public static async Task RunAsync([BlobTrigger("resumeuploadcontainer/{name}", Connection = "AzureWebJobsStorage")] Stream myBlob, string name, Uri uri, TraceWriter log)
{
log.Info($"C# Blob trigger function Processed blob\n Name:{name} \n Size: {myBlob.Length} Bytes");
string[] spiltfilenamewithext = name.Split('.');
string filenamewithoutextension = spiltfilenamewithext[0];
if (IsValidExtension(spiltfilenamewithext[spiltfilenamewithext.Length - 1]))
{
PdfReader pdfReader = new PdfReader(myBlob);
ResumeDocModel resumeDocModel = new ResumeDocModel {
DocumentName = filenamewithoutextension, DocumentUri = uri.AbsoluteUri
};
CognitiveTextAnalytics.TextAnalytics(pdfReader, log, resumeDocModel);
string[] resumeKeysArray = ResumeKeyWords.Split(new string[] { "," }, StringSplitOptions.None);
bool isResumeDoc = false;
foreach (string keys in resumeKeysArray)
{
isResumeDoc = resumeDocModel.keyPhraseBatchResult.KeyPhrases.Any(s => s.IndexOf(keys, StringComparison.OrdinalIgnoreCase) >= 0);
if (isResumeDoc)
{
resumeDocModel.docType = DocType.Resume;
break;
}
}
if (!isResumeDoc)
{
resumeDocModel.docType = DocType.Others;
}
List <CloudBlockBlob> cloudBlocks = DocumentExtraction.ExtractImageUploadToAzure(pdfReader, myBlob, log, filenamewithoutextension, resumeDocModel);
if (cloudBlocks.Count > 0)
{
log.Info("===== Computer Vision Analysis Started ======");
ImageBatchResult imageBatchResult = new ImageBatchResult();
foreach (CloudBlockBlob cloudBlockBlob in cloudBlocks)
{
string blobUrlWithSAS = AzStorage.GetBlobSasUri(cloudBlockBlob);
await CognitiveComputerVision.VisionAnalyticsAsync(log, blobUrlWithSAS, imageBatchResult);
}
resumeDocModel.imageBatchResult = imageBatchResult;
log.Info("===== Computer Vision Analysis Completed ======");
}
else
{
log.Info("The Document doesn't not have any Images to analyze");
}
CosmosDB cosmosDB = new CosmosDB();
cosmosDB.CreateDocumentDB().Wait();
await cosmosDB.UpdInsResumeDocumentAsync(resumeDocModel);
}
else
{
log.Info("Please upload the valid Document in " + string.Join(",", _validExtensions) + " Extension");
}
}
public async Task UpdInsResumeDocumentAsync(ResumeDocModel resumeDoc)
{
FeedOptions queryOptions = new FeedOptions {
MaxItemCount = 1, EnableCrossPartitionQuery = true
};
List <ResumeDocModel> resumeQueryResult = client.CreateDocumentQuery <ResumeDocModel>(
UriFactory.CreateDocumentCollectionUri(DatabaseId, CollectionId),
"SELECT * FROM Resume WHERE Resume.DocumentName='" + resumeDoc.DocumentName + "'", queryOptions).ToList();
if (resumeQueryResult.Count() == 0)
{
await this.client.CreateDocumentAsync(UriFactory.CreateDocumentCollectionUri(DatabaseId, CollectionId), resumeDoc);
}
else
{
resumeDoc.id = resumeQueryResult.First().id;
await this.client.ReplaceDocumentAsync(UriFactory.CreateDocumentUri(DatabaseId, CollectionId, resumeQueryResult.First().id), resumeDoc);
}
}
public static void TextAnalytics(PdfReader pdfReader, TraceWriter log, ResumeDocModel resumeDocModel)
{
// Create a client.
