Exemplo n.º 1
0
        public async Task <TextAnalysisResult> AnalyzeTextAsync(string text)
        {
            // See: https://docs.microsoft.com/en-us/azure/cognitive-services/text-analytics/quickstarts/csharp

            var analysisResult = new TextAnalysisResult();

            if (string.IsNullOrEmpty(text))
            {
                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeTextAsync(): no text to analyze");
                return(analysisResult);
            }

            string textToAnalyze = text;

            if (text.Length > 5000)
            {
                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeTextAsync(): text longer than supported length. Trimming it...");
                textToAnalyze = text.Substring(0, 5000);
            }

            Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeTextAsync(): initializing TextAnalyticsAPI");

            ITextAnalyticsAPI m_TextAnalyticsClient = new TextAnalyticsAPI
            {
                AzureRegion     = AzureRegions.Westeurope,
                SubscriptionKey = m_TextAnalyticsAPISubscriptionKey
            };

            Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeTextAsync(): detecting content language");

            var batchLanguageResult = await m_TextAnalyticsClient.DetectLanguageAsync(new BatchInput(new List <Input>()
            {
                new Input("1", textToAnalyze)
            })).ConfigureAwait(false);

            if (batchLanguageResult.Errors.Count > 0)
            {
                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeTextAsync(): error while detecting language");
                foreach (var errors in batchLanguageResult.Errors)
                {
                    Console.WriteLine($"\t{errors.Message}");
                }
                return(analysisResult);
            }

            analysisResult.DetectedLanguage      = batchLanguageResult.Documents[0].DetectedLanguages[0].Name;
            analysisResult.DetectedLanguageScore = batchLanguageResult.Documents[0].DetectedLanguages[0].Score.GetValueOrDefault();

            Console.WriteLine($"\t\t\tContentAnalyzer.AnalyzeTextAsync(): detected language is '{analysisResult.DetectedLanguage}' ({(analysisResult.DetectedLanguageScore * 100):0.00}%)");

            Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeTextAsync(): performing key-phrase extraction");

            var multiLanguageInput = new MultiLanguageBatchInput(new List <MultiLanguageInput>()
            {
                new MultiLanguageInput(batchLanguageResult.Documents[0].DetectedLanguages[0].Iso6391Name, "1", textToAnalyze)
            });
            var batchKeyphraseResult = await m_TextAnalyticsClient.KeyPhrasesAsync(multiLanguageInput).ConfigureAwait(false);

            if (batchKeyphraseResult.Errors.Count > 0)
            {
                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeTextAsync(): error while extracting key-phrases");
                foreach (var errors in batchKeyphraseResult.Errors)
                {
                    Console.WriteLine($"\t\t\t\t{errors.Message}");
                }
                return(analysisResult);
            }

            Console.WriteLine($"\t\t\tContentAnalyzer.AnalyzeTextAsync(): retrieved {batchKeyphraseResult.Documents[0].KeyPhrases.Count} key-phrases:");
            foreach (var keyphrase in batchKeyphraseResult.Documents[0].KeyPhrases)
            {
                analysisResult.KeyPhrases.Add(keyphrase);
                Console.WriteLine($"\t\t\t\t{keyphrase}");
            }

            Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeTextAsync(): performing sentiment analysis");

            var batchSentimentResult = await m_TextAnalyticsClient.SentimentAsync(multiLanguageInput).ConfigureAwait(false);

            if (batchSentimentResult.Errors.Count > 0)
            {
                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeTextAsync(): error while detecting sentiment");
                foreach (var errors in batchSentimentResult.Errors)
                {
                    Console.WriteLine($"\t\t\t\t{errors.Message}");
                }
                return(analysisResult);
            }

            analysisResult.SentimentScore = batchSentimentResult.Documents[0].Score.GetValueOrDefault();
            analysisResult.Sentiment      = GetSentiment(analysisResult.SentimentScore);

            Console.WriteLine($"\t\t\tContentAnalyzer.AnalyzeTextAsync(): sentiment is '{analysisResult.Sentiment}' ({(analysisResult.SentimentScore * 100):0.00}%)");

            // Extend analysis by estimating reading time for content
            analysisResult.WordCount            = TextTokenizer.GetWordCount(text);
            analysisResult.ReadingTimeInMinutes = ReadingTimeEstimator.GetEstimatedReadingTime(analysisResult.WordCount, analysisResult.DetectedLanguage);

            return(analysisResult);
        }
Exemplo n.º 2
0
        public async Task <ImageAnalysisResult> AnalyzeImageAsync(string url)
        {
            var analysisResult = new ImageAnalysisResult();

            try
            {
                // USING Microsoft provided VisionClientLibrary seems not working in NET Core as-is, a fix is required for ExpandoObject
                // see: https://github.com/Microsoft/Cognitive-Vision-DotNetCore/pull/1/commits/9c4647edb400aecd4def330537d5bcd74f126111

