This is another post in my TV Shows Trivia game serie. In our last post we used TvMaze to retrieve information about various tv shows. After retrieving around 1600 shows, I came to the conclusion that the data was a bit too much – it’s not enough to have the data, we need to be able to know how popular or trendy each show is, in order to provide some form of leveling. One of the lessons I’ve learned from my other games, is that making the questions too hard is a gameplay killer – you can’t expect players to know every isoteric TV show.
So, I decided to look at an alternative data source – TheMovieDb has a very nice API with some C# libraries ready to use.
We’ll use the Top Rated and Popular endpoints to get the most interesting shows. Each show will be leveled by the page where it appeared – shows that appear in the first page will be marked as the easiest. We’ll also rely on the page to level the actors, characters and images – actors and characters will have the level of the show multiplied by 2, and we’ll add the position of the actor/character in the list of actors/characters – that way, secondary actors and characters will have a higher level. Next, we’ll take the images of the actors – we’ll take only images with a vote count of at least 10 votes – we don’t want to have too many images. After that, for each show we’ll also take a look at its seasons. For each season, we’ll take the season itself (leveled to page multiplied by 5 – not that there’s too much to ask about the actual seasons). For each season, we’ll take the episodes – those have names, so there’s a bit more meat to the data which will allow us to ask questions. Each episode has an image, so we’ll add those too. Lastly, each episode also has a list of guest appearances – those are important for us – we want our graph to be as complete as possible, so if an actor appeared in a show, even as a guest, we want to keep that relation. That being said – we’ll set the level of guest actors quite high, because we don’t expect these appearances to be too meaningful. If the actor is well known, her level will be overridden with a lower level where she played a major role. Below is some semi-pseudo code, on how we’re going to do all that:
public async Task GetData()
{
SearchContainerSearchTv container = await m_client.GetTvShowListAsync(tvShowListType, pageNum);
foreach (var serie in container.Results)
{
Console.WriteLine("Getting serie: {serie.Name}");
await HandleSerie(serie.Id);
}
}
public async Task HandleSerie(int serieId)
{
var tvShow = await m_client.GetTvShowAsync(serieId, TvShowMethods.Credits | TvShowMethods.ExternalIds | TvShowMethods.Images | TvShowMethods.Keywords);
AddSerie(serie);
if (tvShow.FirstAirDate != null)
{
AddYear(tvShow.FirstAirDate.Value.Year);
AddRelation(new SerieWasReleasedInYear(serie, year));
}
var tvShowImages = await m_client.GetTvShowImagesAsync(tvShow.Id);
// We'll take only the top 5 backdrops
var imageDatas = tvShowImages.Backdrops.OrderByDescending(data => data.VoteCount*data.VoteAverage).Take(5).ToList();
for (int index = 0; index < imageDatas.Count; index++)
{
var image = imageDatas[index];
AddImage(image);
AddRelation(new SeriePicturedInImage(serie, backdrop));
}
if (tvShow.Credits != null)
{
if (tvShow.Credits.Cast != null)
{
for (int index = 0; index < tvShow.Credits.Cast.Count; index++)
{
AddActor(cast);
AddRelation(new ActorPlayedInSerie(actor, serie));
var images = await m_client.GetPersonImagesAsync(cast.Id);
var profiles = images.Profiles.OrderByDescending(data => data.VoteCount * data.VoteAverage).Take(5).ToList();
for (int index2 = 0; index2 < profiles. Count; index2++)
{
var image = profiles[index2];
AddImage(image);
AddRelation(new ActorPicturedInImage(actor, image));
}
if (cast.Character != null)
{
AddCharacter(cast);
AddRelation(new SeriePortrayedCharacter(serie, character));
AddRelation(new SeriePortraysCharacterPlayedByActor(serie, character, actor));
}
}
}
if (tvShow.Seasons != null)
{
for (int index = 0; index < tvShow.Seasons.Count; index++)
{
var tvSeason = tvShow.Seasons[index];
var seasonDetails = await m_client.GetTvSeasonAsync(tvShow.Id, tvSeason.SeasonNumber);
AddSeason(tvSeason);
AddRelation(new SerieHasSeason(serie, season));
for (int index2 = 0; index2 < seasonDetails.Episodes.Count; index2++)
{
var seasonEpisode = seasonDetails.Episodes[index2];
AddEpisode(seasonEpisode);
AddRelation(new EpisodeIsInSerie(episode, serie));
AddImage(seasonEpisode);
AddRelation(new EpisodeOfSeriePicturedInImage(episode, serie, episodeImage));
for (int index3 = 0; index3 < seasonEpisode.GuestStars.Count; index3++)
{
AddActor(guest);
AddRelation(new SerieFeaturesActor(serie, guestActor));
AddCharacter(cast);
AddRelation(new SeriePortraysCharacterPlayedByActor(serie, guestCharacter, guestActor));
}
}
}
}
}
}
One thing you’ll notice is that I’m adding only one direction for each relation. The reason for that is that each relation has an Invert() method, which will return any complementary relations. For example, the SeriePortraysCharacterPlayedByActor relation can be completed with SeriePortrayedCharacter, CharacterAppearsInSerie, etc. to get me the complete list of relations.