ITextAnalyticsClient client = new TextAnalyticsClient(new ApiKeyServiceClientCredentialsForText())
{
Endpoint = TextAnalyticsEndPoint
}; //Replace 'westus' with the correct region for your Text Analytics subscription
Console.OutputEncoding = System.Text.Encoding.UTF8;
// Extracting language
log.Info("===== Text Analytics Started ======");
string content = DocumentExtraction.GetTextFromPDF(pdfReader);
List <string> splittedList = StringExtensions.Split(content, MaxLengthofCharacters).ToList();
var LanguageDetectAPI = client.DetectLanguageAsync(new BatchInput(
new List <Input>()
{
new Input(resumeDocModel.DocumentName, splittedList.First())
})).Result;
resumeDocModel.languageBatchResult = LanguageDetectAPI.Documents.FirstOrDefault();
var detectedLanguage = LanguageDetectAPI.Documents.Select(doc => doc.DetectedLanguages[0].Iso6391Name).FirstOrDefault();
//SentimentBatchResult result3 = client.SentimentAsync(
// new MultiLanguageBatchInput(
// new List<MultiLanguageInput>()
// {
// new MultiLanguageInput(detectedLanguage, resumeDocModel.DocumentName, splittedList.First())
// })).Result;
//resumeDocModel.sentimentBatchResult = result3.Documents.FirstOrDefault();
List <string> keyPhraseList = new List <string>();
List <EntityRecordV2dot1> entityRecords = new List <EntityRecordV2dot1>();
foreach (string splittedContent in splittedList)
{
KeyPhraseBatchResult keyPhraseBatch = client.KeyPhrasesAsync(new MultiLanguageBatchInput(
new List <MultiLanguageInput>()
{
new MultiLanguageInput(detectedLanguage, resumeDocModel.DocumentName, splittedContent)
})).Result;
foreach (var doc in keyPhraseBatch.Documents)
{
keyPhraseList.AddRange(doc.KeyPhrases.ToList());
}
EntitiesBatchResultV2dot1 entitiesbatchres = client.EntitiesAsync(
new MultiLanguageBatchInput(
new List <MultiLanguageInput>()
{
new MultiLanguageInput(detectedLanguage, resumeDocModel.DocumentName, splittedContent)
})).Result;
entityRecords.AddRange(entitiesbatchres.Documents.First().Entities.ToList());
}
resumeDocModel.keyPhraseBatchResult.Id = resumeDocModel.DocumentName;
resumeDocModel.keyPhraseBatchResult.KeyPhrases = keyPhraseList;
resumeDocModel.entityBatchResult.Id = resumeDocModel.DocumentName;
resumeDocModel.entityBatchResult.EntityRecords = entityRecords;
log.Info("===== Text Analytics Completed ======");
}
public static List <CloudBlockBlob> ExtractImageUploadToAzure(PdfReader pdfReader, Stream blob, TraceWriter log, string name, ResumeDocModel resumeDocModel)
{
RandomAccessFileOrArray raf = new RandomAccessFileOrArray(blob);
List <CloudBlockBlob> cloudBlockBlobs = new List <CloudBlockBlob>();
AzureStorageModel azureStorageModel = AzStorage.GetBlobDirectoryList(name);
if (azureStorageModel.IblobList.Count() > 0)
{
foreach (IListBlobItem listBlobItem in azureStorageModel.IblobList)
{
CloudBlockBlob cloudBlob = listBlobItem as CloudBlockBlob;
cloudBlob.DeleteIfExists();
}
}
try
{
for (int pageNumber = 1; pageNumber <= pdfReader.NumberOfPages; pageNumber++)
{
PdfDictionary pg = pdfReader.GetPageN(pageNumber);
List <Image> IlistImages = GetImagesFromPdfDict(pg, pdfReader);
EncoderParameters parms = new EncoderParameters(1);
parms.Param[0] = new EncoderParameter(System.Drawing.Imaging.Encoder.Quality, 90L);
var encoder = ImageCodecInfo.GetImageEncoders().First(c => c.FormatID == ImageFormat.Jpeg.Guid);
foreach (Image image in IlistImages)
{
MemoryStream ms = new MemoryStream();
CloudBlockBlob cloudBlob = azureStorageModel.Container.GetBlockBlobReference(name + "/" + Guid.NewGuid() + ".jpeg");
image.Save(ms, encoder, parms);
ms.Seek(0, SeekOrigin.Begin); // otherwise you'll get zero byte files
cloudBlob.UploadFromStream(ms);
cloudBlockBlobs.Add(cloudBlob);
resumeDocModel.imageDetails.imageURLList.Add(cloudBlob.Uri.ToString());
}
}
}
catch
{
throw;
}
finally
{
pdfReader.Close();
}
return(cloudBlockBlobs);
}