                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeImageAsync(): initializing VisionAPI client");

                var visionApiClient = new VisionServiceClient(m_VisionAPISubscriptionKey, "https://westeurope.api.cognitive.microsoft.com/vision/v1.0");

                var visualFeatures = new List <VisualFeature> {
                    VisualFeature.Adult, VisualFeature.Categories, VisualFeature.Color, VisualFeature.Description, VisualFeature.Faces, VisualFeature.ImageType                                           /*, VisualFeature.Tags */
                };
                var details = new List <string> {
                    "Celebrities", "Landmarks"
                };

                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeImageAsync(): started image analysis");

                var visionApiResult = await visionApiClient.AnalyzeImageAsync(url, visualFeatures, details).ConfigureAwait(false);

                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeImageAsync(): executing OCR");

                var ocrResult = await visionApiClient.RecognizeTextAsync(url).ConfigureAwait(false);

                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeImageAsync(): performing tag identification");

                var tagsResult = await visionApiClient.GetTagsAsync(url).ConfigureAwait(false);

                Console.WriteLine("\t\t\tContentAnalyzer.AnalyzeImageAsync(): analysis completed");

                // Mapping VisionAPI Client entity to domain entity
                analysisResult.AdultContent = new ImageAnalysisAdultContentResult {
                    AdultScore = visionApiResult.Adult.AdultScore, IsAdultContent = visionApiResult.Adult.IsAdultContent, IsRacyContent = visionApiResult.Adult.IsRacyContent, RacyScore = visionApiResult.Adult.RacyScore
                };
                analysisResult.Colors = new ImageAnalysisColorResult {
                    AccentColor = visionApiResult.Color.AccentColor, DominantColorBackground = visionApiResult.Color.DominantColorBackground, DominantColorForeground = visionApiResult.Color.DominantColorForeground, IsBWImg = visionApiResult.Color.IsBWImg
                };
                analysisResult.Categories = visionApiResult.Categories.Select(c => new ImageAnalysisCategoryResult {
                    Text = c.Name, Score = c.Score
                }).OrderByDescending(c => c.Score).ToList();
                analysisResult.Descriptions = visionApiResult.Description.Captions.Select(c => new ImageAnalysisDescriptionResult {
                    Text = c.Text, Score = c.Confidence
                }).OrderByDescending(c => c.Score).ToList();

                // Merge detected tags from image analysis and image tags
                analysisResult.Tags = tagsResult.Tags.Select(t => new ImageAnalysisTagResult {
                    Text = t.Name, Score = t.Confidence, Hint = t.Hint
                }).ToList();
                foreach (var t in visionApiResult.Description.Tags)
                {
                    analysisResult.Tags.Add(new ImageAnalysisTagResult {
                        Text = t, Score = 0.0, Hint = string.Empty
                    });
                }

                analysisResult.Faces = visionApiResult.Faces.Select(f => new ImageAnalysisFaceResult {
                    Age = f.Age, Gender = f.Gender == "Male" ? Gender.Male : f.Gender == "female" ? Gender.Female : Gender.Unknown
                }).ToList();
                analysisResult.Text = ocrResult.Regions.Select(r => new ImageAnalysisTextResult()
                {
                    Language = ocrResult.Language, Orientation = ocrResult.Orientation, TextAngle = ocrResult.TextAngle.GetValueOrDefault(), Text = string.Join(" ", r.Lines.Select(l => string.Join(" ", l.Words.Select(w => w.Text))))
                }).ToList();

                // Extend analysis by estimating reading time for each transcribed text
                foreach (var text in analysisResult.Text)
                {
                    text.WordCount                        = TextTokenizer.GetWordCount(text.Text);
                    text.ReadingTimeInMinutes             = ReadingTimeEstimator.GetEstimatedReadingTime(text.WordCount, text.Language);
                    analysisResult.WatchingTimeInMinutes += text.ReadingTimeInMinutes;
                }

                // Add an additional default time for estimating how long it will take to the user to watch the picture
                analysisResult.WatchingTimeInMinutes += DefaultImageWatchingTime;
            }
            catch (Exception ex)
            {
                Console.WriteLine($"\t\t\tContentAnalyzer.AnalyzeImageAsync(): an error occured while analyzing image - {ex.Message}");
            }

            return(analysisResult);
        }