public static void ClasifyText(string text)
{
// Google NLP Api Implementation
const string googleAPIKey = "";
const string googleNLPUrl = "https://language.googleapis.com/v1/documents:classifyText?key=" + googleAPIKey;
HttpClient googleClient = new HttpClient
{
BaseAddress = new Uri(googleNLPUrl)
};
var data = new NLPRequestModel(text);
var stringPayload = JsonConvert.SerializeObject(data);
var httpContent = new StringContent(stringPayload, Encoding.UTF8, "application/json");
HttpResponseMessage response = googleClient.PostAsync(googleNLPUrl, httpContent).Result;
if (response.IsSuccessStatusCode)
{
var responseContent = response.Content.ReadAsStringAsync().Result
.Replace("\\", "")
.Trim(new char[1] {
'"'
});
NLPResponse result = JsonConvert.DeserializeObject <NLPResponse>(responseContent);
// Find the best confidence category
double bigestConfidence = 0.0;
NLPResponseCategory bestCategory = result.Categories.FirstOrDefault();
foreach (NLPResponseCategory category in result.Categories)
{
if (bigestConfidence < category.Confidence)
{
bestCategory = category;
}
}
// Set up in the database the category
//
}
}
public string ProcessMessage(string message)
{
string response = string.Empty;
NLPResponse nlpresponse = NLPService.Ask(message);
List <Movie> searchResult;
switch (nlpresponse.IntentName)
{
case NLPIntentNames.DefaultFallbackIntent:
response = nlpresponse.Message;
break;
case NLPIntentNames.DefaultWelcomeIntent:
response = nlpresponse.Message;
break;
case NLPIntentNames.FavouriteGenresIntent:
try
{
saveUserPreferences(nlpresponse.Parameters["Genre"]);
response = nlpresponse.Message;
}
catch (Exception exc)
{
response = exc.Message;
}
break;
case NLPIntentNames.GetMovieDescriptionContIntent:
case NLPIntentNames.GetMovieDescriptionIntent:
searchResult = movieRepository.SearchByTitle(nlpresponse.Parameters["MovieName"][0]);
if (searchResult.Count == 0)
{
response = $"Sorry, I know nothing about the film {nlpresponse.Parameters["MovieName"][0]}";
}
else
{
foreach (var movie in searchResult)
{
response += $"{movie.Title}\tDescription:\n{movie.Plot}\n\n";
}
}
break;
case NLPIntentNames.GetMovieRateContIntent:
case NLPIntentNames.GetMovieRateIntent:
searchResult = movieRepository.SearchByTitle(nlpresponse.Parameters["MovieName"][0]);
if (searchResult.Count == 0)
{
response = $"Sorry, I know nothing about the film {nlpresponse.Parameters["MovieName"][0]}";
}
else
{
foreach (var movie in searchResult)
{
response +=
$"{movie.Title}\tMetacritic: {movie.MetacriticRate}\tTomato: {movie?.TomatoRate?.Rating.ToString() ?? "N/A"}\t " +
$"Imdb: {movie?.ImdbRate?.Rating.ToString() ?? "N/A"}\n";
}
}
break;
case NLPIntentNames.GetMovieReleaseDateContIntent:
case NLPIntentNames.GetMovieReleaseDateIntent:
searchResult = movieRepository.SearchByTitle(nlpresponse.Parameters["MovieName"][0]);
if (searchResult.Count == 0)
{
response = $"Sorry, I know nothing about the film {nlpresponse.Parameters["MovieName"][0]}";
}
else
{
foreach (var movie in searchResult)
{
response += $"{movie.Title}\tRelease Date: {movie.Year}\n";
}
}
break;
//Obsolete
/*case NLPIntentNames.GetMoviewReviewIntent:
* break;
* case NLPIntentNames.GetMoviewReviewContIntent:
* break;
*/
case NLPIntentNames.GetMoviesByGenreIntent:
var genreToFind = nlpresponse.Parameters["Genre"][0];
searchResult = movieRepository.SearchByGenre(new List <string>()
{
genreToFind
},
new PaginationRequest()
{
Page = 1, Size = 1000
}).Data;
//remove already recommended
var alreadyRecommended = prologService.Execute(PrologQueryFactory.Recommended(Username)).Value;
searchResult = searchResult?.Where(x =>
!alreadyRecommended.Any(y => y.Response.ToLower() == x.Title.ToLower()))?.ToList();
if (searchResult.Count > 0)
{
for (int i = 0; i < Math.Min(searchResult.Count, 5); i++)
{
response += $"{i + 1}. {searchResult[i].Title}";
}
}
else
{
response = "Sorry I can not find some fresh films for you :(";
}
break;
case NLPIntentNames.GetTopRatedMoviesIntent:
List <Movie> movies = movieRepository.FindAll().OrderByDescending(x => x.ImdbRate.Rating)
.ToList();
//remove recommended
var alreadyRecommendedTOPRated =
prologService.Execute(PrologQueryFactory.Recommended(Username)).Value;
movies = movies?.Where(x =>
!alreadyRecommendedTOPRated.Any(y => y.Response.ToLower() == x.Title.ToLower()))?.ToList();
if (movies.Count == 0)
{
response = "Sorry, I have no fresh films for you :(";
break;
}
movies = movies.Take(Math.Min(movies.Count, 5)).ToList();
response = "I would recommend:\n";
for (int i = 0; i < movies.Count; i++)
{
response +=
$"{i + 1}. {movies[i].Title} - Imdb: {movies[i].ImdbRate.Rating};\tMetacritic: {movies[i].MetacriticRate};\tTomato: {movies[i].TomatoRate.Rating}\n";
prologService.Save(PrologRuleFactory.Recommended(Username, movies[i].Title));
}
break;
case NLPIntentNames.ThanksIntent:
response = nlpresponse.Message;
break;
case NLPIntentNames.WantWatchMovieIntent:
List <string> preferences = checkUserPreferences();
if (preferences.Count == 0)
{
response = "I don't know about your preferences. Please tell me what genres do you prefer.";
}
else
{
var rec = recommendRandomMovieByPreferences(preferences);
response = rec.Length > 0
? $"I would like to recommend {rec}"
: "I already recommended you all films. Please name me another genres you might likes.";
}
break;
//case NLPIntentNames.WantWatchSpecificMovieIntent:
// break;
default:
response = "Sorry, I don't understand you :(";
break;
}
return(response);
}