:
41 |                 episode = Episode(show, episode["episode_name"], episode["season_no"], episode["episode_no"])
42 |                 session.add(episode)
43 | 
44 |     session.commit()
45 | 
46 | if __name__ == "__main__":
47 |     import_data("movies.list")
48 | 
49 | 


--------------------------------------------------------------------------------
/imdbapi.py:
--------------------------------------------------------------------------------
 1 | from database import Show, Episode, Stats, init_db
 2 | from bottle import route, run, request, template, response, default_app
 3 | import simplejson
 4 | import sqlalchemy
 5 | import urllib
 6 | 
 7 | session = init_db()
 8 | 
 9 | def get_data(show_name, show_year=None):
10 |     if not show_name or len(show_name) <= 3 or (show_year and not show_year.isdigit()):
11 |         return None
12 | 
13 |     show = session.query(Show).filter(Show.name.like(show_name))
14 |     if show_year:
15 |         show = show.filter(Show.year==int(show_year))
16 |     try:
17 |         single_show = show.one()
18 |     except sqlalchemy.orm.exc.NoResultFound:
19 |         return None
20 |     except sqlalchemy.orm.exc.MultipleResultsFound:
21 |         shows = show.order_by(Show.name)[:15]
22 |         show_list = [{"name": show.name, "year": show.year} for show in shows]
23 |         return {"shows": show_list}
24 | 
25 |     episodes = []
26 |     for episode in single_show.episodes:
27 |         episodes.append({"name": episode.name, "number": episode.number, "season": episode.season})
28 |     return {single_show.name: {"year": single_show.year, "episodes": episodes}}
29 | 
30 | @route('/json/')
31 | def json():
32 |     response.content_type = 'application/json'
33 |     show_name = request.GET.get("name", None)
34 |     show_year = request.GET.get("year", None)
35 |     callback = request.GET.get("callback", None)
36 |     data = simplejson.dumps(get_data(show_name, show_year))
37 |     session.close()
38 |     if callback:
39 |         data = "%s(%s)" % (callback, data)
40 |     return data
41 | 
42 | @route('/js/')
43 | def js():
44 |     show_name = request.GET.get("name", None)
45 |     show_year = request.GET.get("year", None)
46 |     callback = request.GET.get("callback", None)
47 |     data = simplejson.dumps(get_data(show_name, show_year))
48 |     session.close()
49 |     if callback:
50 |         data = "%s(%s)" % (callback, data)
51 |     return data
52 | 
53 | @route('/')
54 | def index():
55 |     return template("index")
56 | 
57 | app = application = default_app()
58 | 
59 | if __name__ == "__main__":
60 |     run(host='localhost', port=8000)
61 | 


--------------------------------------------------------------------------------
/database.py:
--------------------------------------------------------------------------------
 1 | from sqlalchemy.ext.declarative import declarative_base
 2 | from sqlalchemy.orm import sessionmaker, relation, backref
 3 | import sqlalchemy
 4 | try:
 5 |     from database_settings import CONNECTION_STRING
 6 | except ImportError:
 7 |     CONNECTION_STRING = "sqlite:///imdbapi.db"
 8 | 
 9 | Base = declarative_base()
10 | 
11 | class Show(Base):
12 |     __tablename__ = 'shows'
13 |     id = sqlalchemy.Column(sqlalchemy.Integer, primary_key=True)
14 |     name = sqlalchemy.Column(sqlalchemy.Unicode(200), index=True)
15 |     year = sqlalchemy.Column(sqlalchemy.Integer)
16 | 
17 |     def __init__(self, name, year):
18 |         self.name = name
19 |         self.year = year
20 | 
21 |     def __repr__(self):
22 |         return "<Show('%s','%s')>" % (self.name, self.year)
23 | 
24 | sqlalchemy.Index('idx_show_name_year', Show.name, Show.year)
25 | 
26 | class Episode(Base):
27 |     __tablename__ = 'episodes'
28 |     id = sqlalchemy.Column(sqlalchemy.Integer, primary_key=True)
29 |     show_id = sqlalchemy.Column(sqlalchemy.Integer, sqlalchemy.ForeignKey('shows.id'), index=True)
30 |     name = sqlalchemy.Column(sqlalchemy.Unicode(200), index=True)
31 |     season = sqlalchemy.Column(sqlalchemy.Integer)
32 |     number = sqlalchemy.Column(sqlalchemy.Integer)
33 | 
34 |     show = relation(Show, backref='episodes', order_by=id)
35 | 
36 |     def __init__(self, show, name, season, number):
37 |         self.show = show
38 |         self.name = name
39 |         self.season = season
40 |         self.number = number
41 | 
42 |     def __repr__(self):
43 |         return "<Episode('%s.%s','%s')>" % (self.season, self.number, self.name)
44 | 
45 | class Stats(Base):
46 |     __tablename__ = 'stats'
47 |     key = sqlalchemy.Column(sqlalchemy.Unicode(200), index=True, primary_key=True)
48 |     value = sqlalchemy.Column(sqlalchemy.Integer)
49 | 
50 |     def __init__(self, key, value):
51 |         self.key = key
52 |         self.value = value
53 | 
54 |     def __repr__(self):
55 |         return "<Stats('%s','%s')>" % (self.key, self.value)
56 | 
57 | 
58 | def init_db(transactional=False):
59 |     engine = sqlalchemy.create_engine(CONNECTION_STRING)
60 |     Base.metadata.create_all(engine)
61 |     Session = sessionmaker(bind=engine)
62 |     session = Session()
63 |     return session
64 | 
65 | if __name__ == "__main__":
66 |     init_db()
67 | 


--------------------------------------------------------------------------------
/index.tpl:
--------------------------------------------------------------------------------
 1 | <html>
 2 |     <head>
 3 |         <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
 4 |         <title>IMDB API</title>
 5 |         <script type="text/javascript">
 6 |           var _gaq = _gaq || [];
 7 |           _gaq.push(['_setAccount', 'UA-51110-20']);
 8 |           _gaq.push(['_trackPageview']);
 9 |           (function() {
10 |             var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
11 |             ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
12 |             var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
13 |           })();
14 |         </script>
15 |     </head>
16 |     <body>
17 |     	<h1>IMDB API</h1>
18 |     	<h3>by Stavros Korokithakis</h3>
19 |     	<p>
20 |     		Well hello. I see you want to access IMDB programatically. Well, this is your chance.
21 | 			This service is updated every day from the
22 | 			<a href="http://www.imdb.com/interfaces">IMDB data files</a> and you
23 | 			can access it with the ?name=<show name> URL parameter. For example,
24 | 			this is the episode list (in JSON) for
25 | 			<a href="/js/?name=how+i+met+your+mother">How I Met Your Mother</a>.
26 | 		</p>
27 |         <p>
28 |         This service only supports shows, it <b>does not have any movies</b>.
29 |         </p>
30 | 
31 |         Instructions:
32 | 		<ol>
33 | 		<li>Use the "name" GET parameter followed by the urlencoded show name. If there
34 |                 are multiple shows with the same name, you can pass the "year" parameter as well
35 |                 to select a year. The script can now return JSONP, you can pass the "callback"
36 |                 parameter to get the data wrapped in a javascript function whose name is the
37 |                 value of the callback parameter (<a href="/js/?name=firefly&callback=imdbapi">example</a>).</li>
38 | 		<li>The URL /js/ returns data as text/html and /json/ returns data as application/json.
39 |                 Both use exactly the same parameters.</li>
40 | 		<li>You can use wildcards by using the percent sign, for example
41 | 		<a href="/js/?name=how+i+met+%25+mother">"how i met % mother"</a> will work.</li>
42 | 		<li>If there are multiple show names, you will get a list of the 15 first
43 | 		that match your string (<a href="/js/?name=%25how%25">example</a>).</li>
44 | 		<li>If the show was not found, you will get "null".</li>
45 | 		</ol>
46 | 
47 | 		<ul>
48 |         <li>The source code for this service is
49 |         <a href="http://github.com/skorokithakis/imdbapi">available on GitHub</a>.
50 |         </li>
51 | 		<li>I have also written a script that will rename your show's video files using
52 | 		this and other services, you can find it here:
53 | 		<a href="http://github.com/skorokithakis/episode-renamer/">Episode renamer</a>.</li>
54 | 		<li>If you need to contact me, my email is can be found on my site.</li>
55 | 		<li>My site is <a href="https://www.stavros.io">stavros.io</a>.</li>
56 | 		</ul>
57 | 
58 | 		<p>
59 | 		Have fun!
60 | 		</p>
61 |         <div style="font-size: 8pt;text-align: center;color: grey;margin: 50px 0 20px 0;">a <a href="https://www.stavrosio">stavros.io</a> project</div>
62 | 	</body>
63 | </html>
64 | 


--------------------------------------------------------------------------------